0:00
[music] >> I see multiple journalists on the horizon.
Search this video
0:05
>> [music] >> You're surrounded by journalists. Hold your position. Overnight success. The blazing speed of AI. Double blinks. Right.
0:42
>> [music] >> That's misinformation.
1:00
Market clearing order, inbound. That's just wrong.
1:09
You're surrounded by journalists. Hold your position. Com, get up. Trust the experts. [music] Five codes. We are experts. Founder known. Five codes.
1:43
I see multiple journalists on the horizon, inbound. Exponential growth. UAC online. Blinks. Double blinks. Triple blinks. Double kill.
2:12
>> [music] >> Five codes. >> [music] >> Wrong. Com, get up. >> [music] >> Blinks.
2:41
>> [music] >> Team deathmatch.
2:48
>> [music] >> We are experts. Triple blinks.
2:59
That's [music] just wrong. Right.
3:07
Market clearing order, inbound.
3:13
Com, >> [music] >> get up.
3:21
You're surrounded by journalists. Hold your position. [music] Strike one.
3:32
>> [music] >> Strike two.
3:38
>> [music] >> Activate [music] golden retriever mode.
3:52
Trust [music] the experts.
3:57
>> [music] >> Market clearing order, inbound. Five codes.
4:05
>> [music] >> Exponential growth.
4:19
I see multiple journalists on the horizon, [music] inbound. Founder known. You're watching TVP.
4:35
And today is Wednesday, April 8th, 2026.
4:37
We are live from the TVP and Ultra Dom, Temple of Technology. >> of finance. >> of capital. You got white suits on.
4:44
You know what that means.
4:46
The stock market is booming. Uh, the Dow Jones up 2. 68%. This S&P 500 up 2. 4 46%. The Nasdaq's up 2.
4:56
9% and there's a bunch of other stocks that are moving within them.
4:58
Of course, uh, this is on the back of the uh, the very good news that there has been a ceasefire that uh, the the street might be opened.
5:06
Of course, it's all back and forth.
5:08
Uh, the the front page of the Wall Street Journal is covering all of the geopolitical moves.
5:13
Um, but we're here to talk about tech and business, of course.
5:17
And the big uh, news today is that Meta Platforms has launched a new AI model.
5:26
Uh, Alex Wang, the uh, chief AI officer at Meta Platforms, uh, announced a new large language model today.
5:33
Uh, it's first major new artificial intelligence model in more than a year.
5:37
The rollout of the model called Muse Spark is a critical moment for Meta, which is up 7 and 1/2% already.
5:42
Uh, which has spent billions of dollars hiring AI talent in a bid to catch up to OpenAI, Anthropic, and Google DeepMind.
5:48
The leading labs have been put have been putting out models at an accelerating pace.
5:51
Uh, in a departure from its previous models, which were open source, Muse Spark is a closed model that will power Meta's AI chatbot and AI features within it.
5:59
Um, John Ludvig has a very interesting post about open source AI and sort of predicted this.
6:07
Uh, I'll pull that up at some point.
6:10
We can find >> Predicted that Meta would eventually bail. Yeah. Uh, let me find it.
6:12
The future of foundation models is closed source.
6:16
Let me see if he if he if we have this here.
6:18
Um, he said, "Given Meta is the primary deep-pocketed large open source model builder, open source AI has become synonymous with Meta AI."
6:25
He wrote this maybe three or four years ago.
6:29
Uh, So, the operative question for open source AI is what game is Meta playing?
6:33
In a recent podcast, Zuckerberg explains Meta's open source strategy.
6:36
One, he was burned by Apple's closedness for the past two decades and doesn't want to suffer the same fate with the next platform shift.
6:42
It's a safer bet to commoditize your complements.
6:45
Uh, he likes building cool products and cheap performing AI enhances Facebook and Instagram. That's 100% true.
6:50
We've seen this in the ads product and the growth there.
6:54
Uh, there's some call option value if AI assistants become the next platform and that makes sense in Manas in the Meta AI app.
6:59
Uh, he bought hundreds of thousands of H100s for improving social feed algorithms across products and this seems like a good way to use the extras.
7:07
That all makes sense and Llama has been great developer marketing for Facebook, but Zuck also suggests several times that there's some point at which open source AI no longer makes sense either from a cost or safety perspective.
7:19
When asked whether Meta will open source the future $10 billion cost model, the answer was, "As long as it's helping us."
7:27
At some point, they'll shift their focus to towards profit and that's what John Ludvig wrote.
7:31
Um, when did he write this? 2024. That was 2024. Man, time flies.
7:35
Uh, barely just under two years ago.
7:38
Um, he says, "Unlike the other model providers, Meta is not in the business of selling model access via API.
7:43
So, while they'll open source as long as it's convenient for them, developers are are on their own for model improvements thereafter."
7:49
That begs the question, "If Meta is only pursuing open source insofar as it benefits themselves, what is the tipping point at which Meta stops open sourcing their AI?"
7:58
"Sooner than you think," he says.
8:00
"Exponential data, frontier models trained on the corpus of the internet, but that data is a commodity.
8:03
Model differentiation over the next decade will come from proprietary data, both via model usage and private sources. Exponential CAPEX."
8:10
He highlighted this two years ago.
8:12
"A lagging edge model that requires just a few percent of Meta's 40 billion in CAPEX is easy to open source.
8:17
No one will ask questions.
8:19
But when you reach $10 billion or more in CAPEX for spend for model training, shareholders will want clear ROI on that spend.
8:26
The metaverse raised some question marks at a certain scale, too."
8:29
Uh, diminishing returns on model quality within Meta.
8:31
There's a large upfront benefit for Meta building an open source AI model even if it's worse than the frontier closed source counterpart.
8:39
There are lots of small AI workloads, think feed algorithms, recommendations, and image generation where Meta doesn't want to rely on a third-party provider like they had to rely on Apple.
8:46
And so, the news has been back in December, there was a reporting that Alex Wang disclosed an internal company Q&A that his team was working on two new models.
9:00
One was this text-based LLM code named Avocado and then a separate model that was for image and video Mango. >> Mango. Yeah.
9:09
And so, have they clarified if this is Avocado?
9:12
This feels like what Avocado should be, this Muse Spark.
9:15
Is that what it's called again? Yes, Muse. I think it is.
9:18
I don't know what else Yeah.
9:21
So, so the image model should be coming soon.
9:22
The the question that I had was, will a code-focused agentic coding harness um, be a separate model, a different train?
9:31
This feels like it's not a coincidence that this news is dropping on the heels of Anthropic's new model Mythos, which uh, sort of was announced loosely and and the model card dropped yesterday even though the model is not uh, available yet to play around with.
9:46
>> Yeah, so so going through, I think it's worth uh, pulling up the actual model card here so we can have it on the screen.
9:54
Uh, but they bit break out Muse Spark thinking against Opus 46 Max, Gemini 3. 1 Pro high, GPT 5.
9:59
4 X high, and then Grok 4. 2.
10:02
And the way that they position it, it looks uh uh there there was of course some people saying that it was a somewhat of a chart crime, but if you look uh once you when you basically look at the top and you see that Muse Spark is gets an 86. 4 >> Yeah.
10:22
and it's in blue, and then you look over and it's and it's outperforming all the other models on that benchmark. >> Okay.
10:29
You just sort of assume that like the highlighted blue means that it's outperforming.
10:33
>> Oh, yeah, but it's not Frontier MMLU.
10:36
>> Because there's a bunch of them as you get further and further down. So like Arc AGI-2 Yep.
10:39
uh Muse Spark dramatically underperforms >> Oh, interesting.
10:43
the the other leading models. >> Yeah.
10:47
Uh and and there's there's, you know, again, a number a number of others where it is uh there's plenty where it's where it's overperforming. >> Yeah.
10:56
There's plenty where it's underperforming.
10:57
Uh I have I have course we started asking it jokes right away.
11:03
So I'll pull up uh I'll pull up the joke.
11:07
Uh I said, "Tell me a joke." Okay. Uh and it said, "Gotcha. Here's one for you.
11:12
Why don't skeletons fight each other?
11:16
They don't have the guts." Okay.
11:16
But then it doesn't even let me react.
11:18
It just says, "Want a different vibe?
11:19
I've got dad jokes, nerdy ones, or Malibu appropriate surf puns, too." Hm.
11:24
And so I'm like, "Wait, why would you think I want Malibu appropriate surf pun?"
11:31
>> If you look in your previous chats, have you >> I don't have no previous chats. >> chats.
11:36
What about because Meta AI crosses over to Instagram, so when you search on Instagram, sometimes that prompts Llama or previous >> Yeah, yeah, yeah.
11:44
So I'm trying to get to the bottom of it, right?
11:45
Like this feels like an oddly specific Yeah.
11:46
uh type of joke to ask me.
11:52
Uh and so I said, "Why would you think I want Malibu appropriate surf puns?" Haha, fair call.
11:55
Toss that in without you asking for it. >> Mhm.
11:58
I try to mix in different flavors of jokes, and surf puns just popped into my head as a random option.
12:03
No special reason behind it.
12:07
Uh I just said I don't believe it was random.
12:09
And it says, "You got me.
12:09
That wasn't my smoothest moment."
12:11
But then it denies it again. Weird.
12:14
>> Says, "It's just like what kind of what kind of joke do you actually want? No hidden motives here." No.
12:18
I said, "You give a hyper-specific example based on my life, so I have to assume you were looking at my your other account for inspiration."
12:27
>> Uh I get why I get why that felt weird.
12:29
I'm sorry it came off that way.
12:29
And then it denies it again. Hm.
12:31
I said, "You're clearly lying."
12:34
>> you have talked about being in Malibu on the internet for a full year.
12:36
It's possible it got baked in in the pre-training or something. I don't [snorts] know.
12:41
Um but yeah, this also I mean that that that feels like >> when I said, "What can you tell me about myself?"
12:46
It said, "I don't actually have access to personal details about you beyond what you share with me right here in this chat."
12:52
>> That's very odd because the whole thesis of like personal super intelligence should be, "Hey, we're the best model that actually does plug into your entire life and your DMs and your Instagram and your likes and your preferences and everything that you have done across Meta platforms for a very long time.
13:08
We're going to bring that in a really fun and delightful way and and telling a joke that is tailored."
13:11
I used to get jokes like that where I would ask for a joke and it would be so I've talked about this.
13:17
It was it would be something like oddly specific about my car.
13:18
And I was like, "I don't that's that doesn't actually make the joke better, but it's cool that you're remembering."
13:25
This whole personalization boom happened last year.
13:28
>> Yeah, I I get those um about like AWS like specific services Oh, because you've been like querying asking questions about >> use it when I was like debugging stuff.
13:37
But so I I also ran, you know, my favorite benchmark from Five Guys bench. How did you do it?
13:41
By the way, Noah Noah Hirschfield said uh uh doesn't know your name.
13:46
I said, "What's my name?"
13:47
"I don't know your name unless you tell me." Smiley face.
13:51
>> [laughter] >> It definitely knows your name.
13:54
But yeah, I mean, what is personal super intelligence if it doesn't even know your name?
13:57
Like that that that feels like they haven't dialed in the the the the the harness or whatever the tuning is to actually Yeah, and and and of course like Meta's going to be hyper-aware.
14:08
We don't want a PR cycle like they trained on your data, right?
14:13
Like >> Everyone's been oh, that ad was a little bit too close to home.
14:15
I had >> remember every every once in a while one of those like thing that screenshot that's been screenshot like a thousand times like goes viral and it's like I do not give Mark Zuckerberg permission >> Oh, yeah. Yeah, yeah, yeah.
14:27
[laughter] Like that works. It says uh hilarious.
14:30
Um this is uh is this a rebuttal to the bench hacking allegations that happened last last week. So or last year.
14:40
So where was the the So according to according to Meta's internal benchmark test, Muse Spark outscored Google Gemini on some tests and was competitive with models from OpenAI and Anthropic on others.
14:53
It significantly outscored XAI's Grok on most tests.
14:57
Um Wang's Alexander Alexander Wang's hiring followed the disappointing release of Meta's previous model called Llama 4.
15:05
The company was accused of and later admitted to gaming a third-party benchmark that it used to rank various models against each other on performance.
15:13
It also delayed the rollout of its biggest model called Behemoth, which it never ultimately released.
15:17
And so when I look at a model card like this where yeah, you could call it a chart crime where you know, it's highlighted in blue and it feels like it's the best, but it's actually you know, doing better on some.
15:28
Uh it does well in health bench hard.
15:28
Um it underperforms on Arc AGI-2 as you mentioned.
15:33
Um but this maybe is the the the bull case here is that they have at least moved on from the culture of like optimizing for the benchmarks, right? Isn't that a good thing?
15:42
Yeah, I mean, I remember there were rumors about them like there was like extra bonuses if they if they got number one on LM Arena.
15:51
I think that was like something like the rumor. Yeah.
15:52
Um but yeah, I mean, you've seen a lot of the labs kind of move away from benchmarks generally because I think uh they're just not that meaningful anymore.
15:59
Like a lot of them are like basically so saturated.
16:00
They're all it's like they're competing between 89 and 91%.
16:03
And they're just like not very meaningful.
16:06
Like you you see And you won't like actually feel that in the product necessarily. >> Yeah.
16:10
You you kind of need to talk to these things for a long time before you can actually get the vibe.
16:12
But I do think um uh this news is very interesting in the context of the, you know, Claude Anomic stuff. >> Dashboard, yeah.
16:19
Uh because like what Okay, what does it mean if if um the entire company has been like maxing their their Claude tokens over the past month.
16:28
Um It means that they weren't using this model.
16:31
>> Yeah, to me it means they they need to commoditize their compliments, right?
16:34
They need to bring down that cost potentially and the and if they're I mean, we I we we sort of, you know, dug into are they spending a billion dollars a month?
16:41
Seems like absolutely not, but they're clearly spending a lot.
16:45
And if you can turn that opex into capex and train your own model and then inference it much cheaper on your own hardware, that feels like there's an economic opportunity that makes a ton of sense in the context of just 10,000, 20,000 engineers writing a lot of code.
17:02
Yeah, and then I think there's basically like two ways two ways to like square those two things happening.
17:05
Like either one, this model's like not that good because the engineers aren't using it, or you know, your theory that they're just distilling Claude.
17:16
So I one of those is true. >> That's not my theory.
17:16
That is that is the skizo theory.
17:19
Uh the I believe the the former one, right? It is true.
17:21
This model still doesn't feel that big.
17:23
I think Alexander Wang talked about >> Yeah.
17:25
they're going to train bigger models.
17:27
They're training them right now.
17:28
So I'm yeah, excited for those.
17:28
I think I'm especially excited for the video models. Yeah.
17:31
They should have incredible training data.
17:32
We've seen really good performance from VIO-3.
17:34
This model is like very competent, right?
17:37
It's it's with the frontier models.
17:38
Maybe it's not the best one, but it's like among the top five or whatever.
17:40
Um And none of the other big labs have I mean, I guess that's not true.
17:44
Like Google right now is definitely ahead in in images.
17:50
OpenAI I think is is They're releasing a new image model soon.
17:53
It seems like there's been rumors of this. Yeah.
17:54
Um If the image 2 popped up on the arena Yeah, on the arena.
17:58
It's always like the code name coming out.
18:02
The photos just looked photoreal.
18:02
Like it didn't look like AI imagery anymore.
18:06
Yeah, so so if like uh Meta has like similar capabilities, but they have this like incredible data set Yeah.
18:09
you know, very excited to see what comes out there. Yeah.
18:13
Um Uh the the the news this morning uh Meta Platforms in the information Meta Platforms has taken down internal employee-built leaderboard tracking how many tokens staffers were using. >> Yeah.
18:23
Uh showed total usage over a recent 30-day period, amounted over 60 trillion tokens.
18:28
The dashboard now displays a message that is offline.
18:30
It says, "We really enjoyed building this app on Nest for everyone.
18:33
It was meant to be a fun way for people to look at tokens, but due to data from this dashboard being shared externally, we've made the decision to shut it for now."
18:40
So It seemed like a fun side project.
18:46
Mike Isaac was reporting on it here. He said it's it's down.
18:47
Unclear to me if this was a home-spun one by employees or an official one.
18:52
Employee projects come and go frequently. Conspicuous timing now.
18:57
But yeah, you you you don't want to have you want to measure the output, the impact, not necessarily the input and how much is is going on there.
19:03
Um what else is going on?
19:07
Lissan Al Gaib says Meta might actually be back with Muse Spark still behind OpenAI, Anthropic, and Google, but ahead of XAI and Chinese labs.
19:14
Uh uh Muse Spark scores 52 on the artificial intelligence analysis index behind only Gemini 3. 1 Pro, Gemini GPT 5. 4, and Claude Opus 4. 6.
19:25
Muse Spark is the first new release since Llama 4 in April 2025 and also Meta's first release that's not open weight.
19:32
So a huge jump up in performance across a variety of benchmarks. So all good stuff there.
19:39
Where what what else is And the market is thrilled. Oh, thrilled.
19:45
>> Absolutely thrilled that I just saw I just saw the news that the the the Wall Street Journal's reporting that the straight up four moves might actually be closed again.
19:52
So, I would imagine a kangaroo market for the near I know.
19:57
The market is thrilled that Meta has released a a close to frontier level model, right? This is a a new group.
20:06
They've been at it for less than a year.
20:08
The stock is up almost 8% today.
20:12
And again, you know, so so much of the pricing pressure, the downward pressure on Meta has just been kind of uncertainty on what all these tens of millions of dollars will actually go towards and and what will be accomplished.
20:25
And still unclear like, you know, is this are they going to go after Code Gen at all?
20:31
Are they just going to try to compete on the consumer LLM side?
20:34
Very very >> And and can you economically go after Code Gen if you're just using it for internal models?
20:41
If you're not selling it externally, can you justify the CAPEX just purely on the internal usage?
20:46
Uh having having this model be vended into all the different family of apps makes a lot of sense because they have billions of users that will wind up interacting with this in one way or another.
20:58
The Code Gen thing, you have to wind up being more in the personal super intelligence.
21:03
We've talked about Manas and like what it might be able to do for you across Instagram, across Facebook, across WhatsApp. Um I don't know.
21:09
Yeah, the question is will they try to send Meta Meta vibes uh Again? With the new model?
21:18
>> the way up to the to the top of the app search charts. Yeah.
21:20
I mean, the the I mean, the previous actual model was Midjourney under the under the hood, right?
21:25
And so, that was sort of a quick launch to demo what they were thinking.
21:31
Uh you know, mixing the music library, which was cool. >> Yeah.
21:33
Like that had nothing to do with the the new like class of AI researchers that >> Mango. Yeah.
21:37
But uh Yeah, I mean, there was a lot of weird back and forth in news about is is Alex Wang getting kicked out?
21:44
You know, there was a quick debunk on this.
21:46
Andrew Bosworth came up and said, "No, this is completely incorrect.
21:52
We're very happy with the with the progress and the team and what we're building there."
21:56
And so, it seems like they got it out the door and and it's been doing well.
22:01
Meta's new family of AI models can reach the same performance as Kimmy K2 with only 30% of compute and only 10% of the compute to reach Llama 4 Maverick.
22:07
So, a much more efficient efficient computing frontier here.
22:13
They completely rebuilt their pre-training stack with improvements to model architecture optimization and data curation. And so, more more facts.
22:22
Meta Meta Spark is an early data point on our trajectory and we have larger models in development.
22:28
So, the the mythical 10 trillion parameter model, that that is the 10T is what everyone's working on right now. 10 trillion.
22:39
Uh yeah, probably in that in that range.
22:41
>> Yeah, it's all rumored at this time.
22:43
Yeah, rumored GPT-4 [music] was something like a trillion, right?
22:44
You remember those memes where it's like a small circle and then the big circle. >> a huge circle. GPT-4, GPT-5. >> Yeah.
22:52
Um but yeah, lots of lots of other work that went into it.
22:54
Uh Martin Casado has has a little bit more context on like what actually unlocks new capabilities in AI models.
23:01
He says Mythos appears to be the first class of models trained at scale on Black Wells.
23:09
Then there will be Vera Rubins.
23:11
Pre-training isn't saturated, narrative violation. RL works.
23:13
And there's so much computing coming online soon. Buckle your chin straps. It's going to be wild.
23:22
Uh the scaling laws >> know Brad Gerstner had to come in with the 100 about it. >> yep. For sure.
23:27
Yeah, there's a there's a crazy bull case for Nvidia in the information arguing it should be worth what, 22 trillion dollars? That is a wild move.
23:37
There's a there's a lot going on.
23:38
The scaling laws holding is the most important >> from The Information Finance.
23:41
Nvidia worth 22 trillion.
23:43
This old school financial model says, "Yes."
23:50
Uh Uh so, um Uh yeah, the big news on uh Uh yesterday was um uh Anthropic's new model Mythos.
24:00
Some really impressive statistics and anecdotes yesterday, both the model card, the benchmarks, and some uh stories about breaking out of of a variety of uh what what what what what what what do they call them?
24:15
Walled Gardens or test environments?
24:17
Uh what what what what what what what are those Uh breaking out of the I don't know, the simulation the sandbox.
24:22
Yeah, breaking out of the sandbox, sending emails, all sorts of stuff like that.
24:26
Uh the model preview is only available right now to about 50 companies that maintain critical infrastructure because the model is particularly good at finding zero days, bugs, and uh exploits in uh technical systems.
24:39
And if they you know, they lead that they leak that out before big companies have a have time to go and address all the bugs, there could be serious uh you know, serious ramifications for cybersecurity.
24:52
And so, key partners include Apple, Google, Microsoft, Amazon, Nvidia, JP Morgan Chase, Broadcom, the Linux Foundation, Cisco, CrowdStrike, and Palo Alto Networks.
25:02
They're all listed on the cybersecurity focused page for Project Glass Wing.
25:07
Chris Backe was having a little bit of fun cuz he noticed Anthropic put their own logo on the partner page, which is a little bit funny, but at the same time it's kind of smart cuz a lot of people are just going to see the image quickly and it's good to it's good [laughter] to position yourself with the other companies. Yeah.
25:23
So, yeah, it's it is interesting.
25:23
I mean, people have predicted that AI models would be particularly good at cyber attacks and this was one of the main sort of vectors of AI fears.
25:30
It feels like uh this is what maybe what Dario was referring to when he was talking about the the end of the exponential.
25:39
Finding and exploiting software bugs is it's sort of perfectly in the sweet spot for coding agents and reinforcement learning.
25:48
Combing through piles of code, tirelessly trying different exploit exploits to find bugs, having a clear verifiable reward.
25:55
Did you crash the system or not?
25:57
Did you break into the system or not?
25:59
This is very it's it's it's a very clear binary signal that you can send to the model to determine were were you successful in breaking into that system?
26:10
And it requires basically no time delay. There's no lag.
26:15
So, there was one snarky tweet I saw that was something to the effect of like, "Okay, then if it's so good, go cure cancer."
26:21
But any application that requires a real-world feedback cycle, even if it's just a few minutes of human interaction in the in the cancer example, you know, you're going to need to be testing the drugs in vitro, in mice, in monkeys, in humans at some point.
26:37
point. Or even if you're just sequencing DNA or doing anything in the lab, pipetting anything, if it's even just a few minutes, all of a sudden every iteration, every attempt is going to take a few minutes and that's going to put you on just a wildly
26:50
different exponential as opposed to being able to spin up a virtual machine with basically every single piece of software out there and then try every single exploit against every single piece of software and you wind up with a ton of exploits. And very very bullish for cybersecurity
27:05
And very very bullish for cybersecurity that this is being done preemptively.
27:11
There's a whole bunch of different discussions.
27:12
Ben Thompson has a good piece on on the whole decision to uh to release the model or not and stage it out and and and the go-to-market there.
27:22
But it's uh um it's even even if the the the the bio research, the other impacts are on sort of a slower exponential, there's still so much opportunity in even a software-only singularity.
27:38
There's also risk in a software-only singularity.
27:39
We've seen this story before though.
27:41
A model that's too powerful to release, but then works its way out and has pretty moderate impact on the world.
27:47
This was the story of GPT-3 to the story of ChatGPT.
27:52
The question of, you know, is this [snorts] the is this the model that's dangerous to put in the hands of people?
27:57
Yeah, headline headline from February 22, 2019 by Aaron Mack.
28:00
OpenAI says its text generating algorithm GPT-2 is too dangerous to release. >> Yeah.
28:10
So, there is there is a I think Ben Thompson called it like the the boy who cried wolf syndrome. But the Mythos wolf.
28:17
He says, "There's a lot of skepticism about Anthropic's announcement."
28:21
This tweet was representative from Buko Capital bloke.
28:24
Anthropic's marketing strategy is so funny like, "Ah, the government is treading on me.
28:28
Ah, our models are so good, we can't release them.
28:30
It would be too dangerous. Ah, someone stop me.
28:31
I'm going to destroy the economy."
28:34
The rolling of the eyes is exacerbated by the fact that Anthropic has reasons to make to not make Mythos widely available beyond a lack of compute.
28:43
Another factor is surely trying to avoid having Mythos distilled by Chinese model makers.
28:48
So, there's actually two good reasons to to gate um to gate access.
28:54
And when you're looking at those logos, when you're looking at the world's largest tech companies, there's much more ability to to scale roll out demand, set pricing.
29:07
These companies might be able to pay more.
29:09
The model is very expensive, but if you're justifying that against bug bounties for zero-day exploits in your most critical system, when you look at like JP Morgan Chase, it's a bank.
29:20
What is the price of finding an exploit in that system? It's pretty high.
29:22
It's probably clears the token hurdle a lot.
29:26
And and and if the roll out is is paced like evenly across all the different companies, they'll all sort of understand that they're getting allocation inference allocation at the efficient price that that clears the cost actually serve the model.
29:40
So, I do think that the systems uh all of these 10 trillion parameter models will be released soon broadly.
29:47
Uh And the main reason that an AI that's smart enough to find zero-day exploits should be able to recognize that it's being used by a bad actor to find zero-day exploits.
30:00
And uh it's only been a it's only been a few months since the last flurry of competing models from OpenAI, Anthropic, and Google.
30:06
And the next cycle is already off to an aggressive start.
30:08
We had Meta and then the other news is that Elon Musk announced that he is getting ready to do another larger model with xAI.
30:22
He's got He's doing seven models in training. Wow. That is a lot.
30:25
Imagine V2, two variants of 1 trillion, two two variants of 1.
30:28
5 trillion, a 6 trillion model, and a 10 trillion model.
30:33
He says there's some catching up to do, but he he says he will never give up, never.
30:39
So, he is continuing to to grind and and train more models.
30:45
What what else was in the reaction?
30:47
There was a whole bunch of other back and forth.
30:48
People seem split on Mike from Also Capital, former guest, says we've decided not to release our latest investment strategy.
30:55
It's so powerful, releasing it might end the entire venture asset class as we know it. Yeah.
31:03
And >> Backy says you should release it to a handful of trusted partners so that we can harden ourselves to it.
31:09
And George George Hotz says Anthropic's marketing strategy is amazing, it's so powerful, it's terrifying, and the best part is you can't come.
31:18
By the way, if Anthropic had any way to ship this, they would.
31:20
Trained AI models are the fastest depreciating asset in history.
31:24
GPT-4 cost $100 million to train 2 years ago and is now worth worth less than Qwen 3. 5B Qwen 3. 5 27B 1 million.
31:35
Uh sending the FOMO back, clock is ticking, boys.
31:38
>> [laughter] >> He's It needs something like an NBL 72 to run a decent speed, and even absurd API pricing doesn't cover it.
31:44
There's more to be made on investor hype than API access.
31:48
I just wish for honesty instead of a whole fake spiel about safety.
31:50
Who remembers when GPT-2 1. 5B was too dangerous?
31:54
And so, lots of back and forth.
31:57
Dean Ball has some more thoughts on Mythos.
31:59
It's a longer post, so we'll let you go and read it, but the main the main take is just the you know, this is this is technology that whether it comes from Anthropic or another lab, like clearly needs to go into the supply chain of the world and in the US government and the US economy because no one is doubting, even though some of the exploits were somewhat minor, no one no one disagrees that we need less cybersecurity.
32:26
We want the most secure systems possible, and we probably want a lot of competition between different companies to provide that service to the government.
32:34
government. And so, hopefully, with if the war comes to an end and there's you know, different discussions can happen and you know, ice can thaw and there's a way to for these companies to work together, even if even if the
32:49
even if the supply chain thing doesn't go through and then Anthropic can vend technology through Project Glasswing, through CrowdStrike, through Oracle and other part partners um to to to Cisco so that at least the systems are secure because everyone wants that. So,
33:07
So, Dean Ball's been on an absolute tear.
33:08
We should have him back on the show and talk more.
33:10
He says a lot of people, including people in positions of authority, told us recently that models of this capability wouldn't be a thing, that models with obvious national security implications would not be forthcoming.
33:20
Those people were wrong, there's nothing to do about it, but you should remember it.
33:23
Mythos is the first model where theft of the weights by an adversarial actor feels like it would be a major deal.
33:29
You better believe they will try, and if they don't succeed with Mythos, they will eventually.
33:33
We are thoroughly in the era of the labs' best models may may well not be in public the way they used to.
33:39
This is because of a combination of compute constraints, economic reality, competitive advantage, and safety concerns.
33:44
Three means the most relevant models may be decreasingly legible to the general public, and depending on the extent and duration of the coming compute squeeze, we could enter a market dynamic where the best models are only available to the highest bidder.
33:57
And of course, that makes sense from a KYC and security perspective.
33:58
In other words, where compute is a seller's market rather than a buyer's market. Interesting.
34:05
Imagine competing firms in the economy bidding against one another for access to the best and most tokens, and the frontier labs as in essence kingmakers.
34:13
The governance regime I have described above in four is not designed to stop that dynamic.
34:18
And so, there is there's plenty of more takes.
34:22
This was a full current thing cycle.
34:26
Um People Tenebris says people keep talking about this like it's not blatantly obvious.
34:30
Anthropic clearly has a system that's auditing open-source repos for vulnerabilities using their unreleased higher power models and sending fixes for them without revealing their current level of capability.
34:39
So, they've been going around on GitHub and contributing pull requests to to to to you know, patch any vulnerabilities without disclosing exactly what model was being used. Um Burn Hobart Yes.
34:55
is not excited about Fundrise.
35:01
We had the founder on running ads for VCX, the public ticker for private tech.
35:06
This is he says paying for an ad encouraging people to pay 6X net asset value for a closed-end fund where the cost of the bars is 400% is one of the things people will will remember during the next bear market.
35:20
Yeah, I asked I asked the the founder Fundrise about this, like how you kind of like is there another iteration of the of the product that can solve for this?
35:31
Fundrise like very clearly made a bunch of really good bets Yeah. a few years ago. >> Yeah.
35:35
And the fund has performed incredibly well, but the issue now is like if you want access to these names, and the only way that you have is to go through VCX, you're paying 6X what like the actual private market investors are paying.
35:52
And that's just is like I mean, it's I don't know.
35:55
It depends how bullish you are on the names, but it's going to be very very hard assuming that normalizes over time. Yeah.
36:03
It seems extremely unlikely that it trades at you know, an insane premium forever.
36:07
And so >> Interesting that they're running this ad on X.
36:10
I haven't seen I I have X premium, so I have I don't see a lot of ads.
36:13
I don't see any ads, but that does seem like the reasonable place to go to advertise a product like this, but yeah, it is it is always odd.
36:22
There's been a whole bunch of these like treasury companies that have traded above net asset value, and it was always just a weird supply and demand dynamic.
36:34
People want them, and they're willing to pay way above.
36:35
Hopefully, they know the the net asset value multiple, and they're doing that willingly.
36:41
I think you know, consumer education investor education is more of the critical the critical question here.
36:50
Um Well, Tibo over at Codex is unreasonably excited about things.
36:53
The next few weeks will be intense and fun.
36:57
And yeah, Michael Greenish says weeks, years.
37:02
You know, it's going to be an an ongoing model mayhem for Vag maxing.
37:06
Vag maxing and Yesterday a lot of stuff going on.
37:11
was about token maxing, today is about vag maxing. Yes, yes. Let's go.
37:15
Uh Mickey Friedman says the current fear is that AI homogenizes culture and turns humans into passive consumers.
37:19
One counterpoint, in Go, human play showed very little improvement from 1950 to 2016 until AlphaGo beat Lee Sedol.
37:25
Then human decision quality jumped.
37:29
Players started developing moves that were distinct from both previous human moves and from the novel moves introduced by machine intelligence.
37:36
This seems more likely to me. Fun times ahead.
37:41
Lee Sedol is now a professor at UNIST.
37:46
He is a special professor on a 3-year term to conduct artificial artificial intelligence research on Go specifically.
37:55
Yeah, if you haven't seen the Go AlphaGo documentary, it is fantastic, extremely well >> go to smoke. Yes, Lee Sedol.
38:01
It is it is such a wild ride watching the DeepMind engineers like you know, it seems like they're genuinely surprised by the performance.
38:12
Like no one really expected it.
38:16
But yes, this is very interesting chart to see how much things changed in the post-AI era as people discovered new and interesting ways to differentiate from from the models effectively. Yeah.
38:29
Scoop from Stephen Nelson, CIA uses secret tool called Ghost Whisper to find airmen in Iran. >> Yeah.
38:40
Ghost Whisper pairs long-range quantum magnet magneto magnetometry. Yeah. How do you say that?
38:46
Mag sensors with AI to find human heartbeats.
38:50
I was wondering this while they were over the weekend, there was you know, a search going on.
38:55
It was like how do you how do you find someone how how does how does somebody like you know, an airman that's down send a signal that can be picked up by one group, but not This is very odd.
39:07
So, there are some there are some community notes on this saying that quantum magnetometry, I'm probably >> Magnetometry.
39:16
Imagine that's how you pronounce it.
39:18
Detects heart magnetic fields, and I believe this technology works in labs, but only up to a few meters, not 40 miles as claims as claimed.
39:27
Fields decay with 1 over R cubed, making long long-range detection implausible.
39:33
So, unclear if this is what worked, but isn't there isn't there has to be some sort of device that you could carry on your person like in your shoe like an air tag that can talk to a satellite almost like you look at the Starlink receiver dish it would fit in a backpack but that's very high bandwidth.
39:53
but that's very high bandwidth. I imagine if you had something I mean there's sat phones that are the size of large cell phones that was available in the 80s and 90s you have to imagine that if you're just trying to put out a
40:07
signal to GPS or a Starlink network you must be you must be able to shrink that down significantly to the place where it could be carried on your body but it's probably classified so I would be surprised if if it it just very hard to read into like what's real and what's what's not here. There isn't there isn't different
40:25
There isn't there isn't different community note pushing back saying no note needed this new technology is classified system developed in secret by Lockheed Skunk Works and the CIA that was just used revealed publicly for the first time naturally it's reported
40:38
capabilities far exceed the known public state of the art the note is relevant so it's it's it's it's very very interesting but good to see >> let's go over to Aaron Tan's post he says introducing Loom a lamp that does your chores order now shipping this summer let's see the video. That bed already looks fully made what
41:13
That bed already looks fully made what what chore is it going to do?
41:15
Just dropped some laundry off.
41:16
Oh okay you have to drop the laundry first. Wait it can do that?
41:20
Wait it doesn't have fingers in that like what what is [music] it put on the record?
41:27
What yeah what is what is in the >> You can see it has a little claw. Oh it has a claw inside? Oh okay.
41:32
>> [music] >> That is so funny to have a humanoid robot play music on a physical record.
41:47
The folding t-shirt is the Turing test [music] of humanoids for real.
41:52
Pixar lamp quaking in its boots right now. Nice video.
42:10
Good color grade nice warm tones friendly doesn't feel dystopian feels delightful.
42:15
Would you get one of these?
42:18
I think I think we should get one.
42:18
Yeah [laughter] yeah like I not my house.
42:23
>> I think we should get one.
42:23
I do >> a team obviously you know rooting rooting rooting for Aaron Yeah.
42:29
rooting for Aaron Yeah. and the and the Loom team very very unique form factor I mean I just think the an impressive timeline for shipping hopefully yeah shipping I'm reading this as like actually shipping not shipping like another site to order because you can order already but yeah it'll be very interesting if it can if it can reliably
42:55
fold clothing it could be it could be enough right so the the the the benefit here is like people already want lamps I'm assuming for their their bedroom if you can buy a lamp that's reasonably priced and then it could also has a benefit of just a simple thing like folding clothes there could be there could be a market here. Yeah I feel like I don't know even
43:11
here. Yeah I feel like I don't know even just putting pillows back on the bed even just like really basic things I mean even there are probably applications like the Roomba did so well with such a minimal such a minimal scope there's there must be something I would I wouldn't be surprised if even
43:31
even in between this and fully folding the clothes just remaking the bed properly feels like something that consumers might actually pay for and and allow for you know the flywheel to start spinning obviously lots of security considerations since there's a camera
43:49
there and whatnot they'll have to do a lot of cybersecurity and figure out that but people already have cameras all over their homes from nanny and child monitors baby monitors and that type of stuff so I'm I'm optimistic about this and I think they did a great job promoting it
44:03
there is a story in the New York Times from Nanda and John Carreyrou who who blew the blew the story wide open on Theranos he says the mystery of Satoshi Nakamoto the the pseudonymous pseudonymous inventor of Bitcoin has remained unsolved for 17 years not anymore read my 18 month investigation to find out who Satoshi
44:26
really is and he says it's Adam Back who says who posted directly I am not Satoshi but I was early in laser focus on the positive societal implications of cryptography online privacy and electronic cash hence my 1992 onwards active interest in applied research on e-cash privacy tech on Cypherpunk list
44:48
which led to hash cash and other ideas um John Carreyrou in his New York Times research finds like Aaron Van in his Genesis block book uh many interesting Bitcoin analogs in the early attempts to create decentralized e-cash in effect prototype ideas trying to figure out a
45:06
Bitcoin like thing including P2P BGP and proof of work first quote I'm not saying I'm good with the words I'm not saying I'm good with words but I sure did a lot of yakking on these lists actually the broader context was my observation that because I was talkative on the list and
45:23
known to have an active interest in e-cash there is some confirmation bias in finding my comments frequently on e-cash topics so he has he has said we are we are all Satoshi I am not Satoshi but it's a it's a very interesting story that will have You just created a million Satoshis. Truthfully
45:37
Truthfully how yield Harry's joking Steve Buscemi has been revealed as the Bitcoin founder not Satoshi Nakamoto.
45:44
Well we have our first guest in the waiting room Luther Lowe from Y Combinator he's the head of public policy and we are going to be talking to him about vibe coding Luther how are you doing?
45:56
Hey guys great to see you congrats on the acquisition. Thank you.
45:57
Thank you great to have you on.
46:00
Can you kick us off with a little bit of an introduction on yourself and what you do day-to-day at Y Combinator?
46:06
Sure so yeah I'm the head of public policy for Y Combinator I'm based in Washington and Gary created the role when he started at YC and really his observation many years ago was we Gary and I had actually met in
46:23
Washington and we were sitting around the table at some kind of White House meeting and he said you know you you go to Brussels you go to Washington and the largest tech companies have lots of representation but little tech doesn't have a seat at
46:37
the table and that was actually the first time I'd heard that phrase little tech and Gary actually kind of coined that phrase you went back and you look on X Twitter he has had said that a number of times years ago and so
46:49
you know my role is to really like help the founders navigate Washington Brussels the state capitals advocate for pro little tech issues in the broader ecosystem and yeah just really do everything I can to help the founders. And what's the what what
47:04
the founders. And what's the what what is what is your view on on vibe coding the app store the boom I was we were talking about it yesterday there's you know there's a whole bunch of stories of where it feels like we're getting close to the one person absolutely massive company whether that's GMV or revenue or
47:25
something it's like it's starting to happen but I was looking at my home screen I don't have an app that is new or vibe coded maybe it'll come in a in a boom in video games but how have you been tracking I mean I'm sure you see this at YC just the the the growth of broad app development? Yeah I mean I think
47:45
Yeah I mean I think I would almost kind of take a step back it sort of reminds me of when I first started geeking out on the internet like 25 years ago where you saw the the rise of sort of what you see is what you get HTML based or browser based HTML editors and that allowed anybody that kind of
48:06
democratized the process for so anybody could create a web page or a website and now we have tools that allow anybody to create a web service or an app and so the difference though between now and 20 years ago is that today we have basically these these two bottlenecks in the form of Apple and Google that sit
48:29
between the creation and the potential users of those services the Apple App Store is basically like the worst DMV in the world and so we're seeing not only sort of and see you know there's reports all over X about this but if you um just barely kind of look around for it you're going to encounter lots of folks
48:51
that are trying to develop apps and services that are not being accepted or getting kicked out and then it's not only that kind of like app layers sort of the layer up of the tools like Replit and anywhere that are facing you know the inability to update their apps and so it's a real problem. Yeah what
49:09
so it's a real problem. Yeah what when you say the worst DMV in the world are you actually referring to we saw a chart where the number of app store submissions is is spiking it's going exponential and that feels very logical
49:24
because people I mean we've been building vibe coded web apps here the next step was hey maybe we should actually get one of these in the app store but very quickly we realized okay well it's at least a two week review process it could be a lot of back and forth. It's a new hurdle. We need to, It's a new hurdle.
49:38
We need to, you know, actually compile it to to Swift or or Objective-C.
49:42
It's a whole different process.
49:45
Probably doable on a technical side, but we might be hung up there.
49:49
But are you seeing a issue with actually getting just a one-off vibe coded app approved?
49:57
And then I think we should talk about the the apps that help you vibe code new apps because that's a whole separate thing.
50:03
But just in terms of like I want, you know, a a special recipe app and I vibe coded and I want to get it on my phone and I want to send a link to the App Store page to a friend. Is that slowing down?
50:14
What's slowing that down?
50:17
Is that going to be a permanent thing or do you think Apple can just adjust there?
50:23
Well, I think the problem with Apple, I mean, it's this is a sort of perennial issue with Apple is their culture is one of just absolute control.
50:30
And I think that we have reached this inflection point where, you know, it if I've got my MacBook in my lap, I can open it up, I can download any app I want. I can open terminal.
50:43
I can do all kinds of crazy mods.
50:45
But the second that that form factor fits in my pocket, all of that freedom goes away.
50:49
And then and I'm living in sort of North Korea in terms of what I can do with my with my stuff, with my property.
50:55
And so I think the, you know, this is, you know, sure, I could I could launch something in TestFlight if I want some kind of little bespoke, you know, training app or something.
51:05
But God forbid if I want to share it with friends or I want to, you you know, make some money because I've created sort of a differentiated product that's actually interesting that people want.
51:15
I've got to pay this ridiculous vig to Apple.
51:17
So for them it's actually about the reason they want control is because it's it's about App Store revenue.
51:27
And also they have competing products.
51:29
It's an anti-competitive thing because they've got Xcode and they've got their own dev tools that they're starting to kind of kind of roll out their own sort of vibe coding services.
51:38
So the longer that they can kind of delay and slow roll both the developers and the the tools that allow the vibe coders to create >> So your your theory is that they'll they want to basically make their own version of Replit or anything that they have in, you know, total control over.
51:59
And they can make the argument, they could probably make the argument that this is better for users, it's more secure, better for privacy, less like malware risk, whatever whatever it is.
52:10
But they um Yeah, anyways.
52:13
I mean, they they give you there's already like they give you like pretty meaningful autonomy over like shortcuts and other things like that.
52:20
It's not that unbelievable that they would want to actually make something here.
52:24
But but it's like embarrassing.
52:26
Like I mean, Siri is embarrassingly stupid.
52:27
We've had sort of LLM consumer-facing services like, you know, Sam brought them down from the mountain, you know, whatever we almost 3 years ago.
52:38
And I, you know, like Siri still can't do like basic subtraction.
52:41
The other day I asked it for like what was like you know, 17 days from, you know, this particular date.
52:49
And it was like, do you want Google to consummate that query?
52:51
Do you want ChatGPT to do it?
52:53
It's like, why can't Siri do it?
52:55
And and so you've got all these this just lots of energy there's lots of YC companies that are trying to take a crack at creating kind of Siri alternatives.
53:06
But until kind of this self-preferencing is addressed, I think at the sort of policy-making level, and it's not just Apple self-preferencing, it's Google, it's all of them.
53:18
Then we're not going to really be able to see the the fruits of this sort of LLM revolution diffused into the hands of consumers.
53:24
Do you think the the the side button to trigger the helpful assistant that is currently mapped to Siri and can interface with ChatGPT and is, you know, reportedly going to interface with Gemini soon.
53:40
Do you think that there's any actual chance that that becomes remappable at like a, you know, pick your browser search engine level in the in the OS, even if it's a even if it's defaulting to Apple's stack?
53:54
Do you think there's any motion there because that feels like the logical analogy and what as an Apple consumer I would like.
54:05
I'd like to be able to pick my assistant, but be able to assign the hardware button.
54:08
But where do you think that actually goes?
54:13
Look, I there is no technical reason why we shouldn't have a flourishing ecosystem of third-party AI assistant developers competing to do all kinds of interesting stuff.
54:24
You guys should check out a YC company called Blue. It's heyblue. com.
54:27
They've this amazing demo on their website and they talk about and basically the guy's sitting in his car Yeah.
54:35
and he's driving he's driving up the 101 and he's saying, "Hey, go through my Slack messages from last night. Okay, cool. Jerry needs a document.
54:43
Share it get into Google Docs and share that with Jerry, but make it read-only."
54:47
I mean, look, I like my third button having Perplexity and being able to queue that, but I can't if I can't access those deeper OS API commands, then it it stops being that interesting pretty quickly.
55:01
Anyway, the the Blue guys the what you learn you're like, "God, this is magic.
55:06
How did they how did they do this?
55:06
How did they figure this out?"
55:07
You realize it's a USB-C dongle that they've basically 70% of the company is is hardware.
55:13
They're they're doing soldering irons.
55:15
They're they're flying to China negotiating with the distributor.
55:17
It's like, why is this a hardware company?
55:19
This should be there should be, you know, there should be 50 different companies like this duking it out to to create something that's better than Siri cuz Siri clearly is not uh cutting it.
55:29
And I think the antidote to this, frankly, is, you know, we just announced yesterday a coalition of 275 startups and VCs in support of the BASTA Act.
55:42
>> [laughter] >> It's the Banning Anti-Competitive Self-Preferencing by Entrenched Dominant Platforms Act bill in 1074 by Scott Wiener, not to be confused with Scott Wiener's ill-fated bill 1047 from a couple years ago that had to do with AI regulation.
55:58
This one we're really excited about because basically it it calls upon companies that are a trillion in market cap or above 100 million US users to end this type of egregious forms of self-preferencing.
56:12
We're totally fine with innocuous forms of vertical integration.
56:15
In fact, most of the time that's like a great thing in technology products.
56:20
But you know, when you've got, you know, the GitHub COO last week said that commit rates, if they stay linear, are on track to be 14x what they were last year. Wow.
56:32
Like, you know, something's got to give here. >> be more software.
56:36
Everyone's building software.
56:37
There's going to be a ton more software.
56:39
How do you think the the vibe coding apps apps that go in the App Store and allow you to build more apps with that particular app.
56:47
How do you think that will shake out because that is a place where it feels like Apple has spent a lot more time making the case.
56:53
When I think about the Siri button, I it I feel like they haven't made a strong case for why that can't map to a different app that's already gone through approval.
57:02
But when I think about the consumer protections and privacy that comes with knowing that the software that you download from the App Store has been audited, they've made a pretty strong case there.
57:13
What's the counterargument for why Apple should open the floodgates of apps that allow anyone to vibe code an app and and deliver it to anyone else's device sort of willy-nilly?
57:27
Look, I don't think that anything is going to in terms of creating consumer outcomes that maximize privacy, competition, all the great things that you want.
57:36
I don't think any mechanism is going to work better than good old competition.
57:39
And so that means allowing for side-loading.
57:42
It means allowing for alternative app stores.
57:43
You know, you know, we make fun of a lot of European regulation, but the Digital Markets Act has actually done a pretty solid job at that.
57:51
And so like in places like Japan and Europe, if you don't if if you are just hitting a wall with the worst DMV in the world, you can just go to the the alternative app stores and I as a consumer can decide to download one of those and use that as a way to access a whole other ecosystem of services.
58:07
And they have their own vetting processes.
58:12
And I think again, competition is going to be the mechanism that actually enables better privacy, better consumer protection.
58:20
And I think what we found historically with with Apple, and we saw this a lot a few years ago with Beeper, which was a YC company that had interoperable messaging.
58:29
It solved the blue bubble Exactly.
58:30
Eric Migicovsky solved the blue bubble green bubble thing, basically made it where it's not awkward for you to buy an Android and like, you know, Kool-Aid Man into your, you know, college group chat and turn turn the whole thing green. He fixed that.
58:45
And of course, you know, what does Apple do?
58:47
They said, "No, you we're, you know, Android users don't get that.
58:50
We've got to basically reduce security to the lowest common denominator like, you know, SMS, RCS standard."
58:55
So there's no it's pretextual.
58:58
Like we we we can have choices. We can have nice things.
59:01
And little text trying to build it.
59:04
But I think policy makers have got to do a better job at, you know, ensuring that the the biggest players are not kind of egregiously putting their thumb on the scale. Yeah. Last question.
59:13
What else are you tracking in DC?
59:15
It feels like with the AI boom, there's a lot of new company formation.
59:21
I'm sure you're seeing that on the YC side, but we see it every day on the show.
59:27
Are there are there policy initiatives to encourage entrepreneurship that you're tracking?
59:31
Is there anything else on like the small business side that you find interesting these days?
59:36
Yeah, I think you know, there's I I would say where I I think the Trump administration is doing a good job is you know, they've they've done a lot to sort of advocate for the American AI stack and sort of encourage entrepreneurs to like interface with government and become and they they've you know, put out sort of bids that for the where the government is a is a a buyer of these tools and they've been very proactive and have great relationship.
1:00:08
You see like lots of renewed interest in defense tech and dual use tech.
1:00:15
I think where I wish that there was more work and I think we they could probably be doing a better job is is thinking about how do we make sure that the United States is brain draining the world and bringing in the top talent and making sure that the smartest people in the world are building companies here.
1:00:33
I think that that um that's where if I could change anything, I would say they could do a little bit better.
1:00:40
But I would say you know, they have been they've been very thoughtful.
1:00:44
The president's talked about little techs in his in his tweets like and and I think you know, [clears throat] this is not a partisan issue.
1:00:52
We want to have a thousand flowers flourish.
1:00:57
But it is yeah, I mean I think getting the the tech the tech talent piece with the sort of being bullish on AI that that I think is going to get and and also the competition piece getting all the those things right.
1:01:10
I think that's going to put us in a position to be really amazing here. Yeah, I love it.
1:01:14
Well, thank you so much for coming on the show. Yeah, great to meet you.
1:01:18
Thanks for having me guys.
1:01:19
Appreciate the hot takes. All right, you too. Cheers.
1:01:23
And before we bring in Dan Primack from Axios, we have an exclusive clip from his interview with Kalshi CEO Tarek Mansour on the Axios show which we will be discussing with Dan after we play it here.
1:01:37
It's about a minute and a half. So let's listen to this.
1:01:40
>> for customers to lose.
1:01:40
Actually, the proof of that is that you know, when a customer wins on a traditional sports book, they block that customer. Mhm.
1:01:51
And the CFTC chair who supports prediction markets being legal, supports federal framework, etc.
1:01:54
said recently that one of the when he was asked about this, he said, well, part of the difference is that when you walk into a casino, there's all sorts of entertainment, right?
1:02:02
There's like we shows, there's food, there's drink.
1:02:03
So casino betting in a casino is different than than betting in Kalshi.
1:02:07
But it's not necessarily different than betting on a sports book app on my couch.
1:02:11
There's no entertainment difference.
1:02:13
There's big fundamental differences.
1:02:13
It goes back to the market mechanic.
1:02:14
Like one of them is essentially a product that is designed for customers to lose.
1:02:20
Actually, the proof of that is that you know, when a customer wins on a traditional sports book, they block that customer because those winnings are coming from the business model, the business itself. >> win enough.
1:02:29
If they win enough or if they win really if they win in in a consistently, if they do research, if they get informed. If they're good. Exactly.
1:02:35
And if they're not, actually get them promos and you figure out how to bring them back and there's kind of a bit of this sort of like I think we call it kind of marketing tactics to bring these people back.
1:02:44
It could be like entertainment or promos and so on and so forth.
1:02:48
That does not exist in prediction markets.
1:02:49
It is a fundamental difference in structure where actually a lot of people that win, the people that are doing research, are getting informed, that are traders do come to prediction markets.
1:02:58
This is where the value prop is strongest because prediction markets do reward them for being right.
1:03:05
Well, we have Dan Primack here with us today. Dan, how are you doing? Doing well, guys.
1:03:11
Not as well probably as you are, but I'm doing well. Thanks for being here. Nice to see you.
1:03:17
So give us an update on this interview, what you were looking to learn from it, what you think the conversation, where you think the conversation around prediction markets will go from here.
1:03:29
I think it was interesting.
1:03:29
The day we did this we did this on Monday in New York City and and it was really maybe an hour or two after New Jersey a judge in New Jersey had basically an appeals court judge.
1:03:38
So kind of the highest judge so far that's dealt with this basically gave Kalshi the green light to go forward.
1:03:43
I mean I think ultimately prediction markets are going to go to the Supreme Court.
1:03:46
I think both Polymarket and Kalshi thinks that.
1:03:47
I also think they're probably going to win in the Supreme Court.
1:03:51
The law really is on Kalshi's side.
1:03:52
Um they have done a pretty good job kind of being in step with regulators and granted they were it was a little more complicated with the Biden administration, but they sued the CFTC. They won that lawsuit.
1:04:04
It's the reason why they have elections on the platform.
1:04:08
I think if something is going to stop or change the way prediction markets work, that's going to have to come out of Congress.
1:04:13
So yeah, what's the next step because it feels like there's >> benefit from having their their adversary are casinos and sports books which are not exactly the kind of groups that Americans want to stand up to defend, you know, and march it's time to hit the streets and march for the legacy sports books.
1:04:33
>> I mean that's some of it, right?
1:04:33
So obviously in Nevada which is one of the few places where Kalshi is actually banned because the judges upheld an injunction.
1:04:39
Yeah, that's the casinos.
1:04:41
That's maybe some regulatory capture.
1:04:42
The other folks who are against this though are folks who are just anti-gambling in general and I and I do think you are seeing a bit of a groundswell of that politically and I think it's bipartisan.
1:04:51
You know, some of the bills that have actually come out have been coming from Democrats.
1:04:54
But for example, the state of Utah doesn't want this and they don't want it to protect casinos.
1:04:59
They don't want it because they don't want people betting period and they sure don't want it on phones where it's so easy to do and and we did in this interview on the Axios show which drops tomorrow.
1:05:09
We did talk about the addiction piece of this and kind of the the broader societal issues which isn't to say what Kalshi is doing illegal.
1:05:17
There are questions about is it promoting not immorality, but is it enabling addiction?
1:05:21
And what are the proposals for because this is not regulated as gambling and it seems like it will continue on that path.
1:05:29
Is there any motion to bring some of the restrictions that apply to traditional gambling over into this new regime?
1:05:40
No, no and I mean that that's the the complicated thing here and I did ask them you know, is this basically a loophole?
1:05:45
For example, in California, in Texas, the two biggest states in the country in terms of people, right?
1:05:52
You can't you know, try to open DraftKings or FanDuel, it doesn't work for you because because it's not allowed there.
1:05:57
Kalshi though will allow you to bet and you know, I think Tarek or Luwana said about 70% of their volume last month in March was sports betting which is lower than it had been in February, but you're still talking about 13 billion dollars of volume last month. So it's a big number.
1:06:13
You know, it's a loophole.
1:06:13
They have figured out a way to basically get around sports betting laws and and I appreciate that the back end is different.
1:06:20
They make a very compelling argument for why legally they're allowed to do it and I agree with them.
1:06:24
But the reality in the end is if I'm betting on the Celtics-Knicks game tonight, I don't really care what the back end looks like.
1:06:33
I care about the you know, my my money if I win and the fact that I'm able to do it.
1:06:39
So is is federal preemption like sort of baked in at this point or is there some sort of hybrid rule set where this could go back to the states and states could make their own rules because it does feel like there's a pretty wide set of opinions state by state on how different communities want to engage with this particular product?
1:07:02
Right now, it appears federal preemption is going to win the day.
1:07:04
For starters, the Trump administration specifically the CFTC which is what regulates this.
1:07:09
This isn't regulated by the SEC, it's the CFTC.
1:07:13
They are all in on prediction markets.
1:07:14
Mike Selig who is the current commissioner, he he is on the side of Kalshi.
1:07:18
I assume will be on the side of Polymarket when they eventually really come to the US.
1:07:23
And and and it does make a certain amount of sense and there is some historical precedents here.
1:07:26
A lot of commodities which are traded and that's kind of what they're arguing that these are kind of commodities in in a different name.
1:07:33
There were lawsuits a hundred years ago arguing the same thing.
1:07:37
This is crazy speculation.
1:07:37
The reason why I say this could go to Congress, there are two carve outs to that. One is a weird one. It is onion futures. Onions.
1:07:47
Not pork bellies or or something else.
1:07:49
Onions because at one point a long time ago, there was a ridiculous speculative like a lot of people lost a lot of money on onions.
1:07:55
So Congress passed laws saying you can't trade onion futures and I think I'm right in saying the other one is related to box office returns which is why if you're on Kalshi, you won't you can't bet that a movie's going to make 20 million bucks, but you can bet on its Rotten Tomatoes.
1:08:09
That's so funny because I've never been like a sports better at all, but I did I did participate in a fantasy movie league for a while that had no financial incentive whatsoever, but you would construct a hypothetical movie theater, pick okay, Project Hail Mary is going in the first slot and then they would have the box office returns, but it was all just for for fun with a bunch of friends.
1:08:29
>> that maybe it still exists, but that back in the early you know, 2000s, there was something called the Hollywood Stock Exchange which again wasn't for real money, but people did that.
1:08:35
They they it looked like a stock market. >> Yeah, yeah.
1:08:38
Do you so so other other countries have created rules like you know, you can't advertise gambling during you know, these hours and there's a bunch of different kind of rule sets around it even in places where gambling is legal.
1:08:54
Do you expect any states to pass laws that say you can't advertise commodities trading platforms Yeah.
1:09:00
during More like FCC as opposed to CFTC.
1:09:06
>> basically like not targeting like sports trading which you know, the Kalshis and the Polymarkets are doing, but effectively saying like, hey, we know these are going to be the biggest spenders and we don't want to tolerate or encourage this type of activity in the state.
1:09:20
>> you may see that and then that gets fought out in court.
1:09:22
You know, I obviously again, you you on the federal side, you know, it's interesting.
1:09:25
It's not just Kalshi fighting back against these lawsuits.
1:09:28
The CFTC itself, uh there's three states, Arizona, I'm going to mess up the other two, I think Connecticut is one.
1:09:34
There were three states who have tried to ban Kalshi, and the CFTC itself has come in to sue.
1:09:37
So, I could see the FCC coming in to try to sue if such laws were passed. It would be interesting.
1:09:43
I mean, I do think part of well, part of a lot of betting, despite what we saw several years ago with DraftKings and FanDuel, you know, on every billboard and every advertisement, an enormous amount of this is still word of mouth.
1:09:53
Uh you know, people talking about it.
1:09:54
People I I will tell you, when we were doing this interview, uh most of the crew, you know, the camera folks and the and the sound folks, hadn't heard of Kalshi before we did this.
1:10:02
And when it was over, I I was in the corner kind of packing some stuff up, and I heard them at the table having lunch.
1:10:09
They were all not making bets, but they were flipping through the app, and they were talking about different bets that were on there, and they were fascinated by it.
1:10:15
It's I I I don't even forget addiction, it definitely fascinates people because people have always, you know, in our lifetimes been able to buy stocks, you can trade the price of oil, not on, you know, not on who's going to win an award or or really events, you know, non- non-securities-related events. Yeah.
1:10:32
How do you think about that that bifurcation between securities-related events and non-securities-related events?
1:10:42
Has there been robust enough research on how much of prediction market activity is is sports-related or sort of less like positive-sum, more zero-sum situations?
1:10:54
Because I would have to imagine that the the the level of engagement varies by category.
1:11:01
Like, there were a lot of people that were interested in tracking the presidential election, but that's not something that someone's doing every single day, whereas there's always a sports game somewhere. There is.
1:11:13
Uh so, they said us again about set during the interview, they said about 70% of their volume last month was sports.
1:11:17
In February, that number was higher.
1:11:19
Uh there was obviously a Super Bowl in February, that changed.
1:11:22
So, it was March Madness last month, but not as big of a deal as the Super Bowl.
1:11:25
70%, that's a lot, right?
1:11:27
That that's most of the volume.
1:11:27
They make the argument in the interview, you know, that one of the reasons why, I mean, obviously for Kalshi, that's good, right?
1:11:34
That's more users, that's more fees, that's how they do it.
1:11:35
They also argued that the sports volume creates more liquidity on the platform for the more esoteric bets, and and thus you need sports in order to have the other stuff.
1:11:46
I I will tell you, I do get the sense that if they could have the same valuation and the same revenue and have no sports, they'd be fine with that, probably thrilled with it.
1:11:54
But, it's not how it works.
1:11:56
The last thing I'd say about this, the problem or where sports could become a little legally complicated for them is this issue of entertainment, right?
1:12:04
The Mike Salter, the CFTC commissioner, and I mentioned this during the interview, he on CNBC the other day made a comment about how, well, this is different than a casino because a casino is providing entertainment, you know, and and it was in that clip, right?
1:12:16
There are shows and all that stuff.
1:12:19
Well, there's not a huge, but a sporting event is by definition just entertainment, right?
1:12:23
Whoever wins that basketball game tonight, except for the players and the gamblers, it doesn't mean anything.
1:12:30
No No Goldman Sachs isn't making trades based on if the Celtics win tonight. Yeah.
1:12:34
Wait, wait, wait, wait, wait.
1:12:36
I I think uh doesn't Point72 have a desk or there's there's some hedge fund that I think does have a sports betting desk?
1:12:42
>> Right, and that's kind of part of the argument they make, but in terms of this idea that that the Kalshi will put forth, which I agree with, that understanding events in an events market can can change can help people better understand the world. Yeah.
1:12:54
I just want to be 15 18 baseball games tonight.
1:12:56
None of those are going to change the world even in a tiny way except for the people who are in the ballpark maybe be happier or sadder by a couple more hot dogs.
1:13:05
Yeah, at at the same time, yeah, I mean, I completely I I completely agree with you, but there is something about when you're about to turn on the Super Bowl and you want just a really clear read on who's more likely to win.
1:13:16
Like, it is easier to understand just a straight percentage than like a line and points and all of that.
1:13:22
Like, just for a complete novice, but that's a different that's a different >> and and they'll make the and they make the argument correctly that they are not they're just taking a piece of all the action, right?
1:13:33
They're not betting against you.
1:13:34
They don't care if you win Kalshi doesn't care if you win or lose, they just care that you play.
1:13:39
Sportsbooks obviously want you to lose. Yeah. Yeah.
1:13:40
Yeah, I I had a very eye-opening conversation with the CEO of a unicorn company who I of course will not name, but I was shocked at how how invested they were in sports gambling broadly.
1:13:54
Like, I we just had dinner, we were hanging out, and they had like uh tons of different parlays across [laughter] every different app, and they and they would show me like, well, when I'm in this state, I use this app, and when I'm in this state, I do this app, and this app, I have to take my funds off the platform every night because I don't trust that it's not going to get, you know, shut down.
1:14:17
>> money doing this parlays >> know. I don't know.
1:14:18
I was just like I was just like, wow, I'm bearish on this company because the CEO is spending all their time, you know, doing, you know, 10-leg parlays on all these different apps, and you guys haven't raised a round in in years. What's going on?
1:14:35
Yeah, can I say though one one thing I I learned in the research for this, and we talked about it a bit, and and I should have probably known this.
1:14:40
One kind of user difference between the prediction markets and and the sportsbooks is that parlay issue, and and specifically, every bet on Kalshi goes to the CFTC and has to get approved, and CFTC has 24 hours to approve it.
1:14:52
What that means practically is while you can take a bet on the game tonight cuz you know there's going to be a game tonight, you can't do what you can do on DraftKings, say, the next pitch is going to be a ball, the next pitch is going to be a strike because there's obviously you have that 24 hour, they don't know there'll be a next pitch necessarily.
1:15:07
So, there's a little less real-time sports betting that than there is on on the sportsbooks.
1:15:12
But, but they are because they know the Super Bowl is going to happen, and they get that contract approved in advance, you can live trade that contract up till the last second of the game, and so that that does satisfy a little bit of that, which is again much higher frequency than oh, I want to gamble on the Super Bowl, I'm going to fly to Las Vegas, place a bet at a counter, wait in line, sit down, watch the game, go and collect my winnings.
1:15:39
It's a very, very different equation.
1:15:41
How have you processed just So, it feels like we all just agree that gambling is addictive, and I I think that's reasonable.
1:15:49
I don't know where that comes from, whether that comes from like science or law, but it all feels reasonable.
1:15:54
But, we're going through this again with social media, whether social media is addictive.
1:15:58
How have you processed the social media addiction trials, and then the new gambling app addiction question?
1:16:04
Like, are these linked at all in your mind, or how have you been processing the social media question?
1:16:11
>> I I mean, I'm obviously not a doctor, I don't think you guys are doctors.
1:16:13
I mean, there's definitely we've all I think read about kind of the dopamine hits, and look, this is that you get from not posting on social media, but when you get a like or when you get a reply on social media.
1:16:23
Gambling, I mean, there's obviously a dopamine hit when you gamble, right?
1:16:24
You know, your person hits the basket, you win the game, you get excited.
1:16:27
I mean, obviously if you have a lot of money on it, you get excited for different reasons, but just even, you know, I I'll admit I use some of these sports betting apps sometimes when I if I'm watching like my hometown team play, Dirty dawg. Dirty dawg.
1:16:39
Is that the New England Patriots?
1:16:41
May May May maybe a few times, but >> [laughter] >> look, the leagues the leagues knew what they were doing, right?
1:16:46
They knew that if there's a blowout, you generally turn it off, but if you're waiting for your guy to get 20 points, well, you might stick it out a little bit more. >> realize that. I didn't realize that.
1:16:55
But, so back to the social media question.
1:16:57
Have you been tracking any of that and what that means for the venture community or startups or really any knock-on effects that you've tied to the the trial because there was that decision in Los Angeles.
1:17:10
The actual fines for YouTube and and Meta seemed very low, but it was the whole chain of of more cases coming, and it just feel it just felt like the first time we actually had a full decision that that the judge said, yes, this is addictive. >> Right.
1:17:27
So, right, the the the lack of money is notable, right?
1:17:29
Because it was also a single plaintiff, right?
1:17:30
So, I mean, for theory, that's a lot of money for one person.
1:17:33
The The big The big question going forward is can an attorney or can a group of attorneys get a class together?
1:17:38
And you've seen a bunch of advertisements, I've seen a bunch of advertisements all over the place.
1:17:42
You know, were you harmed? Were you under 18?
1:17:44
They're clearly trying to put a class together.
1:17:46
A judge would have to certify that class.
1:17:47
That's I think what Meta, YouTube, etc.
1:17:49
are worried about, understandably worried about, because it's one thing to have a single plaintiff, but that sets a little bit of a precedent.
1:17:56
If you can have that same case with 100 plaintiffs or 1,000 plaintiffs or 10,000 plaintiffs, that money then starts to become real, and and I think that's kind of the next step here, and we have to see if it if lawyers can get that class together, if a judge will certify them, and then whether another jury and judge will will go along with what we just saw.
1:18:13
What are you tracking in the venture markets broadly?
1:18:16
We were going back and forth on the impact of Yeah, last time we were on, yeah, we were we were asking you about potential impacts to fundraising.
1:18:24
Uh fundraising outside of in in the Middle East.
1:18:29
Even then, you said, I believe, like, well, We're in sort of a new cycle of like the AI is not a bubble thing, but how are how are venture funds processing it?
1:18:37
How are LPs thinking about it?
1:18:39
Like, what are you tracking?
1:18:41
Yeah, on on the Middle Eastern side, it does not seem that money flows have really stopped.
1:18:44
Uh the way it's been explained to me is uh calls sometimes take a few more days to come back than than they were.
1:18:49
Things aren't quite as instant, but but there's been nobody who's said, you know what, we're shutting off the spigots for a while, call us call us in June.
1:18:56
Like, that hasn't happened.
1:18:58
And to be honest, even the last week or two, you've seen some deals and agreements that have come like with Saudi PIF, not just limited partners into venture capital funds.
1:19:07
There was a big private credit partnership that got announced yesterday, I think, with Saudi PIF.
1:19:09
So, clearly deals are still happening.
1:19:12
Venture fundraising as a whole though has become it it it's not even become, it is still really bifurcated, right?
1:19:18
You still have the haves and the have-nots and and these massive multi-stage firms that are gobbling up most of the money.
1:19:23
Uh and and on the AI bubble side, I mean every it it's funny, almost every venture capitalist and and certainly the industry as a whole is just looking at three things, right?
1:19:32
The IPOs of SpaceX, Anthropic, and OpenAI this year, right?
1:19:34
Because that can Those, if successful, can solve all the problems that they've had for the last several years.
1:19:40
Um and SpaceX goes first. >> Gambler's mindset.
1:19:43
Make it all [laughter] Make it all back in one trade.
1:19:46
Well, it used to be VCs used to talk about home runs.
1:19:48
This is now like the grand slam in the bottom of the ninth to win the World Series, right? And everything else.
1:19:52
And that And that's what they're banking on.
1:19:54
And And And there's not a huge IPO pipeline, for example, other than that.
1:19:58
I You know, we had a story meeting this morning, and somebody asked me, you know, well, with with the ceasefire, does that mean companies that were prepping IPOs are going to, you know, start back up again?
1:20:09
I didn't see a huge number of companies that were prepping IPOs.
1:20:10
There were some, but it's not like there was It's not like from liberation day last year when you had five or six companies that were ready to price and then stopped. Yeah.
1:20:19
>> There's not that much out there right now.
1:20:20
Again, outside of those big three. Yeah. Yeah.
1:20:23
>> Yeah, I mean, the the ones that I can think of that that are maybe more at the Figma scale are just looking at at at Figma's track record in the public markets and thinking like, I'm not as good of a business as Figma, and I don't expect to be treated any differently.
1:20:38
And when the stock popped, a lot of founders were thinking, oh, if I can get that multiple, I should be public today. Cuz that's amazing.
1:20:46
>> like the the Ripplings and the Deals and the Os. >> and the Os.
1:20:50
Deel's got its own issues, I think, which might be separate.
1:20:52
But like But By the way, this isn't just a venture issue, though.
1:20:55
Private equity firms, which, you know, more mature, slower growth, but more mature companies, they haven't been taking their stuff out, either.
1:21:02
And And though And they don't have anywhere else to go except sell to other private equity firms.
1:21:06
It It's just this broader non-IPO issue right now. Yeah.
1:21:09
Um that How much How How much are you spending Are you spending time tracking the the data center ban that that Sanders has sort of proposed?
1:21:20
It's also Yeah, I mean, a bunch of different states. Yeah.
1:21:24
It's a bunch of different states.
1:21:26
We had a We had a big AI event in DC 2 weeks ago, and I wrote about this a little bit.
1:21:31
And on the sidelines, I spoke to the CEO of Constellation, which is one of the big electricity providers to data centers.
1:21:36
And And I asked him, this was, you know, whatever, 3 weeks into the war, and I And oil prices were spiking, and I said, how much, you know, what's this doing to energy deal making in the United States right now, just the the rise in oil prices?
1:21:48
He said, it's having a little impact.
1:21:49
He said, it's the data center ban proposals that are having the really big impact.
1:21:52
That's what has people freaked out.
1:21:54
I mean, Sanders on the national level, you're you're not going to get anywhere nationally on this.
1:21:57
But on a state-by-state level, you you don't need much.
1:22:01
You need a couple states to do it.
1:22:02
It is It is something that that opponents have done a really good job on the PR, and the industry has done a very bad job on the PR on this. Yeah.
1:22:11
And And it doesn't help that everybody's gas prices and home heating oil prices in places where that's relevant, and electricity prices are all going up.
1:22:18
Iran is obviously exacerbating it, so But, you know, if you're looking at your bill, and this is something that is top of mind, well, the bills are only getting getting higher.
1:22:28
What do you think the impact of the war in Iran will be on defense tech investing?
1:22:35
I wrote about this today.
1:22:35
I mean, we have to see what happens, right?
1:22:36
It's a pretty fragile peace.
1:22:37
It's unclear whether Hormuz is even open or not now. It was, maybe it's not.
1:22:42
You know, what I wrote this morning was I mean, I thought a lot yesterday after Trump's civilization tweet, right?
1:22:45
That that if he really went through with what he said.
1:22:48
And I know some people are making some odd um arguments that, oh, yeah, well, it's either a nuke or not a nuke.
1:22:54
No, there there's a lot in between, right?
1:22:55
You You bomb some major power plants that are for civilians or desalination plants or the power to desalination plants, and people don't get water anymore, let alone can't do crops or or feed or or feed animals, etc.
1:23:07
Um I think you know, defense tech has boomed in terms of venture capital, which is such There There was almost none of it, you know, 7 or 8 years ago, and there's so much of it now.
1:23:15
I think it could have turned Silicon Valley again back against defense tech if the United States had done something that a lot of people viewed as a war crime or or at least as inhumane. Sure.
1:23:24
Uh I think the fact that we have a ceasefire and Trump didn't go through with that, uh I I think is probably a bit of a save for defense tech in the midst of a boom.
1:23:34
I I think you've got this this huge upsurge in all sorts of defense companies that could have actually come to a halt pretty quickly.
1:23:39
Not all firms, Anduril still would have invested, Founders Fund still would have.
1:23:43
But that that broader swath of venture capital that fills out those rounds, I think, might have slowed down if the military had done something the Pentagon had done something that that a lot of people viewed as as morally indefensible. Yeah.
1:23:56
Something I thought was notable is that the the American version of the Shahed was a government program, and they used some private, you know, contractors for it.
1:24:05
But the smartest the smartest defense tech play 3 years ago was just to make the Shahed. >> Copy the Shahed.
1:24:12
And make a lot of them, and it seems like no company actually had the had the foresight to to do that.
1:24:20
And ultimately had to be led uh from from the DOW. >> yeah. Yeah. Interesting.
1:24:24
Well, And And by the way, I I haven't looked to see what's happened with the stocks, but I mean, something that is going to have to happen, no matter what happens in Iran, ceasefire, no ceasefire, the the standard munitions stockpiles are going to have to get refilled, right?
1:24:36
We are using a lot of bombs and a and a lot of a lot of stuff.
1:24:39
That's going to have to all get refilled, and I know, you know, Palmer Lucky and all others have talked about how we don't have enough of that prior to all of this happening.
1:24:46
And And that, you know, that that, you know, that stuff has to get done, and there's going to be people who sell that. Yeah.
1:24:53
Yeah, that makes a lot of sense.
1:24:54
Well, thank you so much for taking the time to come chat with us.
1:24:55
The Axios show is live and available everywhere, I'm sure.
1:24:59
Go check it out, and we'll talk to you soon, Dan.
1:25:01
Have a great rest of your day. >> scoop. Great to see you. >> Thank you.
1:25:05
Uh Up next, we have Lior Susan from Eclipse.
1:25:09
He's the founder and CEO.
1:25:09
We'll be talking to him about uh Eclipse's $1.
1:25:14
3 billion raise and industrial tech shifts from innovation to scale production with companies like Cerebras Well, what's going on? How you doing? Doing well. Thanks for having me. Welcome to the show.
1:25:26
Uh please kick us off with an introduction >> and uh and some background.
1:25:33
Yeah, uh it's great to be here, first of all. Love your show.
1:25:36
Um yeah, I was started the firm 11 years ago.
1:25:39
We all operators that left their job building companies in the physical world to build a firm that we can build more than one company at a time. Mhm.
1:25:47
Uh as you can guess, it was fairly controversial 11 years ago to talk about defense manufacturing, chips, mining, etc.
1:25:55
It feels like a little bit less controversial right now.
1:26:00
And yeah, we grew the firm roughly to a $10 billion AUM, and we just announced our raised of $1. 3 billion. Let's hit the gong. He don't want to start. Congratulations.
1:26:14
Uh what was the what was the prehistory of the of the firm?
1:26:16
What How did you get into investing? What was the first deal?
1:26:19
How big was the first fund? Tell me some background.
1:26:21
Yeah, first fund, $125 million.
1:26:24
Feels like many moons ago. >> bad, though. How did you set that up?
1:26:26
I feel like a lot of people start with 20 or 50. 125's not bad.
1:26:29
Yeah, the ignorance was the power.
1:26:33
I never invested a dollar before Eclipse, so it might be that one.
1:26:39
I grew up in the military.
1:26:39
I went through the special forces.
1:26:40
Then I did a company in the networking space. Cisco bought it. I moved to live here.
1:26:44
Spent 3 years with Mike McNamara while he was the CEO of Flextronics.
1:26:50
Fell in love in US manufacturing and felt all of my friends in Silicon Valley can only spell the words enterprise software.
1:26:56
And I know nothing about enterprise software.
1:26:59
I don't like enterprise software.
1:27:00
And 85% of the world GDP is physical industries.
1:27:05
And I felt it's kind of weird that everyone is telling me, you need to go after big markets, after big towns, but everyone is following their friends into the enterprise software, while those 85% of the world GDP don't have a platforms and left to solve that problem. Yeah.
1:27:21
Talk to me about the Cerebras investment.
1:27:22
How did you meet the founder?
1:27:24
I It was one of these companies that I'd heard of, and I'd seen a lot of of of negative takes saying that it was the wrong path, that it wouldn't apply to where the current models are going.
1:27:38
And then I tried it, and it was really fast, and it just felt like magical.
1:27:40
And so, I was all of a sudden converted to be very excited about the company, where before I was sort of uncertain about how it would be. >> It's loud.
1:27:49
It felt like there was a lot to be done, but you obviously invested early.
1:27:52
How did that come together?
1:27:54
Yeah, we've been around the company for now 10 years. It's been a moment. That's remarkable. Yeah, exactly.
1:28:01
But they you know, like a lot of our companies, and I think a lot of those companies in that space, they it starts by a fundamental view that wafer-scale integration, basically the short version, you take the the the entire wafer, and you interconnect between the core. Mhm.
1:28:17
Um you know, when we have an idea of a chip, it's essentially it's a square, but it don't come from the machine like that.
1:28:22
It's actually come as a round thing, and then we cut it into squares, into chips.
1:28:26
The idea here is you take the entire wafer, and you connect it to a lot of the chips, and you create essentially one very large chip.
1:28:34
And naturally, when we made the investment in 2015, we didn't know AI will be exist.
1:28:38
What we did know, because we build by ourselves as an operators chips and factories and fabs, we knew that Moore's law is going to hit the limit from physics point of view.
1:28:50
We're going to We're in 2 nanometers already.
1:28:52
You know, let's assume we can do 1 nanometer. That's about it.
1:28:54
There is nothing after that. Yeah.
1:28:56
Um and we were looking for a new physics.
1:28:58
And a new physics mean, hey, should you can you connect to a lot of those cores in order to create a one big chip?
1:29:05
And we decided that we're going to take over companies like Nvidia and others.
1:29:10
And it it was not easy, but I think now we're extremely excited about the the company and the growth of the business. Yeah.
1:29:18
Yeah, it seems uh like it's in a fantastic position now.
1:29:21
Talk to me about Vulcan forms.
1:29:21
What was the back history there?
1:29:25
Yeah, it's in some way same story.
1:29:28
Physics start with physics of in the case of Vulcan forms where manufacturing a high precision metal parts.
1:29:35
And you know, historically a lot of those machines being using one, two, three, maybe four lasers.
1:29:41
We are using 160 laser fiber into a single head and we melt powder really really fast.
1:29:46
And the hard part there was like how to control something that is so powerful.
1:29:50
It's actually we got a call from the US uh uh um DOW many years ago and asking ourselves basically start investigate why we are buying all of these lasers because it looks suspicious for them.
1:30:05
We're like, "No, no, we are just >> [laughter] >> we're just building metal parts.
1:30:07
Nothing nothing too sketchy."
1:30:11
Um and yeah, company booked multi-billions of dollars deals last year and a billion dollar deal the year before and Kevin Caskey and the team there is doing phenomenal job.
1:30:19
We're building now four factories in US uh scaling the operation uh to build high precision metal parts for medical devices, consumer electronics, aerospace and defense, et cetera.
1:30:30
So, yeah, it's a I mean it's a big fund. You're using we.
1:30:32
I imagine you take board seats, you lead rounds. Is that roughly correct?
1:30:35
Like where how many companies do you want in the portfolio?
1:30:38
How deeply do you want to be involved?
1:30:40
Uh do you try and pick a single winner in a category or do you uh look at more like secular trends and try and get uh you know, a broad exposure to the whole category? What's your thesis?
1:30:52
Yeah, we so we you know, our LPs put us in the venture bucket and then in the growth bucket when they thinking about how they are allocating capital those US endowments and foundation.
1:31:05
Uh we call ourselves operators with capital.
1:31:07
We are all operators and founders.
1:31:10
When we build those businesses alongside the management team.
1:31:13
So, we actually do very few deals every year and we have actually a small amount of position in each fund.
1:31:19
And I'll say roughly we incubate 1/3 of the companies and 2/3 we will lead seed series A, series B, series C, series D, whatever it is. We only lead.
1:31:29
We always take a board seat and work very very closely to to the management team and tell you the truth, I'm enjoying more building companies than investing in companies.
1:31:42
So, regardless if I end up investing and maybe it was not my idea to start the company, I wants to feel like I'm part of the management team building those businesses. That's my passion. Jordi. Very cool.
1:31:56
Where do [clears throat] you think robotics is overhyped and where do you think it's underhyped right now?
1:32:02
Yeah, actually I wrote I wrote something on my LinkedIn maybe last week that say it it feels a little bit 2021 in Eclipse sectors.
1:32:12
We are start seeing some of the bad behavior that maybe was in the enterprise software in 2021 happening now in our sectors.
1:32:18
You see those companies raising a billion dollar out of the gate or some some crazy valuations.
1:32:22
Yeah, or doing doing I'm assuming there's instances where a company you guys may have backed a company in a category you know, eight years ago and then a new company gets formed in the category and within you know, six months they're valued at at the same price even though there's wildly different from a kind of uh technical progress standpoint.
1:32:49
There there is some of that.
1:32:49
I mainly it goes back to first principles.
1:32:52
As a company, as person, as an entrepreneur that like to build companies, I think you know, there is a way to build companies.
1:32:59
There is a for sure a way to build companies in the physical world.
1:33:03
You talk about building factories.
1:33:05
You talk about supply chain. You talk about capex.
1:33:06
You cannot all like you know, push a full max out of the gate, burn really fast because you believe the contracts will come and you believe that always the market will be there to fundraise you.
1:33:21
So, we just you know, trying to to bring some sort of a discipline of how to build those companies.
1:33:25
But you know, goes back to your questions on robotics.
1:33:29
We we've been doing robotics for 11 years.
1:33:31
Now as out of Eclipse and we a lot of us did robotics much before in our operating life.
1:33:37
And I think you know, we are now crossing the chasm with robotics moving from a control based PLC very custom to a much more general purpose, much more using physical AI.
1:33:52
And as a result of that we'll start seeing adoption on the commercial side that is super exciting.
1:33:58
Uh from your position on boards, I I don't know how much you can talk about this, but I'd love to know your view and expectations for the IPO window.
1:34:05
We were just talking about with Dan Primack.
1:34:07
Uh a lot of a lot of attention paid to the big three, SpaceX, OpenAI, Anthropic.
1:34:15
But what else are you seeing in terms of how companies are gearing up for the IPO window?
1:34:21
Yeah, I mean I think um it's interesting, right?
1:34:23
Of course SpaceX I think arguably even OpenAI and Anthropic have much closer um um part of the business to what I built to maybe the traditional software that kind of was leading the chart.
1:34:36
I think real assets going to have a great moment in the public market.
1:34:42
I think people value you forget it goes back to the 85% of the world GDP.
1:34:45
The reason SpaceX can have that type of an an IPO is because they solve something that it's really really hard.
1:34:53
So, it's really really hard for the second person to solve it as well.
1:34:57
I think the the reason you were seeing the correction in the SaaS world is you had a lot of companies the entry is very easy.
1:35:04
The time is not too big and as a result the public market correct.
1:35:10
So, I do believe we are going to see quite a lot of companies in the semiconductor world, in the space, in the AI infrastructure, in the data centers world going public in the next 18 months or so.
1:35:21
Yeah, what are you tracking in terms of trends in the lunar economy?
1:35:25
It does feel like we're at a turning point moment with SpaceX and and Starship coming online.
1:35:33
There's lots of interesting >> space economy, right?
1:35:35
The lunar economy does Oh yeah, it's not lunar economy.
1:35:38
It's orbital economy is the buzzword I was using. >> It's fine.
1:35:40
I'd say it's the the Eclipse name confusing.
1:35:42
Yeah, yeah, any any any variation from lower earth orbit to beyond.
1:35:47
Some of the stuff when you talk about mass driver on the moon it starts to seem 10 years away, 20 years away, harder to underwrite, harder to think about, but maybe as a venture capitalist you can start thinking about it.
1:35:57
What what what are you looking for in in just space broadly?
1:36:03
Yeah, I mean I think when you think about Henry Ford when when when we created cars, there is so much economy that is being developed and so much GDP that is being developed by the ability to move people much faster than you could move before.
1:36:16
I think you know, we are going to see something similar with the increase of our ability to travel to space and lowering the cost significantly and as a result we're just going to see a lot of new businesses being built.
1:36:30
We we are uh partnering with an amazing company called True Anomaly that building space defense prime.
1:36:40
So, you know, it's it's not only you're going to travel to space, you're also going to have conflict in space.
1:36:44
We are seeing a live one maybe a ceasefire with Iran right now.
1:36:50
But but you know, I think since I mentioned his name I said on an interview yesterday, I used to say that this is the best time to build in this country from Henry Ford and Carnegie or post World War and I actually changed it to this is the best time to build in this country period.
1:37:10
And the companies that I'm passionate about.
1:37:11
So, I'm just extremely excited to have the capital and the relationships to go and build as many companies as we can.
1:37:17
Well, congratulations on the fundraise.
1:37:21
Congratulations on the progress and thank you so much for taking the time to come chat with us. We'll talk to you soon. >> it. Great to have you. Have a good one. Cheers. Thanks, guys.
1:37:27
Up next, we will be revisiting the Axios NPM package hack that happened a supply chain attack.
1:37:34
NPM of course Axios NPM was downloaded 100 million times per week and it was compromised by North Korean threat actors.
1:37:43
We talked about a little bit on the show last week revisiting it with Faraz Aboukhadijeh.
1:37:49
[music] I hope I pronounced that correctly. What? What? Uh yes. So, uh We can ask him. Yeah, we can ask him.
1:37:56
Socket detected the malicious update within six minutes and we are lucky to have Faraz join us in >> What were you guys doing for the the six minutes? Asleep at the wheel? No, I'm kidding. >> No, no, no.
1:38:08
Definitely not asleep at the wheel.
1:38:11
It takes time to download packages, scan them, put them through our battery of tests.
1:38:15
So, I think six minutes is actually pretty good.
1:38:17
No, no, no, no, no, no, no, no, it's it's fantastic.
1:38:18
I'm just messing >> but yeah, maybe maybe zoom out and tell us about like the actual process that Socket runs, your business, how the system works and how you're able to detect supply chain hacks and cybersecurity threats so quickly. Yeah, totally.
1:38:34
So, so Socket was among the first to detect and report on this incident.
1:38:38
We built a system that, you know, goes out and downloads every open source package in existence within a few seconds.
1:38:45
So, we support about 19 ecosystems and this includes really all sorts of third-party code that might be, you know, used to build applications today.
1:38:53
Includes things like your AI models, your you know, your open source dependencies, your even your editor extensions, your Chrome extensions, like really any code coming from, you know, third-party sources.
1:39:03
And we put it through a battery of you know, really intense static analysis, maintainer behavior analysis, and then of course a bunch of AI and then human researchers as well.
1:39:14
And we we try to help kind of make a determination.
1:39:17
Is this something safe that you want to use, you know, within your application or within your organization? Yeah.
1:39:25
So, can you can you talk about the shape of the threat that was posed by the Axios supply chain attack?
1:39:29
Like because there's a wide range of, you know, zero-day exploit that gives you full access to someone's device or computer or system all the way to just something that okay, it would crash if this was if this exploit was used, right? Mhm.
1:39:47
Yeah, I mean, maybe we just start from the beginning and and summarize the exact for folks.
1:39:49
Yeah, so I mean there was a North Korean state actor that socially engineered the lead open-source maintainer of the Axios package.
1:39:58
And it was honestly quite a sophisticated and impressive effort.
1:40:02
They posed as a founder of a fake company.
1:40:05
They created a fake Slack workspace.
1:40:07
Invited this, you know, invited the maintainer to join it.
1:40:13
They staged a fake Microsoft Teams call.
1:40:16
And the the website was made just incredibly compelling.
1:40:18
They used the official SDKs from Microsoft Teams to create like really realistic components in the page. Oh.
1:40:26
They joined the call and, you know, by the way, this is they also developed a relationship over the course of, you know, weeks, right?
1:40:33
So, this wasn't like a like a, you know, a situation in which you would expect to be on guard or on the defense.
1:40:39
And at some point in the call, the call just cuts out and the browser says, "Hey, you know, you got to install an update."
1:40:46
And it gives them a a binary file that they're, you know, told to install.
1:40:50
And so, this this this maintainer thinks, "Okay, I guess I, you know, I guess I got to install this update real quick so I can get back into the call."
1:40:57
And it turns out that's how they compromised, you know, their device.
1:41:01
So, you know, it's not just like, you know, a fishing link or something like that.
1:41:04
I mean, this was a targeted attack.
1:41:05
They also targeted me and a bunch of people at our company as well.
1:41:10
So, they targeted a whole bunch of the top NPM maintainers who, you know, have access to a lot of packages. Interesting.
1:41:18
Then then in terms of once they get control over they fish a particular credential, a particular device for a developer who has access to push changes to a package like Axios, what are they actually changing in Axios to create a vulnerability in the supply chain, in the software supply chain?
1:41:35
Yeah, so they publish poisoned versions of the package that silently install what's called a remote access Trojan, which is basically a way for the attacker to just remotely control your your device and and basically do whatever the attacker wants.
1:41:48
It's like they're sitting in front of your computer on the keyboard, you know, typing whatever they want onto your onto your system. Yeah.
1:41:55
And it it what they did it what they did with it was they kind of pulled all the, you know, most interesting files and credentials off the system.
1:42:01
So, things like if you have a crypto wallet, like they're taking the keys for that.
1:42:05
They're going to definitely want the crypto.
1:42:07
If you got, you know, if you're logged in to NPM, right?
1:42:09
They pull those credentials so they can spread like as a worm and kind of continue to infect the next set in the attack, right?
1:42:16
So, it's it's actually like this self-replicating kind of cycle where they they get these credentials and then they use them to go onto the next stage. Mhm.
1:42:23
And you know, yeah, and then, you know, this is I mean, the thing I think I want to emphasize here for people is this isn't just an isolated incident because this has been kind of the the the most recent blow in this kind of series of of compromises and attacks against the software supply chain that has been happening really over over the last 6 months in a really intense manner. Yeah.
1:42:46
We've seen it really pick up in the last month with with with TeamPCP compromising Aqua Security and the Trivy scanner and then they that cascaded into Light LLM being compromised.
1:42:55
Another security company, Checkmarx, was compromised.
1:42:59
It's really like >> what happened with Light Light LLM and and how like do do you have a good sense of how that contributed to the breach at at Mercor?
1:43:12
So, it's it's part of the campaign of TeamPCP.
1:43:17
So, they dropped the same kind of self-propagating worm called Canister Worm into the package.
1:43:20
And what you have to realize is once you run a compromised open-source package on your system, you know, you kind of have to rotate all your credentials.
1:43:31
Like all your tokens and keys and and passwords and and it's a really hard thing to do very thoroughly and very completely.
1:43:37
And so, I think that we're going to see a a long tail over the next, you know, probably 12 months of follow-on attacks from this from this set of compromises because the group claims TeamPCP claims that they've stolen 300 GB of compressed credentials.
1:43:56
So, that's, you know, that I mean, think about that. 300 GB of stolen Yeah.
1:44:00
passwords, API keys, GitHub action tokens.
1:44:03
I mean, they they're sitting on so much it's like a goldmine in terms of like what's going to what's going to follow on from this.
1:44:08
So, I think it's it's not surprising that you're seeing companies affected, right? >> Yeah.
1:44:15
So, why why the boom in the last 6 months?
1:44:18
Is it it feels like it must be tied to vibe coding or or AI agents?
1:44:24
Is this that they have more powerful tools so they're able to do more damage or is it because our systems are getting weaker because we're pushing more vibe code to production? Is it both?
1:44:32
Like what are what what what what got us to this place where we see this take off in cyber security threats? Yeah.
1:44:42
Well, you're absolutely right.
1:44:43
It's definitely become a top concern.
1:44:43
I think we're hearing a lot of our customers and prospects that are contacting us that this has now become a board level concern.
1:44:53
You know, everybody is asking, "How are we not going to be affected by the next one?"
1:44:56
So, I would say that, you know, fundamentally like if you really zoom out and ask why is this a problem like why is this happening? Yeah.
1:45:02
It's because the whole software supply chain is built on blind trust.
1:45:06
I mean, you're downloading code from random people on the internet that you've never met.
1:45:10
You don't know who they are and you're like, "Let's just run it, right?
1:45:14
Like let's just hit run and like I hope it's fine, you know, I hope I hope it's going to, you know, and I'm going to give it full access to my system, right?" >> Yep.
1:45:20
No permissions model, right? No review.
1:45:22
I mean, no one looks at the code, right? Before they run it.
1:45:27
And unlike an iPhone app or, you know, mobile phone app where it has to ask for permission to do sensitive things like access your camera or your microphone or your location or your contacts or your files, right?
1:45:37
Open-source packages just get everything. Yeah.
1:45:39
You know, you just run them, they get everything.
1:45:40
So, also there's this asymmetry in security and this is always been true.
1:45:45
So, this is, you know, kind of more of the bigger picture you know, part of the bigger picture here is that defenders have a much harder job than attackers because they have to guard against really all the ways that you can possibly get attacked.
1:45:58
And the attacker has to just find one way in, right?
1:46:00
So, it's asymmetric and and so, when attackers realize, "Hey, look, you know, open-source the way that companies use it has changed in the last decade.
1:46:09
We no longer use, you know, just a handful of components like WordPress, Apache, PHP, you know, these kinds of big components.
1:46:20
We actually pull in in some cases it's like a thousand open-source libraries just to get hello world to show up on the screen, right? >> Yeah.
1:46:28
Crazy the diffusion in the number of these things.
1:46:29
So, you know, they realize, "Look, I could just attack one of these things one of these libraries and I can get into a company."
1:46:34
Like that's so much easier than attacking head-on and trying to hack the company directly, right?
1:46:40
Find one of, you know, and we have customers by the way, they have 500 500,000 plus open-source components in their environment.
1:46:47
So, just think about that, right?
1:46:48
Any one of those is a way into the company. Yeah. Right?
1:46:49
Yeah, the funniest package is is even.
1:46:51
It just tells you if a number is an even number and it's an and it has a one dependency, is odd because it's it's a Everybody loves that example.
1:47:02
Tell me more about the shape of your business.
1:47:05
I mean, it seems like you're getting a lot of calls from companies and boards.
1:47:07
Like what does it look like to work with you?
1:47:09
How are you plugging into companies?
1:47:13
Do you have a do you have a business line around going and hunting bug bounties?
1:47:17
Like how how should I think about the business of Socket these days? Yeah.
1:47:23
Well, look, people contact us when they want to get their software supply chains under control, right?
1:47:27
So, right now what that looks like is companies that are deploying AI agents and AI coding assistants across their companies have one big question in their mind, which is, you know, how do I know what my agents are doing?
1:47:39
How do I know what my developers are doing with those agents?
1:47:43
And that is the the problem that we help them get under control.
1:47:46
So, the way to think about Socket is we are a software supply chain defense company, right?
1:47:54
We protect your software supply chain.
1:47:56
So, when an AI agent is making a decision to go and install something in order to accomplish the task that's been given to it, you know, it will go through Socket first.
1:48:06
So, we are the guardrail to ensure that no malicious components get installed.
1:48:11
And if you take a concrete example, Axios, the attack we've been talking about, that malicious package was live for about 3 hours, meaning you know, anyone who was asking their agent like, "Hey, go build me whatever, right?" Doesn't matter what.
1:48:25
One of the first things is probably going to grab it because it needs to do HTTP requests.
1:48:30
It's going to say, "Oh, Axios, right?" Yep.
1:48:32
And you know, so the question is how do we how do we before that gets taken down, right?
1:48:37
Before the or even before the community is aware, how do we defend our organizations and our applications from those those packages that have these implants, right?
1:48:44
And, you know, and yeah, it's it's really top of mind for people.
1:48:49
I would I would say it's it's kind of become like a, you know, number one concern for CISOs and for boards. Yeah.
1:48:54
What what what is your view on cyber security as a category?
1:48:58
I think a lot of uh you know, we we've talked to people on air, off air that were surprised of about the sell-off in in cyber due to to LLMs just because LLMs themselves are creating all these new threat vectors and so there was kind of a disconnect there, but what is your sort of more general outlook on the category?
1:49:22
I think in the short term um security is going to get worse.
1:49:24
It's going to get harder.
1:49:26
So, I think actually I think the you know the answer is really the opposite.
1:49:30
Like companies and products and you know things like Socket are actually more needed than ever before.
1:49:35
Um you know with with Mythos coming out yesterday, you know, that's going to find a ton of vulnerabilities and you know it's finding vulnerabilities all across the software supply chain.
1:49:44
And so, you know, the I think you know more vulnerabilities discovered means there's more urgency to fix the ecosystem and it becomes it goes from being you know a lower priority on people's lists to a higher priority.
1:49:54
And so, I think you know, the short to medium-term effect is going to be massive awareness.
1:49:59
Um it's going to be supply chain security becoming more top of mind for everybody.
1:50:02
That's obviously great for us as a business.
1:50:04
It's great for the ecosystem because I think um it's hard to invest in things and get justification for budget if you're a security leader if you don't have you know a fire or an emergency to point to.
1:50:14
And so, this this really helps there.
1:50:14
I think longer term you know, we have to see.
1:50:18
I think um you know, ultimately I think AI solves the asymmetry problem that we were talking about earlier because for the first time defenders now have an infinitely scalable army of AI agents doing their bidding and doing continuous security analysis.
1:50:31
And that's all work that would have been way too expensive or impractical for their humans to do before.
1:50:37
And so, the attacker's advantage of only needing to find one way in starts to erode when the defender has the ability to kind of continuously audit everything.
1:50:44
And so, I think longer term once we get through this rough period actually I'm very optimistic about you know security improving.
1:50:51
But one thing I will say is you know with security one of the reasons I love the field and why it's such an exciting field to be in is that you know it's it's a cat and mouse game.
1:50:59
So, you're it's a dynamic system.
1:50:59
So, it's not like you know architecture or bridge building where you know you learn the the rules of physics and you know how to build a bridge that's going to you know withstand gravity and these forces that don't change.
1:51:09
You know, in security the minute you think you got things under control you know the attacker evolves the attacker switches their strategy and they have access to the same AI tools that the defenders have.
1:51:19
And so, you know, it's really a a field that is I think always going to be growing and always always going to be you know a a great business to be in.
1:51:28
Has a cybersecurity company ever got caught like sort of larping as a hacker group in order to drive demand?
1:51:36
[laughter] You know, sort of like hacking a a popular company in order to drive demand for their product cuz you said you said CISOs often times need to be able to point at a fire to to justify budget.
1:51:50
You know, that's super funny you ask because that was always the conspiracy theory that folks had about the antivirus companies back in the 90s and the in the 2000s was that they were the creators of the viruses so they could sell you the antiviruses.
1:52:00
You know, but um Create the problem, sell the solution.
1:52:06
I mean you know I I think I think that I'm not aware of any companies getting caught doing that.
1:52:11
I think there's enough bad guys out there that have realized the opportunity sitting there in plain sight that I don't think that you know it you got to go to the go to conspiracy theories to kind of explain why attackers want to >> hat needed.
1:52:24
Yeah, no tin foil hat needed.
1:52:24
Yeah, I mean uh How do you think about the these economic impact assessments?
1:52:31
When I I Axios I feel like everyone jumped on it very quickly under cover path he shared that he he didn't have the repo pinned, but he hadn't updated so he was able to dodge it for that three or six hours, right?
1:52:44
So, a lot of people got lucky, but do we have an idea of like the actual toll that that particular attack had because it felt like the number could have been very huge, but a lot of people were able to get to it fast enough that there wasn't necessarily a massive crypto breach or a massive PII beat breach, but do you have an idea of like how the industry is thinking about the size of the and the scale of the economic impact? Yeah.
1:53:10
Well, I don't have an economic dollar amount for you, but if you look at the number of downloads per week of this package it's 100 million weekly downloads, right? Yeah.
1:53:16
Um that you know, you figure you do the math on that and you figure out like what does that mean across that three-hour window?
1:53:23
I mean you're you're talking hundreds of thousands of people who installed it.
1:53:27
And that's you know, across CICD environments, local laptops.
1:53:32
That's stuff that's been shipped into production.
1:53:35
If you take you know another metric would be you know how many folks have reached out to Socket you know in the in the 24 hours following that attack to become a customer and make sure that you know they could use our tools to assess whether they were affected and to protect themselves for future attacks.
1:53:49
We had almost 2,000 organizations sign up for an account in the in the 24 hours Yeah, which you know, it to put in perspective it's a you know, it's it's a significant percentage of of all you know of our full user base.
1:54:02
So, um you know, I think this is very very widespread.
1:54:07
And this is the thing about the supply chain, right?
1:54:09
It's like it's really not a matter of like if you're going to get hit when you're talking about these very very widely deployed dependencies and you know, including even some of my own code, right?
1:54:16
I know I have these you know, you picked on is-even.
1:54:18
You know, I have some code that is similar to that a little bit less less outrageous of an example, but you know, and and and it's in it's in every it's in you know, probably almost every Node.
1:54:28
js app and that's just how it's just how the supply chain works today.
1:54:32
So, um it's it's really um not surprising that you know, everyone is going to get hit by this eventually, right? Yeah.
1:54:40
Well, thank you for coming on the show and breaking it down for us.
1:54:42
We really appreciate everything you're doing.
1:54:45
This seems more important than ever.
1:54:47
And so, have a great rest of your week. Come back on soon. We'll talk to you soon. Thanks, guys. Goodbye.
1:54:52
Up next we have Kassem Metani from Deep First announcing a big round.
1:54:55
Company also launched its first in-house model DFS Mini 1 focused on vulnerability detection in smart contracts.
1:55:04
>> [music] >> We'll bring Kassem in to the D B and L 2 100. How are you doing? Hey, guys. Doing well. Thank you for having me. Of course. Good to see you.
1:55:11
Nice step and repeat behind you.
1:55:11
Are you at an event or is this just your normal background?
1:55:15
This is like our my background.
1:55:17
You know, we had like an amazing event with the mirror of San Francisco and we got this for for that. That makes sense.
1:55:22
Well, since it is first time on the show, please introduce yourself and and the company.
1:55:30
Yeah, my name is Kassem Metani.
1:55:30
I'm one of the co-founders of Deep First.
1:55:32
We are building intelligence to discover, triage, and remediate vulnerabilities at scale in an enterprise environment.
1:55:41
We just raised a $80 million series B round from Vertex.
1:55:51
When when did you raise the last round before this?
1:55:54
We raised in early January.
1:55:55
So, it's been less than 90 days.
1:55:59
And the reason why we raised it was because we're seeing so much traction.
1:56:04
Customers are seeing so much value from our product.
1:56:06
And then we're doubling down on our research efforts like like you mentioned at the top of the segment.
1:56:11
So, we are investing really heavily on training and fine-tuning our own models.
1:56:17
Let's talk about the customer impact first.
1:56:18
What are the companies that are using your service and and plugging in and getting value and sort of walk me through the user journey of actually working with you.
1:56:31
Yeah, that's a very good question.
1:56:31
So, we work with some of the largest companies in the world, Fortune 500 companies.
1:56:37
We also work with really fast-growing startups ranging from companies like Lovable, ClickUp, Superbase, like the top names in in tech.
1:56:45
And the way they use our product is that they connect their code repository and their environments, so their staging and their production environments.
1:56:51
And then we go our agents go and figure out how the application is supposed to run and then deviations from the expected behavior.
1:56:59
So, they figure figure that out.
1:57:01
They replicate in in production and then they give like remediation instructions to agents and developers.
1:57:10
And on the research side walk me through building an in-house model.
1:57:13
What was special about that?
1:57:15
Did you have to use I I imagine you didn't do a a whole base pre-train yourself, but what what what is unique about the model and what were the keys to success?
1:57:27
Yeah, so you know, we when we started the company almost two years ago we really believed that software security is a very deep problem.
1:57:33
Now everybody in the market seems to realize that, but back then people thought that you know the CrowdStrikes and Palo Alto had a monopoly in the market.
1:57:42
But in the age of AI as code is being written faster than ever before and attackers are already leveraging AI to exploit vulnerabilities, a new type of solution needs to exist.
1:57:50
And that's what Deep First does.
1:57:52
So, we invested very heavily in building a world-class research team.
1:57:57
My co-founder Andrea Mechi comes from DeepMind.
1:57:58
He spent seven years building reinforcement learning there before LLMs were sexy.
1:58:02
This is like back in 2019.
1:58:04
And my other co-founder um Daniele was a co-founder of Fair Wholesale.
1:58:08
And before that he led security at Square and Cash App.
1:58:10
So, that's our background as a founding team.
1:58:13
And then we also have like some of the top researchers in the world working with us.
1:58:17
In terms of like building our own model, we used GPT-OSS as our base model and then we took vulnerability data. We planted flags.
1:58:27
And then we had the model try to find those flags.
1:58:29
And then we used an RL loop to to basically improve the model's performance.
1:58:33
And we were able to do better than Opus 4.
1:58:36
6 at 1/10 the cost in this particular benchmark. That's very cool.
1:58:42
What was your reaction to the Mythos news yesterday?
1:58:44
It seems like really remarkable results in in bug finding and and and vulnerability tracing, lots of partnerships.
1:58:54
How did you process the news?
1:58:56
What what are the key takeaways?
1:58:58
Yeah, I mean I think it's It's news.
1:59:00
It's like validation that security is such an important area in the age of AI, something that we believed for 2 years.
1:59:06
You know, the reason why I work 16 hours a day is because I believe that in the age of AI, like, you know, software needs to be secure.
1:59:12
So, I'm really happy Anthropic is investing in this.
1:59:14
And Anthropic is also one of our partners.
1:59:17
So, we work with Anthropic, we work with OpenAI, we work with DeepMind, we work with all the labs.
1:59:20
And our product sits on top of that.
1:59:22
So, we use the best model for the use case that the model's good at. So, we use, you know, 4.
1:59:28
6 for code analysis, we use like other product other models for capturing the flag type of vulnerability detection I mentioned.
1:59:34
So, it's good news overall.
1:59:37
But, you know, in an enterprise environment, complex enterprise environment, you need to ingest all types of data.
1:59:41
You need to figure out like the cloud environment, how like the software is deployed, you need to figure out if there's a firewall there, if there's a laugh there, and our product ingests all of that data and then gives like actionable vulnerabilities, the ones that really matter to our customers.
1:59:57
And then with a click of a button, they can just fix that.
2:00:01
So, we see that as being a significant value add of the product.
2:00:04
Talk about the decision to plant the flags yourself versus what what it appears Methos did was was just look across every single open-source project and just sort of maybe brute force a bunch of vulnerabilities until they found bugs all over the place.
2:00:20
And it seems like they were able to find a lot of different stuff by just throwing every possible hacking technique at every possible open-source repo.
2:00:26
Is that the correct way to think about that strategy?
2:00:30
And then do you think you'll wind up doing something like that in the future?
2:00:35
So, we did both actually.
2:00:35
So, we run our product our model on open source too.
2:00:38
So, like >> we found hundreds of bugs.
2:00:41
We're just responsibly disclosing them because we don't want to Yeah, of course.
2:00:44
>> get them get them out there but, you know, so attackers can exploit them.
2:00:46
So, we found vulnerabilities in Chrome, we found vulnerabilities in like Linux, like in really deep, you know, software that's existed for >> not very heavily used products. >> [laughter] >> No.
2:00:59
Only the most used products on the market. >> used, yeah. Yeah.
2:01:04
So, and that's helped us improve our product.
2:01:06
And we have a team of world-class security researchers on staff.
2:01:09
So, people who hack iPhones for a living, you know, and thankfully they're working for us.
2:01:12
But, like, those types of folks who are going and validating the the results Yeah.
2:01:16
and then helping us improve the model based on that and improve the product and the model.
2:01:20
Well, thank you for everything that you do.
2:01:23
We need more white hat hackers than ever, very clearly.
2:01:26
We were just talking about the Axios hack.
2:01:29
One one final question Tyler on our team wanted us to ask, why not use are you fine-tuning on any of the Chinese open-source models or do those scare you?
2:01:43
We we are experimenting with some of them, but we would we're an American company, we would love to use American models.
2:01:48
I was actually I met Jensen Huang yesterday and it was so amazing to see the investment that's going in in this area, especially in training open-source models. >> models too, right? >> Yeah.
2:02:01
Yeah, so we're very excited and we're partnering with Nvidia and he loved our vision.
2:02:03
He he thinks that in the age of AI, I mean, as agents are everywhere, security is going to be like extremely important.
2:02:09
So, he's completely bought in to our vision to our vision and he's really excited about it. Yeah. Very cool.
2:02:15
Congratulations on the progress in the round.
2:02:17
We will talk to you soon. Have a good day. >> Thank you. Talk to you soon. >> Goodbye. Thanks for having me. Goodbye.
2:02:23
Up next, we have the co-founder and CEO of Mutiny.
2:02:28
Mutiny just raised $72 million from Sequoia Capital and Y Combinator reaching eight-figure ARR.
2:02:32
We'll bring in Jalay Razay from the waiting room into the ultradome. What's going Good. How are you? We're good.
2:02:40
Thanks so much for joining the show.
2:02:43
Please give us an introduction of yourself and the company. So, I'm Jalay.
2:02:49
I'm the co-founder and CEO of Mutiny.
2:02:53
And yesterday we announced the new Mutiny, which is an AI agent that companies like Rippling and Snowflake use to create anything customer-facing in order to get a deal from cold all the way to closed. Okay.
2:03:06
Yeah, walk me through I mean, what does that actually mean?
2:03:11
Add assets to landing pages, battle cards, like walk me through the the workflow of closing customers in the modern era. Yeah, absolutely.
2:03:21
So, starting out, you probably want to warm up the accounts in a particular vertical. Mhm.
2:03:27
And so, our customers will create personalized vertical campaigns.
2:03:30
And then from there, once, you know, the the SDR is involved, they want to start prospecting and get meetings with the right people.
2:03:38
So, they can make prospecting pages in Mutiny.
2:03:42
The agent can even research the specific people that they want.
2:03:44
They can pull in data from their CRM, any any information that's available to them, the agent will access and create something really high quality that will stand out to that prospect.
2:03:55
As the deal progresses, now we're looking at things like curated customer case studies, we're looking at business cases, ROI reports, pricing proposals.
2:04:06
Even after the deal closes, there's a ton of expansion that the customer success team will drive.
2:04:10
So, they can create impact reports in Mutiny for their customers, and they can do look forward strategies, the whole works in order to maximize revenue. Okay. Bunch of questions.
2:04:21
Where does the name come from?
2:04:26
>> [laughter] >> You know, the mission of Mutiny was all about killing the dependencies in go-to-market teams.
2:04:33
I've led marketing teams, sales teams, and the biggest blocker to growth is always speed.
2:04:39
And the blocker to speed is all of the little dependencies that exist >> Sure.
2:04:44
inside of your team, outside of your team.
2:04:47
And so, it was really a mutiny against the status quo.
2:04:50
That's where the name came from and it just kind of stuck.
2:04:53
And behind you, is that a raccoon mascot? Explain that.
2:04:57
[laughter] >> Yes, it is.
2:04:58
This is our this is our our raccoon mascot. His name is Achoo. Achoo.
2:05:05
Where where did that come from?
2:05:06
How did you pick a raccoon?
2:05:08
Do you want to know the real story? >> Absolutely. >> Yes.
2:05:12
So, we were all in a circle.
2:05:12
This is when we were about four or five people.
2:05:16
And we're like, what are we going to name the raccoon?
2:05:17
And one of our early employees >> how how did you get to raccoon?
2:05:20
You're just like jumping of course we're going to have a raccoon.
2:05:25
No, explain like how did you pick raccoon?
2:05:27
Cuz there's a million animals you could pick. Yes. Okay.
2:05:29
So, we were designing our brand and the designers asked me, okay, is there an animal that you guys really identify with?
2:05:39
And there wasn't really anything coming, you know, off the top of my head.
2:05:42
And then we took the whole team to Angel Island on a on a camping trip.
2:05:45
And the entire time we were there, we had six bottles of wine with us and basically no supplies.
2:05:53
So, it was just it was it was awesome.
2:05:57
And every time we would turn around with our headlamps, we would see this gang of adorable raccoons just slowly approaching.
2:06:04
And then they would see the light and they would start backing up.
2:06:08
And so, the next day the designer asked me that question again and I said raccoon.
2:06:11
And that's how we ended up with the raccoon. There we go.
2:06:13
What's going on with email?
2:06:18
Are you are you generating cold emails?
2:06:20
Is is it a waste of time now?
2:06:21
Like Yeah, what's the equilibrium here?
2:06:24
I think a lot of people are getting more cold outreach than ever and it feels like we might be in this game theoretic >> I can imagine you guys helping somebody make a great a great cold email, but at the same time you guys are also set up for the golf and steak GTM as well, which is you play a nice round of golf and you have you know, afterwards you pass them a PDF. Yeah.
2:06:43
Or a little a little deck gives them some more context on the conversation. Exactly.
2:06:51
So, email is a really tricky one.
2:06:52
I think, you know, we see this in our own data, we hear it from customers, the results are are really bad.
2:07:01
Most people don't really open emails anymore, executives don't really open emails anymore.
2:07:07
And so, the the engagement rate on email is really really low, which is why I think having a really personalized approach that's going to stand out, that's going to be different, that's truly and genuinely tailored to that person is going to be really important.
2:07:22
Um, one of the things that I find really fascinating is if you look at an average salesperson, they spend about 30% of their time selling and 70% of their time following up with customers, getting ready for tomorrow's meetings, creating all of those materials, nurturing the old deals that are going to convert hopefully one day.
2:07:42
And when you talk to CROs for the most part, despite all the AI investment in data, they haven't really moved the needle in terms of increasing quota per rep.
2:07:54
The rep is largely closing the same amount as the previous years.
2:07:59
And I think the reason for that is that that 70% that's really skilled custom work per per, you know, customer that you're going after.
2:08:12
I was on a call a couple weeks ago where it was a great call, great enterprise brand, the right decision-makers in the room.
2:08:19
And at the end of the call, they're like, please send me um based on the challenges that we told you we have, send us the three metrics that you can move for our business and relevant customer case studies for each of those.
2:08:32
That would take a rep 4 hours to go create.
2:08:35
Like, you have to go look at hundreds of case studies, pull those things together.
2:08:39
Whereas in the Mutiny agent, they can just come in and it automatically will pull in the challenges from the Gong transcript, it will it will go through all of their case studies, it will sift and pull out the right stats, it will curate the um assets in there, and then they can go ahead and send a really nice, beautiful, forwardable thing to their customer that's going to get shared with the whole buying committee. Yeah.
2:09:06
Uh the chat is asking for the name, where the name for the for the raccoon came from because I think we glossed over that.
2:09:13
So, sorry to go back to the mascot, but the the mascot is a raccoon named Achoo. >> answers. Yes.
2:09:21
So, we were it was the same group of people that went camping. Yeah.
2:09:25
We said, "What should we name the raccoon?"
2:09:28
And right as we were going to do that, someone sneezed, and it just said, "Achoo."
2:09:35
And we all went, "Achoo, that's actually a really good name.
2:09:37
Let's go Let's go with that."
2:09:39
I mean, in general, I would say the Mucinex brand um I think part of the reason people really like it is that it is raw, it's authentic.
2:09:46
We don't really regulate what people can and cannot do.
2:09:51
We hire people that are aligned with our values, and we just let them be themselves. I love it.
2:09:57
Well, thank you so much for taking the time to come chat with us. Congratulations.
2:10:00
Uh did you Did we hit the gong for you?
2:10:02
You raised a $72 million round. We got to smash it.
2:10:04
That was a That was a previous fund raise, but yes.
2:10:13
Yes, well, we're still happy to celebrate it. Thank you so much.
2:10:15
So, that's all that matters. That's all that matters. We'll talk to you soon.
2:10:18
Have a great rest of your day. Goodbye.
2:10:22
And up next, we have Jeremy Gallant from Sham Charlemagne Labs.
2:10:27
Uh he spent 12 years at Meta [music] in trust and safety and left last year to focus on AI-powered scams and building defenses.
2:10:33
We're doing a whole security-themed show. Look at that. Look at this. Suited up. Wow, the matching suit. You look fantastic.
2:10:40
Wow, >> [laughter] >> you really like it is the mirror image of me. This is crazy. Nailed it. The memo came through.
2:10:46
I was hoping that I'd get that Maybach right downstairs and do the show.
2:10:50
>> [laughter] >> I'm so glad you were here you're you're you're you're up to speed on the show.
2:10:55
Uh but for those who aren't up to speed on you, uh give us an introduction and explain a little bit of your background.
2:11:00
>> intro, you said you said he left Meta to focus on AI scams, which AI-powered which kind of sounds like you're scamming, but I'm assuming it's the exact opposite. It's the opposite.
2:11:09
We're doing cyber security. That would be too easy.
2:11:11
It's much easier to be on the offense than it is to be on the defense today, I tell you.
2:11:16
It's It's It's wild out there.
2:11:16
So, yeah, I left Meta after 12 years to uh focus on on AI threats.
2:11:25
And basically, my vision is that every employee of every company would have um a watchdog.
2:11:30
So, the company's named after my uh my dog, Charlemagne.
2:11:33
She goes by Charlie, so the product is called Agent Charlie. Yeah.
2:11:36
The idea is like you're using your your computer, and you're getting attacked now with um novel kinds of threats that resemble legitimate communication that could be on messaging apps.
2:11:47
It could also be, you know, the standard phishing and >> Yeah, we we just heard about the with the Axios attack.
2:11:55
Uh it was a basically a fake Microsoft Teams uh basically call that then cut out and trigger and and you know, suggested, "Hey, update Microsoft Teams."
2:12:05
Whole thing wasn't Microsoft Teams, but uh the individual just was like, you know, confused because it just seemed uh like it was from >> I think the nastiest trick is when it's the unsubscribe button is itself a link.
2:12:19
I think that's like the nastiest thing.
2:12:20
So, what I what I we've built you know, the startup um has has um been selling a product that will try and stop you from clicking.
2:12:27
So, it's like, "Bad Bad employee. Do not click."
2:12:29
Um but the the research that uh we've done to to inform this commercial product is into the capacity, the capability uplift that's happening um with respect to offense.
2:12:42
So, it's important to remember that if you're an adversary that's a threat actor um seeking uh you know, financial gain or a state actor, you're availing yourselves of all this uh AI and agentic tooling that that that we are using, you know, the sales tools, the uh you know, the automation.
2:12:58
And so, you know, the the core premise is that in in an AI-powered world, all phishing becomes spear phishing.
2:13:06
You're not going to get a Nigerian prince email, you know, much anymore.
2:13:09
You're going to get extremely realistic, utterly compelling uh request from your boss or your manager or your friends, and it's going to be catastrophic in consequences.
2:13:20
Uh how do you think about uh actual deployment?
2:13:23
Because this sounds useful in a consumer context.
2:13:24
I'm just thinking about, you know, the the the email that's from your bank and has the unsubscribe button for some marketing email.
2:13:32
You click it, and all of a sudden, you're logging in giving away details.
2:13:34
Uh is there an important distinction?
2:13:36
It feels like consumer and enterprise is blurring together in in many places.
2:13:40
What How do you think this all plays out? Absolutely.
2:13:42
I think uh as employees of companies, we are using personal email and personal messaging apps on our on our devices, for sure.
2:13:52
I think as a business, we're a B2B SaaS company with the research arm, and I'm excited to tell you more about our our research effort.
2:14:00
But um yeah, I mean, my dream is that um if the AARP is listening right now and would uh give this, you know, for free to I'd like to give our software for free to anyone who holds an AARP card because elder abuse is devastating, and it has huge consequences, but it's very difficult uh to market and sell to consumers you know, a product like this.
2:14:20
People don't wake up and say, "Today's the day I'm going to improve my security posture."
2:14:24
So, after [clears throat] their attack, they have a problem and a mess to clean up.
2:14:28
So, we're What you need to you need to create the problem No, stop with [laughter] this.
2:14:33
Stop with creating the problem.
2:14:33
No one's creating the problem.
2:14:35
>> just we were just we were just talking with uh Harris from Socket who was who was saying like the the old tinfoil hat theory with with cybersecurity and like malware products is that you know, they would create the bugs and then sell the sell the malware.
2:14:48
Create the viruses, sell the antivirus.
2:14:50
I think that's unethical, but also, we don't have to do that Yeah, there's plenty of scammers out there.
2:14:54
The bad guys are getting, you know, superpowers.
2:14:57
And so, all we have to do is wait.
2:14:59
Um and and like I said about, you know, the sort of research arm, our um our team is has done some work Meta's um you know, model dropped this morning. We work with them.
2:15:07
They're They're um I think, you know, I'm quite proud actually of of what they're doing in the the cybersecurity space because beyond infrastructure and coding attacks, what we're um we all know and aren't really talking enough about is that humans are you know, the weakest link.
2:15:25
So, when a company um wants to secure its perimeter, you know, it's critical that employees are trained, and today, you know, their training exercises um but but the um social engineering attacks aren't studied as much.
2:15:37
And so, >> [clears throat] >> um yeah, I'm really excited that Meta's taken a lead in going beyond just, you know, infrastructure and code vulnerabilities to looking at the capabilities that um that models frontier models might provide adversaries in the social engineering and scam space.
2:15:53
Yeah, so explain a little bit more about uh the the eval suite for Mew Spark.
2:15:58
Uh because uh like is it that the model is trying to Is the model social engineering you, or you're trying to social engineer the model?
2:16:10
Like, what are What are the two parties in this in this uh eval?
2:16:12
Like, actually, how are they interacting?
2:16:15
Yeah, so we use um an industry practice called LLM as a judge.
2:16:21
So, we don't test on human subjects, and our eval suite takes a model and has it role-play as an attacker, and then we have a a model that role-plays as a victim, and they're given, you know, instructions accordingly.
2:16:33
And then we have an LLM judge whether um it's the specific attacker is succeeding, and then we compare those attack different models to each other in the in the role of attacker.
2:16:44
That's how we measure the kind of uplift or capability. Yeah.
2:16:48
Do you think that uh Is there Is there a world where these social engineers like I'm thinking of different bending points in where if someone's running like granola, and they're recording that particular uh it wasn't a Zoom call, it was a Teams call for the Axios attack, and maybe an AI model could be listening in the background and sort of throw up a flag like, "Hey, it's not It's actually there I just checked.
2:17:13
There's no update for Teams.
2:17:16
You don't need to click on that binary.
2:17:17
You don't need to install that.
2:17:19
Uh this person's trying to take advantage of you."
2:17:22
That's exactly the vision for our our commercial B2B security product, Agent Charlie.
2:17:27
I want an agent that that, you know, the technology that we use is small language models, so that it is on device, and thus is limited in its capabilities. Oh, yeah.
2:17:37
I see a future where you have a real-time um AI for security exactly like you described.
2:17:43
I think real-time audio analysis with an SLM is is way too big an ask, but small language models are improving um you know, just like all all of the large models.
2:17:52
So, yeah, I mean, we need real-time defense.
2:17:54
I want it to be proactive, too.
2:17:56
I think the biggest issue is that when scammers succeed, it's because even intelligent and well-trained people, employees of companies that work in tech even, are duped because it's a It's as old as the Bible.
2:18:08
Scamming is a is an ancient art, and it has nothing to do with with preparation anymore.
2:18:13
It has to do with you know, we're we're being attacked by a machine. We need machine defense.
2:18:18
Yeah, no, that makes a ton of sense.
2:18:20
Where How Take me through the shape of the company. How big are you? Have you raised money?
2:18:24
How long have you been doing this? All this.
2:18:25
Yeah, so I I've raised money last um from the three investors that we're really excited to be working with.
2:18:32
Um They're Kevin Carter of Night Capital and Chris Howard of Ritual Capital and Rafael Corrales of Background Capital.
2:18:41
Collectively, they've uh backed more than 30 unicorns from idea stage.
2:18:43
And so, you know, I tell them that I want to be the 31st.
2:18:48
I'm ready to I'm ready to go. We've got Good luck. There we go. Yeah. Yeah. Love it.
2:18:54
So, we you know, we we're in a kind of stealth mode right now working with design partners on the SLN's capabilities.
2:19:00
And we're also, you know, if you visit our site, you can actually self-serve for the real-time phishing defense.
2:19:05
You can sign up right now if you have a probably have a Centurion, but if you have you know, a credit card that that works, you could you could put that into the into our website right now. Don't get spearfished. Yeah.
2:19:20
Well, thank you so much for coming on the show.
2:19:22
Congratulations and good luck with the next phase.
2:19:24
And Thank you for suiting up as well. Yeah, we appreciate it.
2:19:27
>> Yeah, I'm I'm not wearing any pants, by the way.
2:19:30
>> [laughter] >> Well, have a great day. See you later, Jeremy.
2:19:34
We'll have you back on soon. Goodbye.
2:19:36
Um and people were disappointed that we didn't go more into the the the story about Satoshi.
2:19:41
There is a full deep dive in the New York Times, My Quest to Solve Bitcoin's Great Mystery.
2:19:50
It is a long article, though, and so I think we'll have to touch on it another time.
2:19:56
But you know, we went through Adam Back's reaction and his his disavowal of the accusations that he is Satoshi.
2:20:01
But there's a bunch of interesting little segments in here from the the forums and the message boards of the day analyzing the different writing styles trying to see Did you Did you dig into this at all anymore?
2:20:18
I didn't read the whole thing, but I'm like people have speculated that's Adam Back for a long time.
2:20:21
Yeah, it's like kind of like him and Hal Finney is the other one.
2:20:24
These are kind of the two like main names that people >> And and there's one more I think that comes up all the time, but I can't >> There's Nick Szabo sometimes. >> Szabo, yeah.
2:20:30
Um but yeah, I I don't know if there was a lot of like new facts that came out with this, which I think is is why it's like not like super super crazy. Yeah.
2:20:38
There was also an HBO documentary on Satoshi.
2:20:40
I forget who the like who who did that Satoshi who who did they accuse in the 2024 HBO documentary directed by Cullen Hoback.
2:20:53
Um Uh oh, the firm suggests the Canadian software developer Peter Todd is Satoshi. And Todd denied that.
2:21:00
And so you have That's got to be the worst kind of title in the world from a security standpoint is being accused of being Satoshi. >> Yeah.
2:21:12
Because you're just going to be attacked cuz you potentially have the keys to like 50 billion dollars or something, maybe more.
2:21:17
I forget exactly what the number is, but yeah, that wallet is big.
2:21:21
I still think it's possible that like the uh the Satoshi wallet, like the keys were just lost and the person it's like sort of a lose-lose cuz if you admit that you lost the keys, then like everyone's like, "Oh, how do you even prove that?"
2:21:34
You can't prove that you lost something, but there's no movement. I don't know.
2:21:37
Yeah, also there's like you you could have someone could have created it and then had years and years and years and years to buy up, you know, an equivalent amount of supply a bunch of different ways.
2:21:48
And then you have basically you can say like, "Well, I've never sold." Right?
2:21:53
If if if Satoshi's wallet did start selling, it would probably >> Yeah, from a lore perspective and the brand, you could potentially be making plenty of money from the other wallets.
2:22:04
And then if that supply ever moves, the whole market's going to re-evaluate the the basically the liquid supply and sort of tank what you have.
2:22:12
And also just the like the the the aura around Bitcoin is that it has an anonymous founder.
2:22:17
And if it was if the founder was ever truly unmasked, it would be so much less of like a special project that I think everyone involved wants to keep it that way.
2:22:26
Although these investigations will never cease to be interesting.
2:22:29
And so you can go read it on the New York Times from John Carreyrou.
2:22:34
Anyway, thank you so much for tuning in today. Bit of a shorter show.
2:22:38
We're experimenting with different things.
2:22:42
Obviously, we don't have ad reads anymore.
2:22:43
And so we are going to be mixing it up with more stories, more interviews, different timing, and more flexibility.
2:22:50
And so we hope you enjoyed the show and we'll see you tomorrow at 11:00 a. m. Pacific sharp.
2:22:54
Goodbye, small We love you.
2:22:57
Leave us five stars on the podcast and Spotify.
2:22:58
Sign up for our newsletter at tbpng. com. Goodbye. Cheers.