0:00
Almost all of investigative journalism is solicited.
Search this video
0:02
In other words, it's someone coming to you, the journalist, with a story they want you to run.
0:07
It's been like that since Watergate, right?
0:11
Watergate famously a solicited story.
0:14
And I'm not saying that it's never good.
0:16
It's never ethical to run with a story that's solicited.
0:19
Of course, sometimes people come with and come to you with a story and that's really important news, but that actually warps the entire ecosystem of what you see in front of you. Well, hello everyone.
0:37
It's Jim Oanis with yet another Infinite Loops.
0:39
I am very excited to have Brian Chow, the founder of Epic News, the innovative open-source investigative journalism site on as a guest today.
1:00
Brian, I love this idea so much.
1:00
We have our own on-prem [music] AI installation hardware software models that we use.
1:10
Uh, and this was one of the ideas and you beat us to it. So, congratulations. >> Thank you.
1:17
It's been really surprising both the stories that have gone around and people just didn't investigate at all and the underlying infra because people have been talking about AI and journalism for basically like as as long as I've been around like this is pre- chat GPT people are talking about AI and journalism and just no one has done it to any effective degree at this point.
1:44
I think everyone tries to do like AI for writing, right?
1:48
Everyone tries to make AI write the articles.
1:51
And I don't know if you've noticed this, but when AI writes the articles, they are just terrible.
1:57
They are they are utterly terrible.
1:59
And no one has really made a dent in using AI to investigate these troves of legal and financial documents, which is what we're doing, and we're changing that.
2:13
So for for listeners and viewers who are not familiar with uh your your new uh endeavor, explain exactly uh for our listeners and readers what you're doing, why you're doing it.
2:28
I know that I I'm a huge fan of tech making things possible that were impossible before, right?
2:34
So, the first book I wrote was called Invest Like the Best: How to Use Your Computer to Unlock the Secrets of the Top Money Managers.
2:44
And it was literally, hey, no human being can go through the 1600 annual reports of every company uh that get released on a quarterly basis, but it's trivial for a computer to do.
2:57
That was way back in the 90s.
3:00
Now, we've got the supercharged AI.
3:03
It can go through tens of millions of records.
3:06
But talk about a what was the motivation?
3:10
Uh b how you put it into practice effectively.
3:14
And another thing I really want you to comment on is something that I love and that is you published the null sets.
3:22
I think that's really really really important. >> Yeah.
3:29
It's really shocking even before you get to the null sets.
3:32
The way that journalism is shock is sourced is totally shocking to me.
3:38
And may maybe you've been around the block enough that you you kind of just know this and have known this forever, but almost all of investigative journalism is solicited.
3:51
In other words, it's someone coming to you, the journalist, with a story they want you to run.
3:56
And I'm not saying that it's never good.
3:58
It's never ethical to run with a story that's solicited.
4:02
Of course, sometimes people come with and come to you with a story.
4:05
And that's really important news.
4:07
It's genuinely important and it's like totally a good thing that that they run that.
4:12
And it's been like that since Watergate, right?
4:15
Watergate famously a solicited story.
4:18
And it's been been like that actually long before that.
4:20
But that actually warps the entire ecosystem of what you see in front of you.
4:27
of what you see in front of you. And until very recently, it was physically impossible to have stories that came basically any other way because to do something like what we're doing now in which for example we query every single row of the uh HHS grant table and that's
4:48
how we find find a story and we do that literally exhaustively literally looking at the same criteria for all grants saying okay which grants have spiked in the last uh five years, which grants have grown a ton either proportionately or as as an absolute value and what relationship those might have to something that's newsworthy. And
5:13
And you know, I don't I really don't blame journalists preai because like you said, if you were to do that manually, that would take you literally a lifetime.
5:25
And so just based on the constraints of practicality, not necessarily due to any particular, you know, motivation, it was not possible to do that.
5:33
And now we can have um journalism that we source through these exhaustive investigations into legal and financial records.
5:42
And we can actually have a more scientific way of approaching these things where it's like, okay, here are the data sources that we've searched.
5:50
Here are the um here are the prompts that we used.
5:53
Here are the results that we have.
5:56
And here are the investigations that we tried to do and that we actually, you know, hit a brick wall.
6:02
We didn't we found nothing. Yeah.
6:04
And the I I love the nothing because uh you know, learning via negativia is something that is not natural to we humans, right? For the most part.
6:15
I was a big Sherlock Holmes fan when I was a kid and I remember the story.
6:20
Did you want the Sherlock Holmes cold cases? >> Yeah. Yeah.
6:23
[laughter] Um, and I always love the story where he knew that the intruder was known to the family because the dog didn't bark.
6:31
And that's something people just have a hard time getting their heads around, right?
6:39
All of the null cases that sound very enticing or like a motivated reasoner could say, "Well, of course there's a conspiracy.
6:48
Well, of course this is going on.
6:50
And then null said, "Yeah, we we scoured a 100 million data points and we're not finding anything."
6:59
Um, I wonder though is you mentioned the a traditional journalist will generally look to sources, right?
7:11
and then basically go try to either confirm or deny what the source is telling them.
7:20
Here the sources are you know a cornucopia of documents that have been required for filing for public information etc.
7:28
What what about uh situations where there are very little reporting requirements and or for example the size of the offshore empire.
7:43
I call it the empire of elsewhere uh because of the way um it gets cleared through the city of London uh but settles in uh jurisdictions that used to be part of the British Empire but have very very lax uh financial uh diligence and and uh paperwork requirements.
8:07
How how do you get into things that aren't in a Freedom of Information or or or aren't one Freedom of Information Act request away?
8:18
>> I I want to acknowledge that there's a lot of truth in in what you're saying, but also a lot of the time these barriers are more paper barriers than you might think.
8:29
So we had one story where we uncovered the donors behind this degrowth organization in Europe called uh partners for a new economy P4NE.
8:41
And what we did there is that they were actually hiding their money.
8:45
It's funny enough this this degrowth organization calls for calls for an end to growth a shifting of economic policies away from prioritizing growth. Who are they funded by?
8:54
They're funded by these 19th century European heirs who are hiding their donations through literal Swiss bank accounts.
9:03
So they had this uh big essentially money mixer legal version of a money mixer called Swiss Philanthropy Foundation.
9:11
And what Swiss Philanthropy Foundation does is that it fiscally sponsors 106 different uh essentially like NOS's.
9:23
And whenever anyone makes a donation to any of the 106, including the degrowth organization we were investigating, on paper, that's a donation to Swiss Philanthropy Foundation.
9:33
In theory, no one knows who it's going to.
9:35
It no one knows if it's going to P4 or anyone else or any any one of the 106 different foundations that they have in there.
9:42
So, how did we crack this case?
9:45
We ended up using those donation rows to trace back the donor side disclosures.
9:51
In other words, in a lot of cases, even if P4NE doesn't have to disclose their donors because technically it's being done through the Swiss Philanthropy Foundation, a lot of the time the donors have their own disclosures that they need to do.
10:06
and you can go go line by line and see what matches and then go to the donor rows and then see if they have a record for those same transactions to Swiss Philanthropy Foundation and that's how we trace back a lot of the donors to P4NE.
10:22
So there's both cases where you're right, there's more of a layer of obscurity, but I think there's a big juicy middle there where there's a lot of obscurity that does enough to offiscate you from traditional journalists or or from people who are just doing it by hand or just doing it using traditional software and from some of the tools and the search mechanisms that we've unlocked at effort.
10:48
And I think that's true on a national and also an international level.
10:53
You mentioned you mentioned routing offshore.
11:00
We know where a lot of these accounts are filed.
11:02
A lot of them are Hong Kong. A lot of them are Dubai.
11:04
A lot of them are um kind of like vaguely Saudi related accounts.
11:09
Like I I don't know if I'm saying anything I shouldn't on this, but this is all all pretty wellnown.
11:13
And you know they're they're well-known operators that that'll teach you how to do it as well.
11:21
And I think that there is a lot more to uncover there.
11:25
Maybe that's not the ideal.
11:28
Maybe that's not the pinnacle of 100% transparency, but I think just as a practical effort, we can put together a lot on that front still. >> Yeah.
11:38
I'm I'm I'm finally writing a fictional novel that I've wanted to write.
11:42
I had the idea for it like 34 years ago, and I'm finally writing it.
11:47
And we're using a very uh or experimental technique where we have a writer's room.
11:51
It's me generating all of the story and all of the uh ideas behind it, but then the writer's room is like, "Yeah, Jim, this this doesn't make any sense."
12:01
And how do you ever think about that?
12:03
It's the way a lot of TV shows are produced.
12:06
>> Yeah, I was just about to say, >> "Yeah, >> you're writing the Seinfeld sci-fi novel."
12:10
>> We are actually putting AI editors and writers into the writer room, too.
12:12
So it's I'm a big believer in the senitar model which is man plus the tool right I look at AI as just an incredible tool that empowers us we humans uh to uh to uh make the story better etc.
12:31
One of the things I bring that up because as I mentioned >> if you can say how how well is that going I I I think a lot about using AI to improve my writing and I've not really gotten anywhere with this.
12:44
So I' I'd be curious on how that's going with AI editors.
12:48
>> I it's it's uh the editing on uh certainly on copy edits uh it it is virtually perfect on factchecking now that hallucinations have gone way down.
13:02
Again, it's difficult because we have our own onrem which has been fine-tuned for the things we're looking for.
13:07
So, it's not we're not using a just a commercial model off the shelf.
13:11
Um, and and so it's also very good at that.
13:14
On the writing, it's a very good critic of writing, but that's up to you.
13:20
By that I mean you've got to come up with a prompt where you get rid of all the syncopanti where you get rid of and so I came up with one that was like I want you to embody like one of the most ariodite well- read uh uh critics and I want him to hate everything that I'm going to put in front of him.
13:44
And oh man did that hurt.
13:48
And the thing that is really interesting is it caught things that I would not and my human writing team and and editors also missed like oh we didn't even we didn't even think about that.
14:02
So it's really really good at all of that.
14:05
On the writing side it's getting a lot better.
14:06
Uh but what you want it to do is you just have to constantly reinforce no this is not the voice. This is not my voice. Sorry. Gone.
14:18
uh but I suspect that as it as it goes along uh it will improve dramatically.
14:26
But the the point I was going to bring up was as now remember this is a fictional novel but I run it through not only a panel of experts that we generate.
14:37
So if a chapter happens to be about um offshore accounts, we spin up expert the panelist members who are expert in that particular thing.
14:46
We spin up experts who in the city of London and the remembrancer and and the bank of international settlement etc.
14:54
And then the reasoners are the ones that check all the facts.
15:00
And we use an adversarial system there.
15:03
We have one group of agents saying this is true and we have another group of agents saying no it's not and here's the evidence.
15:10
Uh, but I ran a fictional chapter through it just recently and it kind of blew me away because remember this is fiction but it's looking for the facts around the town I'm using, right?
15:25
And it comes back one of our villains starts a terrorist campaign in in Europe and one and one of the things that is part of that is an electrical grid attack, right?
15:38
and and the AI came back with, "Yes, that actually did happen at that particular time."
15:46
But what you maybe didn't know was the family that electrified that town in in this particular country, they had their own house on top of a hill that they had their own electrical system built for.
16:00
So even if the entire town went black, their house would still be lit up.
16:06
And I'm like, "Wait a minute."
16:10
[laughter] And so I went the traditional route and went and found the paper records and found all of that.
16:15
I'm like, damn, that it it was right.
16:18
So I suspect that in what you're doing, you you can find all sorts of similar items that don't intuitively appeal to us. >> Yes.
16:31
And there's all sorts of things that become notable via the combination of information.
16:36
The most obvious example is plagiarism, which is now in the news for for other reasons.
16:41
But you have one person write a thesis, that's no big deal.
16:46
You have two people write the exact same thesis, now that's news.
16:48
And it's a lot like that for these entire categories of crimes or of newsworthy findings.
16:56
Take take fraud for example, right?
16:59
You have one ledger that's just all of the expenses.
17:01
Okay, that's like, you know, uh m maybe you can say maybe you can look at that and say, oh, they're they're like paying too much for these services.
17:08
They're they're they're running inefficiently.
17:10
Okay, but it is another level of news newsworthiness if you have two ledgers.
17:15
One for the people who are getting the getting paid and one for the people who are paying and they don't match.
17:21
Now, that that's a that's an entire different category of information. >> Yeah.
17:26
And one of the things I thought is, you know, the the old way of doing things was, you know, a reporter would start with a hypothesis and then he would go he or she would go and try to find evidence in favor of that hypothesis and ideally uh information that was at odds with that uh hypothesis.
17:48
But aren't the prompts you're using?
17:52
Isn't this just like your prompt is your hypothesis?
17:56
No, >> I think that is true at an abstract philosophical level and I think that is less true in practice where there are there are degrees to this.
18:09
For example, when we found the uh office of refugee resettlement grant story, this was our biggest story at like hundred or sorry, it had a million and a half views on Twitter.
18:24
The gr the prompt that generated that was something like I I had at that point already set up the infra to to efficiently fetch and create a mirror of the HHS grant table but it was search this database identify any newsworthy findings transactions or
18:44
evidence of wrongdoing look carefully for um something like look carefully for rapid jumps in absolute and relative values of grants and um cross reference that with with other investigative uh reporting or or or legal findings. And what we ended up
19:02
And what we ended up finding was something that was very specific was related to these particular religious charities which claim to represent the viewpoints of of religious Americans, both Catholic and and other Christian and Jewish Americans, but really had the vast majority of their revenue, 81.
19:25
8% in one case, coming from these federal migration services grants.
19:31
So, you know, I I I know a lot of people with different perspectives.
19:33
You can have whatever perspective you want on refugee resettlement, but the organizations that are are moral arbiters for this issue should not be organizations that are actively profiting from these these federal grants.
19:48
You know, they can provide those services or they can be the the commentators, but that's a clear conflict of interest.
19:55
But the way that we got these stories was with this open-ended query.
20:00
And in many cases, the open-ended queries are um are just more effective.
20:06
They they are just better.
20:08
And it's a way that the information flows are shaping the direction of journalism, I think, for the better.
20:15
Because if I set out and I'm like, "Okay, find me fraud in, you know, find me fraud in like daycare centers ahead of time."
20:25
And that's all I say, then probably would have missed the story. It's not overtly fraud. It's technically legal.
20:31
Uh it's not it's not in daycare centers, obviously.
20:33
It's in a different it's in a different funding stream.
20:37
And um and in this case, actually having a more exhaustive and more expansive prompt at the start leads you to having a higher likelihood of finding a newsworthy story.
20:52
And that is something that is inverted from journalism in the past where in the past it would almost always be the case just because of the sheer effort involved, the sheer human quantity of time and labor that the more narrow your hypothesis, the more likely you were to get somewhere quickly.
21:14
And that is just the opposite of true at this point, at least in in our experience.
21:20
And I think that this will have a permanent reshaping on what stories end up getting published. >> Okay.
21:31
So, let me ask a question.
21:31
Imagine we gave your harness and your models to a highly ideological person and it doesn't matter what what extreme they're on. Okay?
21:44
So, they're are either a communist or they're a mega free market type, right?
21:52
and they were able to run a million queries against a government database.
21:58
Couldn't they construct an extraordinarily persuasive but misleading story using only actual factual data that they cherrypicked from those government databases?
22:14
I think that there is a much I think that that's true in theory, but I think that um I think that that's a better equilibrium than than what we have now, right?
22:25
The equilibrium that we have now is basically the um I made it the [ __ ] up equilibrium, right?
22:33
It is the um you know, all all the all these stories about like data center and and water usage.
22:41
It's like, you know, the real number is like 1% of what uh of what like or the the the real p the real amount is like 1% of total water usage. It's or less than 1%.
22:53
It's like a it's like a fraction of what like almond farmers alone use. >> Yeah. >> Yeah. Yeah.
23:01
If if you if you ask me like, okay, for Lent you have to give up everything that you use data centers for or you have to give up uh you you have to give up almonds.
23:09
What would be the harder choice?
23:12
[laughter] And and you know, it would obviously not be almonds.
23:16
Um and it was this incredibly effective campaign of not even laundered like misconstructions necessarily.
23:28
You could call them misconstructions.
23:30
You can call them mis misreadings or misrepresentations of denominators.
23:37
But really what they were was there was this one author who basically made it the [ __ ] up and was promoted by the New York Times was promoted by all these journalistic institutions who clearly had a story they wanted to tell and was grasping at straws in order to tell it.
23:55
And I think it's theor in theory possible that they could do the same thing using the tools that we have, but in some sense that would be a restraining force, not an amplifying force.
24:10
And and here's what I mean.
24:10
We end up making literally hundreds of falsifiable claims when we set out.
24:15
falsifiable claims when we set out. And actually not even falsifiable claims at the level of here is the actual amount of water a data center uses, but falsifiable claims in that we're we're saying here is this financial record that is proving what we said and uh here here is you know here's the link to that record and really if we were
24:39
just full of slop, if we were just making [ __ ] up, then you could click on those those links and you could verify that and and you could very easily dismiss SS you could very easily generate this kind of reputation obliterating um point which is like oh you claim to have found this invoice but actually
24:59
this is just completely fake and that is actually something that has happened in some cases when it when it comes to either you know defendants or um you know lawyers uh defense attorneys or attorneys in general making like chachi PT hallucinated citations or other news stories or or famously in the um Nathan Kaufnness case, the person who suspended
25:24
Nathan Kaufnness had this university speech which made up quotes by people like Albert Einstein and uh apparently this was because the that person uh used uh this is Petra desitter uh used chatbt to write the speech which which itself is a little bit suspicious I think because if you're doing that with modern chatbt will not fabricate quotes so brazenly in that way. I I almost think
25:51
I I almost think it's like some kind of um some kind of deflection, but it's this information ecosystem where it is the I made it the [ __ ] up information ecosystem and there are all sorts of historical institutional legitimacies that are basically tied to the practice of making things up at the moment.
26:16
And also to be fair, a lot of new, you know, institutions or just influencers or whatever who who are tied to making things up.
26:23
And I think that when you get to the systematic production effort of actually trying to verify and source and evaluate largecale records.
26:39
that that imposes a discipline on you that doesn't make it impossible, but that just by having this quality filter makes it less likely that that's the kind of story that you're doing.
26:54
>> You you make a very logical argument.
26:54
Uh but the the fact is we live in a world where a lot of people are very busy.
26:59
Uh they have uh their priors.
27:02
Uh and we all, myself included.
27:07
>> Yeah, I love my prior. I'm a huge fan.
27:10
>> We all love our priors.
27:10
And that turns us into confirmation bias machines where we literally are blinded to information that conflicts with our priors.
27:20
So let's just say that's the average voter, right?
27:25
If we keep this in in the political realm.
27:28
Uh so you know an unkind term would be lowinformation voters, right?
27:33
And nevertheless, you could make I think I I I'm asking you really I think you could cherrypick real facts, not I made it the [ __ ] up, but real facts that are verifiable in a let's say one of these many government data sets or private data sets or whatever and you could turn that into an incredibly persuasive story.
28:05
And look, I am 100% on board with AI.
28:10
I think it is potentially the greatest tool that has certainly been invented in my lifetime.
28:14
I've been waiting for it since I was in my 20s.
28:16
I keep journals, handwritten journals, and like I found one from when I was 22 and I didn't call it AI obviously, but uh this supercomput etc.
28:26
But every technology is dual use in my opinion.
28:33
and and and so I worry about the cognitive load that is coming for your average person and I think I I I don't know that we are adequately prepared for it because as you said uh the the idea that logically you can go back and tie it into all of the footnoted etc.
28:57
By the way, like I I'm I'm amused by most conspiracy theorists because the real conspiracy is there really is no conspiracy except in filing. >> Yeah.
29:11
A conspiracy of incompetence.
29:11
My like least favorite phrase in the entire English language is do not attribute to malice what can be attributed to incompetence.
29:19
Because if you've ever spent like a moment in the real world, you would know that the most in and the most malicious people are also extremely incompetent and very often incompetent people are also malicious.
29:32
It's not it's not mutually exclusive and and yeah, I I I think you're totally right.
29:38
You're totally right on this.
29:40
You're totally right on the human nature points.
29:43
>> And so I I think that there's going to be a lot of potholes.
29:46
there's going to be a lot of things that were, whoops, we didn't mean to do that.
29:50
Uh, but rather than, you know, subscribe to the idea, no, shut it all down, you know, make it illegal, etc.
29:58
, it's it's like the example I always give is fire.
30:00
Boy, fire was incredibly useful technology for we humans.
30:06
Uh, I've read reports that when we started cooking our food, it's what uh developed the prefrontal executive function of the prefrontal cortex.
30:16
But fire is also really [ __ ] dangerous.
30:20
And so rather than try to ban fire, what we got was fire department, fire alarms, fire extinguishers, fire uh exits, etc.
30:30
And I think that we're going to have to think out really thoughtfully.
30:36
We're gonna have to do we're gonna have to do this here as well because I could defin I don't know whether you saw the report that said that um I haven't read the full report yet so let me add that footnote but that uh they were finding in multiple studies that large language models are actually more persuasive for a majority of people than human beings are.
30:59
I you do you do not have to be a very insightful person to think, oh, what would happen if I put that in the hands of my greatest enemy, right?
31:12
>> Oh, but you you should do a double click on those reports.
31:13
I I've read very many of those papers.
31:16
It's an area that's fascinating to me, practically useful to me as well, as you can imagine.
31:19
But uh do you know how the LMS tend to be better persuaders than humans?
31:26
>> Do you know what are the the causal mechanisms that they found? >> No.
31:29
It is by like being patient, not being condescending, uh ironically empathizing with the the other person's views.
31:38
It's basically by the same epistemic technologies that we've known, you know, since the time of the since the time of the Greeks make for good persuasive rhetoric and good persuasive conversation. Yeah.
31:54
And a lot of people didn't really Ex. Exactly. Exactly.
32:00
And a lot of people just don't want to put in the effort to do that.
32:04
And that might be the frontier at the end of the day is that the LMS, no matter what your problems with them, are infinitely more patient, are by design infinitely more patient than humans are.
32:20
And that's going to come with, as you say, its own set of consequences, both positive and negative.
32:26
And and do you have any ideas about how that unfolds and and how >> Oh, yeah.
32:33
I've been thinking about this for for so long. >> Enlighten me. Enlighten me. >> Yeah.
32:39
I remember I I remember there were these two dueling camps back when I did AI policy in DC.
32:45
I was one of the few people who were doing AI policy in favor of, you know, having more AI.
32:49
Uh this is a funny anecdote as well, but when we were deciding on the name, the or was ended up being called Alliance for the Future, but when we were deciding on the name, we decided not to put AI in the name because every single organization that had AI in the name was anti-A this amazing amazing thing.
33:06
[laughter] Um but that really gave me an insight into how policy makers were thinking about this.
33:15
And there tended to be two camps.
33:15
And there tended to be two camps. there tended to be one camp which was something like oh AI is not a big deal maybe it'll be good for GDP on net but it'll be fundamentally uninteresting and therefore we shouldn't regulate it and uh that ended up being like most most of my my allies in that situation and then there was another camp that was
33:39
something like AI is going to fundamentally change the world and therefore we must ban it and I don't know if it's because I've been always like a pathologically optimistic person, but my way of thinking about it that really solidified after leaving DC and building more with AI for myself was something like actually the apocalypse was in the past. Actually, this is like
34:01
Actually, this is like this is like reinventing postmillennialism, right?
34:07
But but actually that there has been this massive wave of antisocial and destructive behaviors and actually most of that is going to be revealed.
34:17
There was this concept that was floating around for a long time of something like an encryption jubilee.
34:26
A time in which you know we finally crack RSA.
34:29
Maybe that's with quantum computers.
34:31
Maybe that's somewhat with the assistance of AI.
34:34
And we leak everyone's encrypted messages and then everyone can be okay with tolerating each other's like strange porn tastes and Google search history and whatever, right?
34:50
And I think that there is something analogous with people's um either laziness or biased forms of argumentation or even just like we were talking about the lack of effort that many political commentators have in really trying to persuade people.
35:08
And when the LLMs become widely diffuse, I think you're going to see a wave of things that could be that that are on their surface incredibly appealing and invigorating in a sense that just get demolished on the thinnest push back.
35:29
A lot of those arguments are just going to vanish or are just going to or or or are going to become these like anti-markers of uh of intelligence or of status.
35:45
They're going to be much better socially penalized because any single person can go, you know, at Grock, is this true?
35:52
And there's a certain class of arguments wi that will just completely be evaporated by grock or by you know chat GPT or claude or any of these models probably even by you know these like you know SE 7B models these like tiny models from Google or whoever um like Gemma where even like very like not even just existing technology but technology that is well below the frontier of what we have now can just demolish these arguments.
36:24
And I think that the biggest societal change that we will undergo is unveiling all of this essentially clown behavior that has gone on for the past really centuries. >> Yeah.
36:39
And uh I I [snorts] agree that you know I I I joke with people that we've always been living in the Truman Show.
36:50
uh just now we are having the tools uh that say oh wait wait wait a minute this all seems a lot more managed than we would have thought and I think part of it came from this anomalous period and and I stress anomalous period after World War II through the ending of uh I I don't know when the FCC changed the Fair Doctrine Act but where literally in America >> who's under Reagan Right.
37:18
It was uh I think so 82 yeah the 1980s but I I definitely think that >> 87. >> Thank you.
37:32
>> It was it was second second Reagan term.
37:35
>> One of one of the uh I guess I would call it a bug.
37:38
Other people might call it a feature of human OS is the desire for certainty and a hatred of uncertainty or chaos or what have you.
37:49
And and so we we had this period between the end of the war and 1987 when in America there were three major networks that were the source of the news for the vast majority of Americans.
38:01
There were two national newspapers, the New York Times and the Wall Street Journal. And that was it.
38:08
And you think like shooting fish in a barrel, if you want something buried and you have access to those three networks andor those two newspapers, you you can manage society in a much more uh seemingly orderly way.
38:26
Now, that started to break down right when when television went and started filming the [ __ ] we were doing in Vietnam.
38:34
And that's another example.
38:36
And I'm wondering your take on this.
38:38
Is what you're doing in any way kind of analogous to what happened when we introduced reporters who, by the way, for much of Vietnam were not censored.
38:52
They were not they they were just riding along.
38:54
and the American public for the first time saw what was actually happening and they're like, "Holy [ __ ] we we don't want any part of this."
39:06
Could Do you see something similar happening with things like uh effort News?
39:11
>> Yeah, I I think it's totally revolutionary and it was actually this unfulfilled promise of the Nixon era.
39:18
The idea was sunlight is the best disinfectant.
39:20
we'd finally have transparency and that would bring a whole new wave of accountability to uh to governments.
39:27
And funny enough, we got the transparency, at least much more than we had before.
39:33
And that led to almost nothing.
39:38
It led to like the Bill Clinton sex scandal.
39:41
Um and there there was some degree that that was useful.
39:44
um in some degree um but but it led to no real systematic change and some some people would argue I think correctly uh that there there's been since then a systematic decline in accountability in all sorts of government areas and I think in part that is because infrastructure to take advantage of that transparency never really materialized.
40:12
you had all of this dependency on sources reporting and um you know once again this isn't to fully condemn sources reporting but especially if you're censoring your own stories a as an institution in order to get better sourcing that itself would be something that is a lot more worthy of condemnation where the transparency was
40:34
there but at that point the news organizations the centralized news organizations I think actually um by by that time both TV and print were so attached to um getting stories through source reporting that they're no longer able to effectively comment and really effectively counterbalance the measures that were taken in place to control information. And that only became
41:01
And that only became stronger in the um in the COVID and in in the kind of post social media platform era as well.
41:10
And I think that that is changing for for a number of reasons.
41:16
Number one is just underlying technology, right?
41:19
It's a lot easier to make these queries.
41:21
It's a lot more e it's a lot easier to just process a sheer amount of documents, many of whom will be hits, right?
41:27
Or sorry, many of whom will be misses.
41:28
uh there will be a much greater number of misses per hit, right?
41:34
For for example, if we're just looking through every grant in the HHS database, a lot of those grants would just be fully legal and not just fully legal, but you know, fully non-controversial stuff that we at least on balance, I don't know if it it's fully good as a policy goal, but on balance do not see as contradicting the the purposes of HHS.
41:56
And that costbenefit calculus has totally changed where if it's something like, oh, maybe we'll get one scoop, but we'll really have to burn through like thousands of hours and hundreds of thousands of documents of things that we will completely not get anywhere with.
42:14
Before it would just been, you know, are you kidding me?
42:16
Let's move on to the next area.
42:17
But now that can actually be investigated and actually investigated fairly diligently.
42:22
So I think the overall trend like I said my underlying theme to all of this is that yeah there are going to be problems there are going to be misuses of the technology but the overall trend is towards something better than what we had before.
42:38
>> Yeah and that brings me to the economic question I had for you.
42:42
So if in the past uh there was a big story that a news organization wanted to uh own so to speak, like the Boston newspaper that that basically broke all the stories about the child uh uh abuse rampant in uh the Catholic Church and they made a movie about it, etc. , etc.
43:06
um that you can quantify.
43:07
You can say, "Okay, we need to hire six, a dozen, however many journalists.
43:14
It's going to take them six months, a year.
43:16
Okay, we can tote up exactly what their salaries are, what their overtime is, what the cost of travel, etc. , etc."
43:24
Um, but you have costs too, right?
43:27
In other words, you have costs of inference, engineers, uh, a variety of things.
43:32
How How cost advantaged are you over the model I've just described?
43:41
>> The stories that we've broken so far, whether it's the preferred communities grant, whether it's um the the degrowth donors, all all of them are stories that would have just been straightforwardly infeasible under the previous model, which would have if you were to put a dollar number on it.
43:56
And and it would probably not even be feasible to just recruit that many journalists.
44:00
it would be this kind of market that eats the entire market uh and eats the entire old market.
44:08
It it would have been like upwards of tens of millions of dollars at least if you could even uh if you could even like physically assemble like find that many journalists who who had those competencies in those areas to pay.
44:26
So it is a total sea change in terms of what is capable and what is cost.
44:32
Speaking of cost though, I this is actually more of my traditional educational background.
44:37
I had a degree in pure math and started in machine learning at a really young age basically when I was still in high school.
44:44
But anyways, I I had been a machine learning engineer before and this is mostly what we optimize on because a lot of the harnesses make extremely inefficient queries to databases.
45:00
It's shocking because at the same time I can look at, you know, I can look at some of the cyber security applications.
45:07
I can tell that like there are a lot of things that the LMS are a lot uh better than me on.
45:11
But in terms of making API requests, not all of them, but but many of them or making efficient database queries, they just will put out tons of [ __ ] slop queries and will consume way more tokens, will just take longer, right?
45:28
We'll have higher latency.
45:28
uh on on these tool calls and there is just a lot that you could optimize for in many dimensions.
45:37
That's part of what's so exciting about this and that's been super exciting for me.
45:41
It's it it brings me it kind of like points to that environment that a lot of people talk about this like actually before my generation but of of the early internet where it seemed like everything was so unoptimized.
45:54
You could just sit down and think about it and you could optimize these things.
45:58
you you as basically anyone, right, who was reasonably smart, who who had a computer, was reasonably familiar with these things, and that's just this massively inspiring thing.
46:08
And that's something I feel like we're picking up a ton of momentum on every single day on our team.
46:16
>> So, you know, what about the idea that journalism has been mclassified for a century?
46:23
And by that I mean like we always think of journal I think of HL Mein, right?
46:28
He was the king of the muck rakers and he was always great with his quotes and everything else.
46:32
Uh but I I want so at least I maybe others don't but I sort of associate journalists with writers with you know uh the the actual act of writing or producing if it's TV or radio but m maybe maybe they were actually search and verification engines and using writers was just a hell of a lot cheaper because we didn't have the technology.
47:03
Um, your model changes the economics of finding needles in haststacks, right?
47:11
But >> yes, that was one thing that we thought of putting on the website.
47:12
We find needles in haststacks.
47:14
And but my my question and push back on that is like if I was your opposition, right, if I was trying to get up to something nefarious, why wouldn't I, if I have my own tools that are at least as good as, if not better than yours, why wouldn't I just make a [ __ ] ton more hay and make it much more difficult to find the the true needle, so to speak?
47:41
>> Oh, it's already happening.
47:41
Uh yeah, I I think that that's totally fair and that's totally happening and actually that's been happening for 30 years.
47:46
That that is the intelligence agency um um defense department that that is the documented strategy uh of just producing so many documents of maliciously complying with foyer requests.
48:00
Uh that that's totally happened and I think there will be another escalation of this.
48:06
Like imagine you make a a foyer request to to the CIA and they just give you like pabytes of AI slop.
48:13
That could totally happen.
48:16
Uh and I think that probably will happen.
48:18
So um I I don't really have a solution to that necessarily.
48:23
To to some extent we will have to AI harder and it is this red queen's game where you're uh where you're uh staying in the same place but you're running faster than faster.
48:32
I I think in some sense we can out compete the CIA at least in this narrow uh area or we can we can out compete all of the people who would otherwise be trying to offiscate information and in part that's just the pace of technological development.
48:50
Of course I think a lot about incentivizing our organization and building up our capacities in that way.
48:55
In part that's innovative dilemma.
48:57
You could go, you know, there there's this in there are all these tech theories and why, you know, why the incumbents will always win or why the startups will always beat the incumbents.
49:07
It's always treated as this kind of predetermined natural course of history thing.
49:11
I think we just we just have to win.
49:13
I I think we just have to fight as hard as we can.
49:15
And I think you're absolutely right that it's not just going to be us.
49:19
It's going to be people on both sides.
49:20
It's going to be people both maliciously complying, maliciously producing information, subverting transparency laws, um or or at least kind of like maliciously complying with transparency laws.
49:30
That's totally going to be happening.
49:32
And at the same time, we'll just have to race race on faster and faster. >> Yeah.
49:38
But again, I I I come back to good old human OS and I don't know anyone who reads the terms and conditions uh before clicking. Okay.
49:46
[laughter] And I agree with you, by the way, that the patient large language models will read all of them and they will point out all of the ones that go against you.
49:57
Um, but I I'm I'm I'm worried about the people, you know, they're busy with other things in their life, right?
50:11
They've got a family, they've got a job, they've got, you know, hobbies, etc.
50:13
And [laughter] meanwhile, I kind of see this massive battle between two philosophical po points of view if we keep the extremes right and and the weapons are getting a lot more uh precise and and I just wonder and it's it's like again I the Pandora's box is open and so trying trying to say no no we we we we want to rethink that that's not going to happen.
50:48
But I do wonder if we should spend a little bit more time on, you know, cog sec cognitive security, right?
50:54
Uh like m maybe uh freely disseminated if we could figure out a way to really be able to judge deep fakes, for example.
51:07
Wouldn't it be cool if you just freely distributed to everyone something on their phone that was like the minute a video showed up and it just went deep fake across it?
51:15
At least that would condition them to start thinking, "Oh, you're absolutely right about history."
51:25
But people have been manipulated by religions, by political ideologies, by a variety of incentives for all of human history. Like that's nothing new.
51:34
But now we're doing it at weapons grade and it scales.
51:39
And so I definitely think some form of cognitive security for everyone is might be a good idea.
51:50
>> Yeah, you're definitely the best kind of critic, Jim, because I I I think I really share the same like fundamental fundamental view of human nature.
51:59
I have a very Hobbesian view of human nature.
52:04
Hobbes is still my favorite philosopher to this day.
52:07
>> That's brutish and short, right? [laughter] >> Yeah.
52:10
And so much more than that.
52:10
You know, there were both the the dual powers of the sovereign, the origins of human reason goes so far.
52:16
But um I think you're right in that I I think that you want to draw a distinction between the scale of deployment where you're absolutely right.
52:37
It's it's bigger than ever before and the equilibrium harms.
52:40
and the equilibrium harms. I think something that every major tech policy fight has sort of gotten wrong in the past few decades has been this conflation of the possible harms and then the equilibrium where people will say this for example with cyber that's something that's more played out people will say oh there are so many potential hacks And then you ask the straightforward
53:15
next question which is well how many of those potential hacks will be prevented by AI and not only that how many of the potential hacks from past generations or you know just just possible with previous levels of software or previous
53:29
levels of uh human human hackers will be uh prevented by AI and then that equilibrium that's not zero right that's not zero harm but it's very very different And man, this could really be a long long answer. I I I will try to keep it from
53:47
I I I will try to keep it from being like a 30 minute hobbs monologue. >> Yeah. Yeah.
53:52
No, obviously we let's try to uh make it a little more compressed, but I think this is important.
53:57
And I think that uh I think that people who are in favor of AI and all of the tools and the powers that they unlock need to examine and acknowledge the other side of the argument.
54:12
Don't get me wrong, I I'm maybe one of the most pro- AAI guys out there in terms of the potential for what it can do.
54:20
We just made a movie uh about how uh the Vuvius, you know, the uh volcano that destroyed Pompei uh is the schools. >> Yeah. Yeah.
54:33
And >> we made a movie about it because to me it's the coolest thing in the world.
54:39
It required a particle accelerator and AI to be able to translate these scrolls.
54:46
But now we're able to transcribe ancient Greek and Roman scrolls that if you touched them in the past, they just disintegrated.
54:55
And like I'm giddy about the use cases for AI in terms of expanding our knowledge, expanding innovation, expanding discovery, etc.
55:08
But I I definitely think that you need to be I I call myself a a pragmatic optimist.
55:16
And by that I mean I think you need to consider the unintended consequences and try to come up with good solutions for them because if you don't the people who are on the other side of the argument are going to make hay with those and they're going to say yeah but right like and most of most of the objections are going to be you know way deep in the tail and very unlikely to happen.
55:41
But if we who are in favor of this new expansive rocket ship for the mind that AI is and and also I hate the artificial intelligence I Brian Romeli calls it intelligence amplification which I much prefer um because really that's in my opinion one of the things it's doing andor uncovering right like you're doing um but I I do think we need to think about these other use cases and have good ideas because it's going to happen. >> Oh, absolutely.
56:15
It's it's totally going to happen.
56:17
And I think one of the most important questions for any philosopher to ask is to step back and ask, is this a matter that will be decided by philosophy or is this a matter that will be decided by action?
56:35
And to a large extent I think that the part that has been decided by philosophy we've already talked about it's kind of the nature of the story and and and that kind of ep epistemic field of what comes and the part that will be decided by action is essentially what are the names in the media ecosystem people will trust.
57:03
what are the solid ground, right?
57:06
And and that can never be fully reliable.
57:09
You know, you shouldn't put your absolute faith in any kind of news reporting or or or any kind of, you know, human institution in general.
57:16
But, you know, you you have to you have to have something.
57:22
And I think about a lot of that on a daily basis as well where effort is not merely trying to produce a good like technological platform, right?
57:38
We're not merely trying to say like, oh, here is this technology.
57:41
Everyone will have this technology. Uh voila.
57:46
I'm really concerned and I think this is very different than a past generation of more optimistic mo more maybe like rousoian founders even I don't know if they're they're like explicitly rousoanian but like implicitly rousoian but I'm a lot more concerned about institution building that this is the specific group of people who will be employees and founders ers and contributors to effort.
58:16
This is a specific kind of culture that I want to calibrate on uh epistemic openness and of hypothesis testing and of the scientific method and of just a very high standard for everything. Not publishing slop.
58:31
That's the most important thing.
58:33
We do not publish slop and we do not publish slop in all sorts of ways.
58:36
We do not publish slop in that we do not publish, you know, AI writing.
58:40
At least at this point writing still, at least in our experience, not great.
58:42
uh we we we we we rigorously like hand hand write um we we human write all all the articles and kind of thoroughly check for any improvements we could make to the phrasing any improvements we could make to the clarity.
58:59
We also of course don't publish slop in terms of records.
59:02
We verify all of the records um both with a lot of AI tools and by hand.
59:05
We don't publish slop in the sense that we um we we don't bury stories, but we want to make sure we take our stories to completion.
59:16
We don't want to just be like wildly flinging accusations.
59:18
We want to get to the point where we have a clear border of this is what we found.
59:26
This is what we didn't find.
59:26
Here is how we can prove what we found.
59:28
Here is, you know, the steps that led to us not finding these other things.
59:34
And I think that at the end of the day, part of that will be decided by technology.
59:40
Of course, we would not be able to do most of what we're doing now without it.
59:45
But there's also a large part that will inevitably come down to a form of human judgment.
59:50
And that is something in my opinion that you have to be proactive about.
59:56
You can't just completely leave the human judgment to the wind.
1:00:00
that you have to be proactive about as we're building out effort.
1:00:02
And to me, in the long term, there is no perfectly, you know, philosophically consistent defense against all bad actors.
1:00:14
That's just not feasible.
1:00:14
Maybe maybe there's a solution to it.
1:00:17
Maybe, you know, GPT7 will will come up with one, but it is just beyond me and it's beyond every philosopher in history.
1:00:24
But it the the practical solution to that looks more like the institution building that I think a lot of people in the previous generation have neglected. >> Yeah.
1:00:35
And that that's an interesting point because it also leads to objectivity.
1:00:39
I I think that that's a term that we like to use a lot and misuse mostly uh because I I I I think that um your organization for example, right?
1:00:56
Like I would love to see that you don't get cap because you're a subscription model, right?
1:01:00
Um, there's always audience capture and like are you are you dedicated to exposing whatever kind of corruption andor uh story that got buried regardless of which political side it helps. >> Yeah, for sure.
1:01:20
And I think that at the very beginning you're going to see actually you're going to see a different kind of selection bias which is that we are reporting on essentially stories have been that have been dropped because for example the uh refugee resettlement story that's something that could have been reported on for literally five years.
1:01:42
So the fact that we were able to report it is a direct consequence of it not having been reported.
1:01:48
And honestly, like it's it's funny because this is going to be a a justification for stories that are perceived as more um rightcoded.
1:01:57
Um but honestly, it's because a lot of the right-wing media ecosystem are just a bunch of muppets who do not seriously complete their investigations and are are kind of comfort are comfortable with just like publishing slo publishing accusations, not actually digging into the documents.
1:02:14
And I mean digging into the documents not be completely easy especially without AI but it is just this completely different equilibrium where you know there are obviously biases with the New York Times but they do actually do investigative reporting at least and they do actually break stories and they break stories in a certain direction and that's going to leave actually a gap that is not going to be equal.
1:02:43
So I think that in the short to medium term you should expect that to be unequal via negativa.
1:02:49
Uh let me give a more concrete example um related to that though.
1:02:56
Um we're we're going to be breaking a story.
1:02:58
I don't know exactly when this will come out but we will be breaking a stories within a few days of reporting on something that the Trump administration has been trying to keep secret which is the whole um freedom fuel uh network. Do you know about this?
1:03:14
Do you know about Freedom Fuel? >> No, I don't.
1:03:17
>> It's this it's this fascinating mystery.
1:03:19
I think it's been this fascination of journalism world because it's this like particularly enticing thing.
1:03:26
Essentially, the Trump administration has been promoting this series of uh of gas stations in New Jersey and Pennsylvania that have been selling gas almost certainly at a loss.
1:03:38
They originally sold at 347. Right.
1:03:40
This is this is because Trump is the 47th president and they've in general been somewhere like 40 to 50% below the going gas rates in those areas almost certainly losing money compared to selling those uh selling that gas on uh even on the open market.
1:04:03
>> So there's this fascinating mystery. >> Yeah.
1:04:06
And economists would call that dumping to gain market. >> Yeah. Yeah.
1:04:09
And so there's this fascinating mystery of where's the gas coming from?
1:04:15
>> And yeah, and all the king's horses and all the king's men have tried to figure out where the gas is coming from and they have not figured it out.
1:04:23
And we have not I I I think we have meaningfully moved forward the freedom fuel case. I'll put it that way.
1:04:32
We we we've not fully found evidence of government subsidy, but we have found fairly conclusive evidence of self-deing. >> Interesting.
1:04:44
Um, of the stories that you've dropped so far, which one got the most uh reaction, both positive or negative?
1:04:55
And then I do want you uh to because when I was getting ready to chat with you, I read your story about how how to uh hijack the new X uh or as us old-timers call it, Twitter uh algorithm.
1:05:11
Uh but whi which >> am I an old-timer? Oh [laughter] man. >> Maybe in spirit.
1:05:19
Maybe >> learning a lot today.
1:05:20
[laughter] But but which which stories uh are the ones that like really the lightning hit the rod? >> Yeah.
1:05:31
Uh it it was definitely the um it was definitely the uh refugee resettlement story because this one was just something where the financial record was so damning and also where the players involved were these already like very big media entities.
1:05:47
US Conference of Catholic Bish Bishops, Church World Service, um, HASS, um, Episcopal Migration Ministries for a long time, they've been trotted out.
1:06:00
They are the most public of public figures or public entities, they've been trotted out to be very critical of any kind of migration restrictions.
1:06:10
And for literally for you know upwards of 5 years in terms of the grants that I reported on but upwards of 10 years at least for the total body of grants it had just gone completely unnoticed that they were being majority that the vast majority of their revenue 81.
1:06:28
8% 8% in the case of uh US uh CCB um 89% in the case of church world service was being funded by these resettlement grants.
1:06:38
In other words, they were explicitly benefiting from the policies they were advocating for.
1:06:46
It would be like taking you know a dairy farmer and having them be the object treating them as the objective moral arbiter of whether the government should subsidize milk.
1:06:56
It's like, you know, you you bring these these there are these like well-known names who go on TV every time there's like a milk policy debate and they're treated as the representatives of Americans of of like Christian Americans worldwide or or Christian Americans, you know, across the country and they come out and they say, "We morally and condemn in the strongest terms the president's opposition to subsidizing milk."
1:07:25
And then it turns out like five years later that all this time they're selling milk.
1:07:30
It's like [laughter] what?
1:07:34
This is a totally crazy story.
1:07:34
And it also goes to show like despite all of the media attention, despite the endless commentary on these organizations, for a long time, no one went and actually dug into the financials. And now we did. We we finally done it.
1:07:52
And yeah, the the reaction to that was overwhelming.
1:07:56
I think actually it was interestingly there's a certain type of story.
1:08:01
There's a certain type of I mean I I don't want to be mean to any individual specifically but but you know as as an institution there's a certain type of New York Post story that very effectively draws uh kind of counterciticism that breaks a real thing but also taints it in a way and fills it with like other sort of like inaccuracies or bias or basically gives this like massive attack surface to to critics of that story.
1:08:31
And of course, the most famous one of these was the Hunter Biden laptop, which at its core was a truthful story.
1:08:37
It was a real finding, but was distorted in all these ways, which eventually became used in order to, you know, weaponize um social media against it.
1:08:48
Anyways, uh we we got very little of that with the um with the uh uh refugee resettlement story, I think, actually for that reason.
1:08:59
And you know there there's actually a perverse incentive there.
1:09:03
I'm going to do my best to avoid it.
1:09:06
I think that we shouldn't be like lowering the quality of our story intentionally in order to get more kind of counterreaction from the other side.
1:09:14
But I think there's a certain type of story that gets a ton of negative counter reaction because it like I don't know if it intentionally uh introduces flaws but you know because of the flaws that are there.
1:09:24
And I think that this was sort of the opposite part kind of story where the financials were so clearly laid out, the conflict of interest was so undeniably established that the other side didn't even want to touch the story. >> Interesting.
1:09:41
And you we're we're talking a lot about incentives, right?
1:09:44
And and another point of view that is kind of consistent with my quip about we're all living in the Truman Show and always have been uh is like what percentage of in quotes news is just propaganda uh and it's literally the various sides vying for power trying to win the the news cycle.
1:10:12
Like that's another one like winning the news cycle that that's as old as we have the media, right?
1:10:17
And and I just wonder one of the things that I do worry about is all of this is so new and I'm not just talking about AI.
1:10:28
I'm talking about Bill Bryson has a great book called America One Summer in which he's talking about, you know, the first time a single human spoke to more than, you know, a thousand humans, right, was the guy on the radio announcing Lindberg's return from flying across the Atlantic.
1:10:51
And and so telegraphs like it it's in the scheme of human history.
1:10:55
All of these new technologies are brand spanking new and now we've got like you know we've gone from spears to thermonuclear multiple warheads re really quickly.
1:11:14
Is there going to be some form of way that that humans just like >> Yes, I I think it's all resolving now.
1:11:21
I think it's all not not resolving to a final conclusion, but a lot of that builtup technical debt is now being re-evaluated.
1:11:31
And that's the conflict we see.
1:11:33
That's what people are angry on on social media 24/7.
1:11:35
And not all of that is kind of directed in a productive way.
1:11:40
A lot of that is directed towards slop.
1:11:42
Those criticisms are totally fair, but largely what I see is a re-evaluation of those fundamentals.
1:11:47
You know, as Nietze called it, a re-evaluation of all values.
1:11:51
And I think where that starts is something I call the crisis of nominalism.
1:11:56
There's this big phrase that goes around, you've caught me at like a great time because I was just on like a 4-hour hike and I was like thinking about this and I have all these thoughts fresh in my mind.
1:12:06
fresh in my mind. But there's what's called a crisis of liberalism which I think is really mislabeled and the theory behind calling it a crisis of liberalism is like oh there there's a ton of premises but the story is like oh we've lived in what we now call liberalism for roughly 250 years
1:12:25
basically since the American founding and that is now coming to various flaws whether that's fertility whether that's social media whether that's foreign influence whether that's uh migration And uh you know what people will say is that liberalism is collapsing o under its own values. And I think a
1:12:42
And I think a pointby-point deconstruction of that claim almost every premise is false where we are not living under we're no longer living under liberalism.
1:12:51
It's not a continuation of older values.
1:12:54
Um liberalism is not causing the issues that we see.
1:12:58
And actually it is not even a crisis.
1:13:01
It is re a re-evaluation of a slow decline.
1:13:04
And I can get into each of those points specifically, but but let me try to give a core kind of like something that's like a useful takeaway from this is that you know nominalism what it actually is is people taking a name for granted, right? This is science.
1:13:22
It was published in a title in a journal called science.
1:13:24
Therefore, it must be science, right?
1:13:27
And not actually going looking underneath is this the scientific method.
1:13:31
Are these um neutally tested hypotheses?
1:13:35
Were these in a controlled environment?
1:13:37
Were were these you know what was the statistical probabilities that this was actually a relevant finding and so on and so forth that at the very surface level you kind of look at the cover of the box and you never actually look inside the box.
1:13:51
That problem has persisted.
1:13:53
It has persisted certainly scientifically.
1:13:56
It's persisted in uh news media in in both the older institutions and the kind of influencer slop type media that's come up um to to rival it.
1:14:09
Both of those domains are domains where you kind of just get the slop and you never [clears throat] look on inside the box.
1:14:14
You get that and this is really the most uh disastrous version in spending and you get this in [clears throat] a lot of government spending.
1:14:25
lot of government spending. I'm I'm very sympathetic with the with the Doge people where they had a complete communication failure and they they they just got kind of so internally overwhelmed that they ended up kind of
1:14:41
being driven crazy and not really being productive in in their political actions where if you are someone who who's had experience putting together financials for a startup I I've done this both in my own company and with uh previous companies that I've worked for. I I've
1:14:57
I I've done this personally and you are used to that level of accounting and you know that if you do not present this quality of information to auditors, you will go to jail and you see the standard that is set in Treasury or or or in in these other government departments then you will have a meltdown.
1:15:18
It is just a completely different standard.
1:15:20
is a total double standard.
1:15:21
And they were expecting to go to the media or to go to Congress with this and for them to react with, oh my goodness, this is a total this is like a five alarm fire.
1:15:31
This is something that, you know, if you were done at a private company, you would go to jail for.
1:15:36
This is this is some crazy stuff.
1:15:38
And they were completely not prepared for the reaction that they got, which is this is how things have always been.
1:15:44
And you you know both of these sides are kind of act acting rationally in their own narrow context.
1:15:49
It is simultaneously true that if you were doing this in a private company or or public company if you're doing this in you know um any kind of normal corporation you would go to jail for this.
1:16:01
It's that that's both simultaneously true and the kind of protests of the senators of this is how it's always been.
1:16:07
This is maybe not good but there's not much we can do about it individually.
1:16:12
I think that's also true.
1:16:14
So it was this huge kind of Shakespearean tragedy that played out.
1:16:18
But that's a good illustration of what I mean by the crisis of nominalism that we have these names, right?
1:16:23
We have the these names of accountability or auditing or or you know financial recordkeeping and under the surface they've completely diverged.
1:16:30
Under the surface you actually look inside the box and it's like what is even going on?
1:16:35
It is this like cathonic mass and a lot of that is going to be re-evaluated.
1:16:42
A lot of that is going to be either split into different names or there is going to be this type of philosophical colonialism where people will go in essentially like people who have been more competent over the past 10 or 20 or 30 years will go in and have to clean up these these institutions.
1:17:02
Part of that is also that this new trend of software rollups which could be a completely different rabbit hole.
1:17:09
But there's a lot that's going to shift and that's going to hinge both economically, politically, and philosophically on this crisis of nomalism of the big question of what happens when you finally look inside the box.
1:17:23
>> Well, you you make uh now now we're down to linguistics in my opinion.
1:17:28
Uh so for example uh I think it was Kirkagard said when you label me you negate me and and you were using uh words like accountability etc.
1:17:40
Of course good old George Orwell understood that long time ago when he talked about newsspeak and you know war is peace, freedom is slavery etc.
1:17:52
And you know and then of course the classic Bill Clinton under questioning during the monarch.
1:17:59
>> What do you mean is >> what? Yeah.
1:18:00
Well, it depends on what you mean by the definition of is is right.
1:18:06
But you yourself, you you've said slop I don't know how many times. Do you see? We've always had slop.
1:18:12
We except it was always just human slop, right? Of of the millions. >> Yes.
1:18:20
I I I when when I say slop, I am inclusive.
1:18:22
I am, you know, not discriminating based on species. >> I I understand.
1:18:28
But but I want to say we get captured.
1:18:30
One of the dangers of labels in my opinion is it puts thinking aside, right?
1:18:38
You label something slop.
1:18:42
It goes into that bucket. We call it slop.
1:18:44
You you you no longer really try to disambiguinate.
1:18:51
Is this really slop or is this really I mean like would would Joyce's early works get put in a slop bucket? Probably.
1:18:59
You know, Finnegan's Wake and Ulisses and it was because he was so new, right?
1:19:04
His style of writing was so new people literally didn't know how to read it, right?
1:19:11
And now we know how to read it now because we evolved into understanding that kind of post-modern way that Joyce approached his novels.
1:19:23
But I I just think that I I worry too because labels, the power of labels, you you yourself have just made an excellent case, right?
1:19:34
accountability, transparency.
1:19:36
Like when when I hear transparency and accountability, I think, okay, wow, it's going to be accountable.
1:19:43
This is going to be transparent.
1:19:44
Uh and and so the power of language in and of itself, the semantic meaning, the semiodic meaning is something that we're going to also have to like kind of dig into.
1:19:55
One of the ideas I had would be how cool would it be to have like an AI companion that was I'm calling it the reader, right?
1:20:05
And when you're reading a news article, it just points out who like factually incorrect uh appeal to emotion.
1:20:14
Uh you know, I don't know if you've read influenced by Chelini.
1:20:20
We talked about it a little earlier.
1:20:22
So, and and I joked one of the reasons why the large language models were pretty good at it was because they've been trained on the entire uh uh corpus of influence and Chelini is kind of the king with his book influence.
1:20:37
Um, but I I definitely think if you had a reader there that was saying like who benefits from this?
1:20:46
I I used to say to younger people when they're asking me like how do you like navigate reading the news this preI was just like ask yourself the question who benefits if I believe this um what are the objectives of this particular writer?
1:21:04
Uh are they for a particular outcome?
1:21:07
Are they just informative? What are they?
1:21:10
But if you actually had a com reading companion right next to you or a watching companion, right?
1:21:15
if you're I mean I if we had this I' I'd put it on this podcast and it would probably contradict me a lot.
1:21:22
Well, Jim is doing an appeal to emotion.
1:21:26
Jim is doing an appeal to, you know, authority, all of those things.
1:21:31
But I definitely think that like organizations like yours, that's why I was so excited to talk to you are you're you're the beginning stages of this unfolding.
1:21:45
Yeah, I I was just so when you talked about Joyce and uh his early works, I immediately flashed back to this experience I had two months ago.
1:21:57
So, so I'm going to read something and uh tell me what you hear. Okay.
1:22:04
The old pressure professor Mdash Calguz by name Mdash aimed his glass at one of the ships still lit by the sun, then patiently focused the lens until the image was as sharp as he could make it.
1:22:17
Like a scientist over his microscope, peering in to find his culture swarming with the microbes he knew all the time must be there.
1:22:25
The ship was a steamer, a good 60 years old.
1:22:28
Her five sacks straight up like pipes showed how old she was.
1:22:31
Four of them were locked off at different levels by time, by rust, by lack of care, by chance.
1:22:39
Mdash, in short, by gradual decay.
1:22:42
She had run around just off the she had run a ground just off the beach and lay there, listing at 10° like all the ships in the Phantom Fleet.
1:22:52
There wasn't a light to be seen on her once it was dark, not even a glimmer.
1:22:57
Everything must have gone dead.
1:22:57
M Dash, boilers, generators, everything, all at once.
1:23:01
M Dash as she ran to meet her self-imposed disaster.
1:23:04
This is from the second page, I believe, of Camp of the Saints.
1:23:09
>> And I was going to read this.
1:23:09
I was going to reread this >> and I just couldn't do it.
1:23:16
>> It was too AI coded for me. It just broke my brain.
1:23:19
I know it's obviously not AI, but it just like totally broke my brain.
1:23:23
>> I was It's so funny, serendipitous.
1:23:23
I was just rereading or trying to reread that book for people who are not familiar with in the camp of the saints.
1:23:35
Uh it was a book written I think when in 1970s right uh >> yeah a long time ago >> and it envisions the very migration problems that we are having now done differently.
1:23:48
I if my memory serves I I read it back then and was horrified and I tried to I I picked it up again and like you I'm like wow uh but it I had a different reaction than you.
1:24:01
I I my reaction was why do large language models?
1:24:06
And if you look at my Twitter account, you'll see that I have used M dashes copiously my entire I've written four non-fiction books obviously all before AI.
1:24:18
You're going to find a lot of M dashes in them because I love them.
1:24:21
And and so on Twitter I say you can come and pull my M dash out of my cold dead hands.
1:24:29
But isn't it fascinating, [laughter] right?
1:24:32
It's it's like uh you know the the uh the same with the it's not X it's Y.
1:24:39
Go back and read mystery novels and things.
1:24:42
The reason they use that technique and the reason they use M dashes is because of the training literature they were trained on.
1:24:50
We use M dashes all the time previously and you know you you have uh uh the Orient Express like it's all it he wasn't this he was this you know when she's describing the villain and and so it's another part of that axiomatic reaction right like >> I find I want to add a little nuance to that >> because I think that the these are actually different cases.
1:25:24
actually different cases. I think for m dashes you're mostly right on that is just kind of part of a training data for the it's not this it's that that's more downstream I think of a particular technique called trans called um oh my goodness just like
1:25:46
struggling to pronounce this correctly um called contrastive learning there we go called contrastive learning where the idea was that the models would have a better directional update if they were given both positive and negative examples of something. And this was a specific
1:26:06
And this was a specific technique published in papers by OpenAI, by Google, all the biggest names um that specifically uh used this format and was actually effective at least in the short term for for post-training their models.
1:26:23
So I I think that those two those two cases are actually actually have slightly different origins.
1:26:30
But I think the underlying idea, >> you know, you shouldn't necessarily dismiss the these uh these literary patterns that they're legitimate literary patterns.
1:26:40
I think the underlying idea is right. >> Yeah.
1:26:42
I think if you are an Agatha Christie fan, that is who I was referring to when I talked about Murder on the Orient Express. >> She does that.
1:26:50
It's not this, it's this all throughout her books, right?
1:26:53
And because she's wants to surprise the reader, right?
1:26:58
You thought that he was a good guy.
1:27:01
No, he's actually the villain.
1:27:05
You know, you thought he was steadfast.
1:27:07
You thought he was noble. No, he was the villain.
1:27:11
[laughter] And and so I I just think that there's a lot of ignorance uh about the idea that um the the models do what they do. you ultimately right. Agria's trillemma.
1:27:28
You familiar with a grippus trillemma? >> No, I'm not.
1:27:33
>> Ah, well that would take us forever.
1:27:33
But the the the short version of he he he destroys logical systems because he ultimately and Lewis Carol a mathematician and logician actually wrote a book uh not a book he wrote a piece called what the tortoise said to Achilles and it made the case even better than a grippa trillemma but basically it's like every time you have a logical rule you got to say well what why Why do why why why do we have that rule?
1:28:05
And then you keep going down down down down down and you finally get to that base rule where you said earlier in our chat, I just made it the [ __ ] up.
1:28:15
[laughter] But in in in Carol's what the tortoise said to Achilles is ultimately everything is a human choice or a decision at least now, right? This may change, right?
1:28:31
going forward and you mentioned chat GPT7 or 10 or 12 or whatever but for now and historically a grius trillemma is the problem of you know infinite regress of there's a let's it would take us an entire different podcast to talk about a grippus trillemma but I think you would definitely be interested in it.
1:28:57
>> Yeah, I'm I'm reading it now.
1:28:57
Yeah, it is this like it reminds me a ton of like formalized set theory. This would be okay.
1:29:06
Like it would be very it would not be entirely surprising to me, but it would be like a life-changing moment, I think, if um if the LMS discovered a a reduction in the set theory axioms.
1:29:27
That that would be that would be a moment that I'm here for.
1:29:31
It would be like my personal 911 is when [laughter] is when AI discovers a way to reduce the set theory the set theory axioms. >> Yeah.
1:29:40
Well, I I I mean again I on because I am very optimistic about the even knowing that we're going to have a lot of bad with the good.
1:29:49
I definitely think that the power and discovery that uh AI uh allows us to use as tools.
1:30:00
That's the other thing that drives me a little crazy, right?
1:30:03
like the on I I was looking up the thing on a on our system which we've trained on a bunch of different things and you had you had said like 80 to 100% is uh is of news is source going to the newspaper.
1:30:23
Well, it it it gave me the whole history and it was like, yeah, media scholar Jim McNamera reviewed 150 to 200 uh studies and he found that up to 80% of media content is sourced from or significantly influenced by PR with estimates of 50 to 75% common. But then it keeps going.
1:30:48
It goes back to Herbert Gan's classic 1979 content analysis and it gives me all of the notes that I can actually go and go to the actual book or study and find.
1:30:59
And I definitely think that for a certain type of person, the ineterate rabbit hole diver that is really, really cool.
1:31:10
My question is, is it cool for non- nerd for like I'm totally willing to call myself a nerd, right?
1:31:18
And I love rabbit holes and I love doing all that, but and I would guess that you would be coded nerd. Yeah, >> sure. Yeah.
1:31:27
[laughter] >> I But I just wonder how much of a difference it's going to make to the non- nerds.
1:31:33
In other words, you you I I think of the flatearthers.
1:31:42
If you if you could if if you're if you could do a news story based on all of the data uh and convince the flatearthers, right?
1:31:53
That would be of interest to me.
1:31:53
There are again we come back to kind of the scientific method, right?
1:31:58
If if it's not falsifiable, it's a religion in my opinion, right?
1:32:04
uh it's a belief system and and what I've seen historically at least and I don't mean to pick just on the flat-earthers uh there there are a lot of cult-like beliefs that are patently untrue but when confronted with data and I you know I made my career in asset management on doing algorithmic investing using data um so I'm a big believer but one of the things you see is here is Mr.
1:32:37
flatearther, the million reports showing you why you're wrong.
1:32:42
And what happens is in many cases they double down on the belief and they're like, "Yeah, no, that you know, even though you you can show every footnote, every study, everything.
1:32:57
No, that's that's all wrong." Aggressive I don't know.
1:33:01
I don't want to say stupidity, but a aggressive disbelief.
1:33:06
Do you think you ever change those minds?
1:33:08
>> I think that on operational grounds, on anything where it's like, oh, you actually have something at stake, you can change those minds.
1:33:15
I think when it when it's this kind of like when it's this Truman show thing and it's pretty clear that like, you know, if you think the Earth is flat and the entire system of uh of, you know, physics that's been developed around a circular Earth is wrong, like do you avoid flights?
1:33:35
>> [laughter] >> you know like a lot of the time the answer is no and uh the ones who the ones who who the answer is yes actually I would probably respect more um because of course that entire system is based off of an assumption of the flat earth uh not an unfounded assumption of of course but you know that all comes along with the system in play and so you have this really like fake virtualized game of debating about the flat earth.
1:34:08
And I think in that environment, you're never really going to get anywhere with most of the time.
1:34:17
And you know, maybe the AIS will have enough patience to do that and eventually win someone over.
1:34:21
But in all practical grounds, that is actually not a particularly effective use of our time, right?
1:34:32
Like the the flatearther is not actually particularly harmful to anyone and in a lot of cases can go about their own life without that cogn cognitive dissonance, right?
1:34:43
They'll they'll go on the flights.
1:34:44
They'll, you know, talk to their round earth colleagues.
1:34:46
They'll in a lot of cases actually work a stable job and so on.
1:34:51
And that's like kind that's like basically fine.
1:34:56
And actually, I think there will need to be a slow political reorienting around kind of what what Leo Stros calls the the the cave below Plato's cave, right?
1:35:12
that that we kind of have all this scientism.
1:35:13
We we've now gone to the extent of applying the scientific method to all sorts of ideas and and concepts and philosophies and and all kinds of new moral rights that are not the ground of the scientific process, are not falsifiable, are not things that you can decide via microscope and are slowly being unrolled and seen as such.
1:35:41
And I think that political process involves um actually it's interesting this this is kind of a I think it's a tension but not necessarily a contradiction.
1:35:55
You know we we can have the um philosophical maturity to to to realize that there are forces acting on the world that act in opposite directions at once.
1:36:03
opposite directions at once. you know or or you know that that you know the different forces will act in different directions even if they're acting at the same time but I think there's both a tendency towards developing greater standards for truth as well as actually an admitting that there there are these issues that have been treated as truth
1:36:27
or is treated as scientific that are really not scientific at all and are rather more moral or or or theological issues And I I I think to to add one more thing on this, I think that there's a big unraveling that goes back to Mill, who I think was actually a traitor to classical liberalism by redefining the goal of classical liberalism around consensus. This was the classical Mills trident,
1:36:52
This was the classical Mills trident, right?
1:36:55
Which was an argument for free speech that I think ultimately undermines free speech when it comes to the modern day.
1:37:01
And the idea was okay if you think that uh if you think that what someone is saying is uh is wrong then you they should be allowed to say it because then the then then you know the free speech the debate will will prove them wrong.
1:37:16
You know you'll have good information to beat the bad information.
1:37:19
And if you think it's it's correct then you know obviously they should be allowed to say it because they should be allowed to tell the truth.
1:37:23
And then if you think that it's somewhere in between or if if it's uncertain then you know it's either case one or case two.
1:37:30
So, you should still let them uh you should still let them have free speech.
1:37:33
And and to be clear, I'm for free speech.
1:37:35
I'm not anti-free speech.
1:37:37
But what this did was it reef defined the old classical liberal rights established by Loach and I think established before we even had the term classical liberalism by Hobbes which was much more around uh survival and around peace.
1:37:52
that the idea was that you would have classical liberalism, you would have freedom of speech, freedom of conscience, the acknowledgement of of reason, uh, and and you would have that under the grounds of avoiding war that there were these intractable beliefs that people would have particularly around religion that if not dealt with in a mutually tolerant way would inevitably lead to war.
1:38:19
And this was not a hypothetical, of course.
1:38:21
This was the hundred years war.
1:38:23
There's many religious wars before that.
1:38:27
And um and the idea was that you would have these um natural rights that were enshrined in a mutual understanding that if we did not have these rights, it would be a violent free-for-all.
1:38:41
And that slowly got displaced with this idea idea of consensus that if we have these rights and not only is it possible but it it is inevitable that we will reach uh a total consensus.
1:38:57
And that broke down in all sorts of ways.
1:39:01
Most notably that people realized that that wasn't what was happening.
1:39:03
That as social media gave more people not just the right but the potential to speak that there was less consensus.
1:39:08
and and there, you know, as we talked about at the beginning, some of those people were just making [ __ ] up.
1:39:14
But it actually undermined this new theory of classical liberalism, which I think was not a theory of classical liberalism at all.
1:39:24
It turned into um a theory that was deeply antithetical to classical liberalism.
1:39:30
You know, Brian, uh, we could make this How long I I'm I'm not a listener usually, but how long does like, uh, Chris William Well, I do listen to Chris.
1:39:40
Uh, some >> This could be the eight hour Lex podcast.
1:39:44
>> Yeah, this could be the eight [laughter] hour, but but I am getting my hook from my producer.
1:39:48
So, what I will do is we've left so much left to be discussed.
1:39:52
We'll arrange for you to come back on.
1:39:54
Uh but for now, first off, tell tell everyone where they can find effort news and how they can subscribe. >> Yeah. Uh effort.
1:40:04
news is where you find us. effort. news/subscribe.
1:40:08
And uh to find some of the new stories that I'll publish on on socials, you can just find uh Brian Chow57 on all of the socials like X uh uh Instagram, so on.
1:40:20
Um that's B R I N C H AU57. >> Perfect.
1:40:24
uh and next time around we'll have to talk about all these other fascinating things because I I think uh I think we are in a very tumultuous and yet fascinating and innovative time and and I think the more we talk about these things and and discuss them the better people can understand them and reframe them.
1:40:44
uh you know we didn't even get into the difference between the real scientific method and scientism but I guess the shorthand there could be if anyone tells you the science is settled they're an idiot because science is never settled right it it just it keeps like yeah this works this works this works oh [ __ ] right Newtonian physics worked worked and worked until we the quantum guys came along and said actually.
1:41:13
So, it's a ongoing process, right?
1:41:16
And it should be falsifiable.
1:41:19
It should lead to better explanations.
1:41:22
It should lead to more explanatory power, but it's a process, right?
1:41:28
It's it's not anything that is settled, right?
1:41:30
I I often think of the true scientific method as being like a pure punk rock energy.
1:41:36
Uh it's like, no, I'm not going to take your word for it.
1:41:41
No, I'm gonna actually go >> and Verba. >> Exactly. Exactly.
1:41:44
Uh if you've seen uh or listened to the show, Brian, you know that we do have a final question for you.
1:41:52
And and that is we're going to make you the emperor of the world.
1:41:55
Uh you can't kill anyone.
1:41:57
You can't put anyone in a re-education camp.
1:41:59
But what you can do is we're going to hand you a magic microphone and you can say two things into it that is going to that are going to incept the entire population of the world whenever their next morning is they're going to wake up and they're going to say you know what I just had two of the greatest ideas and unlike all the other times I'm actually going to act on these two ideas.
1:42:24
What are you going to incept in the world's population?
1:42:28
First has to be the crisis of nominalism. I think that's right.
1:42:34
You have to ask the question of what's in that what's in the box.
1:42:36
And not everyone will come to good answers.
1:42:39
Some people will ask that question and they'll answer it by making stuff up.
1:42:45
But I think even asking that question, a lot of things cannot survive contact with that question.
1:42:50
And almost all of them that cannot survive contact with with that question of is is this nominalism uh should not survive contact with that question.
1:43:03
Second thing is I think this maybe this is this is this was primed by by listening to your your podcast with uh Jonathan B, but I think that that's like a question that that's like a question for everyone.
1:43:18
And then a question for people who are already doing things about it or who've set out to do things about it is uh more like the question we started on what is the domain of philosophy versus what is the domain of um versus what is more more the domain of action where they're definitely problems that you can be solve that can be solved by philosophizing around them. I love those problems.
1:43:46
Those problems are great in theory, but there are also a ton tons of problems where you just need to go out and do the thing.
1:43:52
And and I would have them very clearly distinguish. What is the thing?
1:43:56
What is which of these problems are problems of action versus problems of philosophy?
1:44:04
>> Both both good ones on the nominalism.
1:44:07
Uh it's funny my son and I were just having a long conversation about like what is wrong with people and and one of his observations was you know they they have not they have sufficiently made it so that they don't have to come into contact with reality.
1:44:26
In other words they they have sufficiently made it so that the box is there as you put it and they're not opening it.
1:44:33
They're not going to look in that box.
1:44:34
They're just quite happy with assuming what's in the box and and saying, "Yeah, what's in the box is true."
1:44:41
And and that is uh that is not uh going to get you very far.
1:44:46
You're going to have a >> It's a very Lovecraftian theory of knowledge, right? >> Yeah.
1:44:51
And and it's and you're going to have a look, George Bach said, "All models are wrong. Some are useful."
1:44:56
Uh if you don't upgrade yourself to a more useful model, you're going to suffer, right?
1:45:04
If you, as my son puts it, refuse to engage with and contact reality, guess what?
1:45:10
You're not gonna things aren't going to go your way.
1:45:13
And I just put up on Twitter the other day that Jed McKinn quote, you know, if if you're having a little tea party between Lord Lion and Lady Gazelle and somebody comes along who doesn't buy your fantasy narrative, it isn't that they are mean, it's just that your fantasy narrative is a bit fragile.
1:45:32
And contact with reality destroys that fragility, I think. >> Oh, absolutely.
1:45:42
>> Brian, this has been a ton of fun.
1:45:42
I wish you the greatest of success with effort News and uh we will reschedule you to continue the conversation.
1:45:52
>> Yeah, this was fantastic.
1:45:52
You're really an amazing interviewer, Jim.
1:45:54
And uh see you next time.
1:45:58
>> See you next time, Brian. Thanks for coming on.