0:00
[music] >> I'm here with Eddie Kim, Gusto's co-founder and head of technology.
Search this video
0:09
Gusto was in YC's winter 2012 batch and recently crossed a billion dollars in annual revenue and now serves over 500,000 small businesses in the US.
0:18
In fact, one in five new businesses started today are Gusto customers.
0:23
Today, we're going to talk about a new product they're releasing, which Eddie actually built himself, called Gusto co-founder, which uses AI to automate running a small business. Welcome, Eddie. >> Thanks for having me. >> Cool.
0:34
So, why don't we start I mean, actually, tell us a little bit about what exactly is Gusto co-founder.
0:36
What does it do and who is it for?
0:40
>> Yeah, so simply put, uh, Gusto co-founder is for small businesses and it will automate most of the business processes that a small business has.
0:47
Um, many of the things they're already doing on Gusto, but Gusto co-founder can take those things and automate it fully.
0:55
So, anything from, you know, running their payroll uh, from start to finish, approving time off, reminding your employees to submit timesheets, or even things outside of Gusto.
1:05
Uh, one example I like to give is texting you if it's going to rain today, cuz if you have a business that's like a tour guide business where it's highly dependent on the weather, you want to know that and then have Gusto co-founder actually email all of all of your customers for the day to let them know that it's going to rain and they should bring an umbrella.
1:23
>> I mean, I think that's actually a really interesting point.
1:24
Like, you guys are in a unique position cuz you build for small businesses.
1:28
They're really less technically sophisticated than a lot of the customers that, um, at least certainly YC companies today or other technology companies are building for.
1:36
So, what do you see as some of the challenges when you're trying to design a product for a small business owner, um, and especially an AI product that's, you know, living at the edge?
1:47
>> Yeah, I think one of the issues that I see is, especially, uh, for small businesses, is they really still use AI as sort of a glorified search engine.
1:56
Kind of ask it a question, they'll it'll reply.
2:02
Uh the more advanced small businesses out there, and I would argue like not just small businesses, but like most people who are using AI, the more advanced version of this is like maybe you have it um do some research for you, generate a report, or summarize something that you wrote, right? Or author something.
2:15
But they're still kind of largely like I have this like thing I have in mind, and then I'm expecting a response from it.
2:24
And this is like sort of a a agentic world that we've been promised has never really materialized for most people out there.
2:31
I think there's like the tech world that's on the cutting edge, folks that are like creative enough to come up with all these ways to automate, you know, all parts of their lives. But the remaining 99 99.
2:41
9% um are still kind of like using this as a glorified search engine.
2:46
And so what we wanted to do with Gusto co-founder is bring all of these like agentic promises that we were received that we received uh and turn them into into reality.
2:56
And so one of the ways that we've done that with Gusto co-founder is instead of starting with like a open claw type experience where uh it has a lot of powerful capabilities, um but it has a main has a prob- has a problem which I call the blank canvas problem.
3:14
We actually start with all of the things that Gusto's already solving for our customers.
3:18
You know, those are things in payroll, HR, time scheduling.
3:23
And then we sort of suggest ways that we can automate their running payroll, their approving of timesheets.
3:32
These are things that they're doing sort of on a regular basis.
3:34
You know, payrolls run once a week usually for hourly if you have hourly employees or twice a month if you have salaried employees.
3:43
Um and we we'll suggest ways that we can actually take it end to end without the user even having to like log into Gusto and and run their process.
3:53
>> Um something I'm really curious about is cuz you guys are already at like massive scale.
3:56
Um How What is it from the inside?
3:57
What does it look like trying to get a product like this is essentially new product like trying to get it out there?
4:04
Like um How did you come up with the idea and then just how do you sort of at your scale again like sort of figure out if this is something customers want?
4:12
Like how does it live alongside the existing product and just that cuz I feel it's so much more it's probably more complex there's more pieces at least than if you're just like starting from scratch and it's like you and your co-founders building it on day one. >> Yeah.
4:26
The original idea came from when uh I actually set up an open claw myself. >> Me too.
4:33
>> And like probably like you like I spent 8 hours just setting the the thing up.
4:38
Um I had to buy a Mac mini cuz at the time of the world where you could still get a Mac mini so I was able to actually get a Mac mini within 2 days.
4:44
Um installed everything on it.
4:46
I wanted to make sure you know I heard all the horror stories of it deleting people's email so I was like very careful to air gap it and I set up this you know pretty solid secure open claw which only like an engineer could could really do um and once I had like set all of that up um I kind of really didn't know what to do with it right?
5:08
I actually found myself also using it as sort of a glorified uh search engine right?
5:14
And I thought there were some really cool aspects of this like there were like the ability to like just text was like sounded so simple and like not that like innovative to me at the time but when I actually like had hands-on experience like doing it myself it was just like kind of like this is such a way better experience to like interact with AI than having to open up a web browser and and like log in to like uh Claude or or or ChatGPT.
5:41
>> I like I had exactly the same experience with my open claw and I've just like been saying this a lot around here at least that I am.
5:47
People just really underestimate chat as the interface and if anything I found that as I try to build out more functionality from Open Claw, like as an engineer or a tech person, your natural inclination is always oh, I can like let me add more functionality or let me make
6:01
the client better so that I know so it has a project management view or something like that, but I think in the world of AI you actually want to figure out ways to make the agent smarter so that you can do what you want from Telegram without needing like a fancy UI. Has that kind of played out for you
6:17
Has that kind of played out for you guys? >> Yeah, absolutely.
6:18
The other learning by the way was like, you know, everybody it's not the first time I heard about this.
6:24
Like people were saying like yeah, the ability to like talk to this thing with Telegram is pretty cool and you we read about it and there's like it was lots of people making YouTube videos about this.
6:34
And but then you really don't get it until you actually set it up yourself and you actually have hands-on experience.
6:40
And that was like one like important learning of like it is really important to still get into the weeds.
6:46
Like if you're a technical leader, you should be coding, but if you're not, you should also be like not just be reading about these things like Open Claw, but you should be you know, going out there and setting it up yourself.
6:56
Because like I just realized that there's this big gap between like what you read about and what you actually experience, right?
7:02
They're the same thing, but like that this like you they're they're completely different experiences. >> I agree.
7:05
So how did you go from kind of hobbyist interested in Open Claw, let's see what this can do to the initial idea for co-founder and building version one of that?
7:17
>> Yeah, so I had set this all up and you know, had my Open Claw running.
7:22
Actually went on vacation.
7:24
Went to Madrid and on my way back at the end of vacation, my layover was in London and I actually missed my my the the the second leg of the flight that was flying from London to San Francisco.
7:37
And so then I had all this time.
7:40
I I kind of pissed, but I had my computer on me and I was like, you know, writing a lot of code using Cloud Code.
7:48
And I was just kind of like blown away at what you could do with Cloud Code and like I could just literally just kind of like share my intent of wanting something and then Cloud Code would build it for me.
7:56
So then I asked myself, why can't we bring this power into the hands of our customers?
8:00
Why do they have to wait for Gusto to, you know, build its road map and deliver to our customers?
8:06
Why can't our customers actually like say, I want this from Gusto, put in a prompt and Gusto actually builds that experience for them.
8:13
So I kind of had that idea like kind of percolating in my in my mind.
8:18
And finally, because of this missed flight, like I actually had 5 hours of like uninterrupted time where I could actually try to build this.
8:23
So I, you know, took out my computer, I used Cloud Code, actually built a prototype of this in that 5-hour like sitting in the lounge at the airport.
8:32
And I was pretty I was pretty excited about what I had built.
8:36
And so when I got back to work, I started showing it to a few, you know, engineers, designers, our leadership, and I spent the next couple weeks kind of just sort of refining that idea.
8:50
And through conversation, it sort of evolved like there are pieces of this that like we really liked, pieces of this that like didn't work so well, and eventually led to the start of development of of Gusto Co-Founder.
9:00
>> What did that first prototype API just have some more what did it what did it do and then how did you shape it with the team?
9:07
>> Yeah, so it was really just a, you know, a a chat prompt.
9:11
The customer would say, I want a user experience where I can send out a form to my employees and they can fill out the survey.
9:18
Or I want to build a to-do tracking application.
9:21
Or I want to build a CRM.
9:26
So literally the customer would just say like I want this and the prototype that I had built would actually build the a web app essentially.
9:34
And it would use our, um, Gusto design system, so it actually look like Gusto had actually built it.
9:40
And then it would just be like an application that they would add to Gusto.
9:44
So, you can like literally have any application that you wanted.
9:47
>> And is all the context and data being pulled on from what uh their Gusto account, or do you have to go elsewhere?
9:52
>> The use case for this prototype was really just like a CRUD it it was a CRUD, uh, web app builder.
9:58
It like wrote code, it saved it into a database column somewhere, and then it would just like execute that code whenever you visited that that page.
10:05
So, it didn't actually, uh, and then I I added like the ability to like store data, but it would be like the data that you would bring in into it.
10:13
>> Cuz that seems like a huge advantage for you and any company that's in, I guess now it's been called like the system of record, but like you have you already have a ton of data, um, on the businesses.
10:21
[clears throat] At least where you're talking about Open Core, personally for me I found the unlock of Open Core is once I found a way to kind of securely give it access to things like my email and use embeddings, and just it has lots of context on me, like >> Right.
10:34
>> personally, professionally, >> Yeah.
10:36
>> it can it just orders of magnitude more useful.
10:38
As I'm get like for your customers, like how how does how does it sort of use the context that you have on them already to >> Right. >> help them do things?
10:47
>> Yeah, and and so the prototype was really cool.
10:49
Like we had this ability to like build any kind of like CRUD-based web app for our customers, but you're right, it didn't like really leverage anything about what we knew about the customer.
10:57
It didn't like, you know, populate it with data that we know about the customer.
11:01
And so that's how this idea sort of like went from prototype and and it sort of evolved to being less about like building an actual web app, um, to building more, uh, workflows uh, that we would automate for our customer, right?
11:15
So, we do have a lot of data on what customers are doing on Gusto.
11:18
Uh, we have some data on, you know, what you know, ba- what based on like what industry they're in, like if they're a dentist, like what things they would typically want to automate, uh, as, you know, someone in that vertical.
11:30
We so, we have all all this data like in aggregate across all of our customers and then also like for each individual customer what they're actually doing on Gusto.
11:39
And so, the idea sort of evolved to like, what if instead of building a prompt that builds you a web app uh we leverage a prompt plus the data that we have about our customers uh to build them an automation for a business process that they're doing every single week.
11:56
And so, that ultimately became the uh that's the main concept of Gusto co-founder which is centered around an automation.
12:04
And Gusto co-founder has many automations for a customer and they run on sort of a heartbeat that, you know, gets triggered on a regular basis.
12:16
Um that code by the way Yeah, and so, like I mean part of that was also like when I was setting up Open Co I was like curious how it worked.
12:20
So, it you know, it's open source so, I just read the source code of like, how is it actually like running this like automation for me?
12:30
And it was like surprisingly simple.
12:32
It's a cron job that runs an LLM every 30 minutes.
12:34
And that's exactly how it works in Gusto co-founder, right?
12:37
So, it wasn't even just like some of the concepts of like, you know, being able to text it through SMS or talk to it through Slack but it was literally some of the like technical designs uh of co-founder were heavily inspired by some of these tools that I started using personally.
12:53
>> will say the Open Co heartbeat idea is great but I found that um it's it's not actually I end up in like having to write a lot of just cron scripts like um outside of the gateway myself cuz it's just like not um there's some things you actually want to be deterministic.
13:07
And so, like payroll probably one of those things. >> Exactly, yeah.
13:09
And like I found the heartbeat um not being deterministic means it's like probably going to fire.
13:15
Like if I tell it to like do something in the morning it'll probably do it.
13:17
I just can't guarantee like when exactly it's going to do it. >> Exactly, yeah.
13:20
Yeah, so I actually for that exact reason we kind of started with a heartbeat.
13:24
We kind of like improved on it and added like multiple different ways to like trigger a job.
13:29
So, like there are like also in Gusto co-founder the ability to just schedule a job via a regular cron, right?
13:37
And actually will detect if this is a type if the thing that the customer's asking for is more suitable for a cron versus like running a you know, a prompt every 30 minutes, which also can get pretty expensive. >> Yeah.
13:47
>> [laughter] >> Um I think you guys are in such a unique position given your customer base.
13:50
So, some things I'm also curious about from just like a sales and talking to the customer perspective.
13:56
Um AI feels like it's such a hot topic in the enterprise world right now.
14:01
Like everyone knows what like a workflow is or what tokens are and what their token budget and it's just sort of like in the consciousness.
14:09
I'm imagining when you like your customers, I'm sure they've read about AI, but they probably haven't like I don't know if like automating workflows means as much to them.
14:17
So, how do you when you first go and show these products to them?
14:19
Like do you do you have to think about messaging it differently? Like Yeah.
14:25
Like you just what's that process like?
14:27
>> I I think businesses on Gusto like are I think we're pretty in a pretty unique situation because the things that people do on Gusto are sort of recurring tasks like running payroll, like approving time off, like things that they do week in and week out.
14:43
And they know that it actually wastes a lot of their time.
14:49
Um so, running payroll is actually like a pretty involved process that usually involves you know, exporting data from one system, uh manipulating that data, and then putting it into Gusto, and then running payroll.
15:04
So, for example, like there's a business like a massage uh spa business that um I talked to, and they export data from Mindbody, they put into Google Sheets, they have all these calculations on how that data from Mindbody gets translated into uh commission, tips, and um and hours, and and things like that.
15:25
And then, and only then, they put it into Gusto, and they run payroll.
15:28
That last part is obviously like very, very simple and straightforward, but it's all like kind of like this work before the work that takes them a lot of time, and they do like literally every single week, right?
15:39
And so, when you like actually talk about this to your customer, like they actually instantly get it because they have so many of these like, "I do this every week.
15:45
Like it's on my calendar, and it takes me like 1 hour every single week."
15:48
And if Gusto co-founder can just like do this for me, and it just texts me like a summary with a final approval where I just say yes or no over either SMS or Slack, like they actually really get that.
16:02
It doesn't really take much like selling at all.
16:05
>> Yeah, that makes a lot of sense.
16:05
And actually, I guess now I'm you're hearing you talk about it, I realize there's um there's a headwind you don't have, which is if you're selling into sort of like an an enterprise, there you never see face some resistance from people maybe worrying about how it's going to affect their jobs, and like what it like uh what it means for their role in the company.
16:22
Whereas, if you're working with small businesses, it just seems like it's all upside cuz none of them want presumably want to be like exporting data out of my body and putting it into QuickBooks or or whatever. >> Yeah, totally.
16:31
They just want to be able to do more with less.
16:33
Like small business owners, they're entrepreneurial, they're scrappy, they're like founders, right?
16:38
And so, like anything that you give them that can like save them time so they can go spend more time on like building their business.
16:44
Like that's that's that's a that's a that's a hard yes for them.
16:49
>> Yeah, I'm curious on that front, extending it out.
16:50
Like what do you think happens?
16:53
Like this product Co-founder successful, which I'm sure it will be, frees up a lot of time for the these small business owners.
16:59
Kind of where do you think that time goes, or maybe what are going to be knock-on effects of that?
17:05
>> I mean, I think they're going to, you know, spend this time looking for ways to grow their business, like expanding into more products, thinking a little bit more about strategy, how do they expand?
17:15
So, all of a sudden, it like frees them up to do all these things that uh they want to do, but they just didn't really have time to do.
17:20
I think that's like step one of like what we're trying to do with Co-founder.
17:24
Step two is like we also want it to be like and the reason why we called it Co-founder is we want it to be more than just a business process automator.
17:33
We want it to actually proactively let the customer know about things that, you know, they aren't doing that they should be doing, right?
17:42
That can range from like compliance tasks to like, you know, um stay, you know, make sure that they're aware of like the latest that's happening in like business compliance.
17:52
But, another is like, "Oh, you're doing some qualified research in your business.
17:57
There's a tax credit that you can get that the government offers called the R&D tax credit.
18:00
Um based on what I'm seeing in your company, you're probably eligible for this, and I can fill out the forms for you and make sure that like you get this credit.
18:07
Just I just need you to kind of like review it and and like let me give me permission to go. Let me do that."
18:12
And like that business owner might say, "Wow, that's amazing.
18:16
Like I didn't really know that existed.
18:18
Go ahead and do that for me, right?"
18:20
And that's that's not a hypothetical story.
18:21
Like there was a company called Cabana Pools that like we actually found them $50,000 in R&D tax credit.
18:28
They didn't really know that that was possible.
18:31
>> I bet you guys are in a really interesting spot to like you like level up the businesses, too.
18:34
Like in another lifetime, I used to build software for eBay power sellers.
18:38
Um and I found that um there's like there was like a sliver of power sellers who understood the value of like software and tools to, you know, like automate listings. Right?
18:46
In particular, what they would do is look at their um like listings in the same category and just constantly checking like the pricing and if they were priced appropriately. >> Right.
18:55
>> Um but they they those were like the like top 1% of technically sophisticated eBay power sellers.
19:00
I just like now you guys are in a good position where you could just take any small business and give them like market intelligence.
19:04
Like what like what should you be selling?
19:06
Like how could you optimize your business? >> Yeah. Yeah.
19:09
I mean, one of the automations I saw was um every week like take a look at what my competitors are doing and and generate a report for me and give me some suggestions on what I could do to stay ahead, right?
19:21
That's like one of the like more creative automations that I've seen.
19:27
But, you know, there's also no reason why we shouldn't be able to like say, "Oh, you're an eBay seller, like this is what you should be doing, right?
19:37
And and then kind of like ask the user if we can run this for them like every week.
19:41
That's a great example of like kind of maybe more of a proactive business partner that Gusto Co-Founder could become.
19:48
>> How's it going so far?
19:48
Like how many Can you share anything about how many businesses are using it and what anything that surprised you?
19:54
>> Well, so we just launched it and we added 500 customers to to Gusto Co-Founder.
19:58
And before that, we had like a kind of a pre-launch with We We call it a small business council, like a really small group of like 20 customers that we give like really early access to to things to to give it feedback.
20:11
And the feedback has been incredible.
20:13
Like as expected, like they everybody is blown away by the ability to run everything I'm doing on Gusto through text message, right?
20:19
This is the same experience.
20:21
So, same experience I had with Telegram with Open Claw, like I saw it in their face when when when we when I said, you know, try this thing, give us the phone number and like put in the eight-digit code to set it to set up the connection.
20:33
And now you can like run your payroll or contractor payment through through text message, right?
20:37
So, that was like one like very clear feedback that we got. Like this is incredible.
20:40
Like I can do this anywhere I'm at and it's so accessible.
20:45
It's been really cool to see just the automations that that companies are creating.
20:50
It's been What's been surprising is a lot of the automations don't necessarily have to do with things that Gusto is already doing for them.
20:57
So, obviously like the obvious ones are like payroll, time approvals, expense approvals, things like that.
21:02
But like if, you know, like you're you're a tour business, like text me the the what text me every morning if it's going to rain um today and if so, send an email to my customer list, which sits in this Excel file here, and let them know to bring an umbrella for today's tour, right?
21:18
That's like has nothing to do with what Gusto does for them at all, but is absolutely possible in what people are doing with with Co-Founder. >> Um what's next?
21:27
Any what what's next in the road map for Gusto Co-Founder?
21:31
>> Uh there's so many things.
21:31
Like right now we have SMS and Slack.
21:32
I want to add like a lot more channels.
21:37
Um so I do want to add things like uh Telegram.
21:40
I think a lot of businesses do things uh through WhatsApp.
21:44
>> a big [clears throat] one.
21:44
>> SMS is kind of a pain because like you're just limited on the on the character count.
21:47
So like just having Telegram or WhatsApp, I think it it are going to be like really important channels.
21:52
We want to add a lot more we call uh connectors, which are like right now today we have things like QuickBooks and Notion and Google Workspace.
22:01
Um but there's a lot of like vertical specific connectors.
22:02
Um I heard of a company I was talking to a dent- dental practice that uses um a patient management system called Curve Dental.
22:10
And so like why couldn't we get information in and out of there?
22:13
Um so there's probably like hundreds of connectors that we could be um adding.
22:14
I I think we're in this mode where we're just going to really observe what people are doing and you know, it's it's almost like when we launched where we we don't really know exactly how people are going to use it, but uh we're going to let the customers guide how we evolve the the road map.
22:31
>> Yeah, I mean I might a lot of what you're talking about right now seems like it you it will also pour fuel on the fire or help someone's existing business, but I wonder if there's ways to take it just get more people starting businesses who wouldn't have done >> Yeah. Done otherwise. >> Yeah.
22:45
Yeah, and and actually uh one of the things that we want to do in our next version of this is open it up to um people that don't necessarily have a company today.
22:55
Uh right now you have to be a payroll admin on Gusto on it's basically have to be an administrator for a company uh to use Gusto Co-Founder.
23:03
But we actually built it we built the data models in a way where that doesn't necessarily have to be true.
23:08
You could use Gusto Co-Founder even if you don't have a EIN yet, right?
23:13
So if you're thinking about starting a business or maybe you already have like a side hustle, like you should be able to use Gusto Co-Founder to automate and help you grow that business.
23:22
And of course at some point like your business grows and you might say to Gusto Co-Founder like can you help me get an EIN or can you help me register to be an employer in California and we would go do all that for you because we actually already have those capabilities as a company.
23:37
>> How has this product launch felt different for you personally?
23:39
Like I'm now you've always been obviously very close to the product and and things are launching but I'm guessing now like in the era of vibe coding you're able to like like you could build the prototype yourself versus like trying to get a team to rally around it.
23:52
>> That to me was even more mind-blowing than like kind of the what we built is like kind of how we built it.
23:58
So obviously it started with this like prototype.
24:01
I just kind of like showed a few of our engineers and designers and ultimately just kind of like organically roped them into working on this thing for me.
24:08
So one big change was like I was This is actually built by five people, five AI builders over the course of 10 weeks. From start to finish.
24:19
Just a whiteboarding session and 10 weeks later like we went to through a full-blown like tier one launch in the company.
24:28
That would never have been possible I think without AI.
24:29
And it was and the fact that it was just done with five people was also like equally and I was one of the five to in this team was even more mind-blowing to me.
24:41
And so yeah, I mean I think with AI just the level of like that someone like me who hasn't been hasn't coded in many many years like that's it's just changed so dramatically.
24:52
And then we also had we had it was four engineers and one designers one designer.
24:58
Uh and this designer like was contributing as much code as any of the engineers. >> Right.
25:04
>> And it was really, really cool to see a designer actually be able to like write like production grade code.
25:09
And then also like engineers really be comfortable with designing, right?
25:14
And what happened is the engineers would kind of let's just like prototype something, like do their best at design.
25:21
They wouldn't be afraid of They wouldn't get their, you know, wrist slaps because they didn't, you know, design something perfectly or use Figma in the beginning.
25:28
Um and the designer would know how to use cloud code enough to sort of like refine it and make it look really, really good, right?
25:34
And at the same time designer would like she wants some functionality, she would write the code.
25:40
Um and it was pretty good, but then the engineers would then take that and like sort of like get it, you know, the rest of the way there.
25:45
And so you kind of have this like merging, like designers are going more in engineering and engineers going more into design.
25:51
And so the craft really became secondary and everybody's like shared like responsibility was literally to write and commit code.
26:00
And we did that for 10 weeks and we were able to like move with speed I'd never seen before.
26:05
>> What's your advice to other people running companies already at scale and they're trying to figure out how do we like get that kind of product velocity and how do I like um get my team more AI filled?
26:16
>> It's something that I'm talking a lot about and and we're discussing a lot at the company.
26:19
Like what are what what can we learn in the process of how Guesstimate was was built?
26:27
Um and I think a lot of it is kind of like defined a little bit by like what did we Not necessarily what did we change, but it's more actually what did we not do.
26:34
We didn't have meetings, we didn't have any text backs, we didn't have any Figmas.
26:38
We set the goal of like we're not going to write any docs.
26:40
Now, I wouldn't recommend that for every single part of the organization, uh but for what we were doing, which is like kind of a zero to one type thing, like all we had If we didn't have like Jira board, we didn't have like sprint planning, we didn't have retros.
26:56
Uh we just had a uh literally the only thing that we had was one um perma-zoom.
27:00
So, we had a Zoom meeting that was going 24/7.
27:02
Uh people would just like hop in and out.
27:05
And um and Claude code, lots of Claude code tokens. And that was it.
27:11
Like it was almost more defined by like what we didn't do versus like what we changed in in how we build software.
27:19
>> And it was funny cuz I in a in a sense there there's no docs cuz you don't sort of sit down and formally write the docs and the specs.
27:25
But like in a way there's more documentation ever cuz like the transcripts of Claude code are effectively like documentation of like thoughts and then direction.
27:34
>> Yeah, and another thing we learned is that like I I mean first of all like all of all of us, myself included, threw away so much code, right?
27:39
So, you would have an idea and before discussing it or writing a doc about it, I would just like literally open up a pull request for it.
27:48
Then I would go to the Zoom and be like, "Hey, like I made this thing.
27:52
Like what do we all think about it, right?"
27:53
And I'd say like there's a good percentage of time it's like actually like that's not what we want.
27:57
We just delete that code. And that's okay.
27:59
Like that actually if you do that and like you have a hit rate of 50% it actually is significantly faster than like trying to like you know, write a PRD and a Figma and like getting a green light and like all that stuff.
28:12
Like it actually it all that stuff ends up taking longer than like just trying to write it three three different ways and you know, picking one of the three.
28:20
>> One of the things we've been talking or debating at least internally at YC is now I mean AI makes it so cheap to write code, like you said, that a clear benefit there is that you can just get your ideas out there and and discuss something tangible.
28:31
Um uh a downside is like now it's so cheap you could build like why only build one thing?
28:38
You could build like 10 things.
28:39
And um we've been debating whether that's actually good or not.
28:41
Like do you actually in the world of AI is it better to try and build 10 things at once because you can or should you still be trying to do one or two things well?
28:50
I'm just curious if you have any thoughts on that.
28:53
>> I think you still have to be really disciplined about like what the product looks like.
28:57
Like if anything you have to be much more disciplined than than before.
29:01
So in that sense like you can't just say oh I could I could throw every like you know bell and whistle into this thing and give it to our customer.
29:08
>> Yeah you have a lot more things to say no to now like >> Exactly yeah which makes it in some ways a lot harder.
29:12
But I think the the way you get to that like really disciplined set of functionality that you ultimately give to your customers I think the best way to do it is to actually like write the code in every single like permutation that that you can think of and then have a discussion about it.
29:30
When you just sort of like talking about it like from like a product requirement or UXR type perspective there's actually a lot of information you lose when you don't actually have the implementation in front of you.
29:46
And so I don't know if you necessarily make a better decision by like kind of having a bunch of meetings and writing a bunch of docs and running a bunch of user research up front because it's you're still missing like a huge part of the input which I think is the actual implementation cuz you learn so much in term when you actually like implement something.
30:03
>> Yeah I think there's a certain kind of like high bandwidth there's sort of like this data that you get from like just high bandwidth one-on-one conversations with customers and watching them use something that at least yeah Claude like doesn't quite have like the AGI to to replace humans on that. >> Right yeah yeah.
30:20
>> Like you were saying like essentially now we live in this age of abundance right like we can think anything we can make it happen um as engineers we've like we're learning to channel that but how um what's your advice for maybe small businesses like now it seems like uh it's an easier time than ever to start a business run a business um what's your what's your advice to them to make the most of this opportunity?
30:41
>> Yeah I mean I I think we've been on the secular trend for the past 15 years where it's become easier and easier to start a small business.
30:47
I think Gusto's been a part of that story with you know making parts of the back office like running payroll a lot easier than it was before.
30:56
And now I think with AI like it's like kind of a step function easier than even before to start a business.
31:02
All these things that you still had to like that were easier to do all these business processes compliance setting up benefits figuring out like arcane HR rules like all that stuff was still work for a small business despite it being much easier than it was before.
31:22
But now with AI like I think you can literally like automate many of these things and like focus on what business owners want to do most which is like just really focus on you know building their products getting more customers
31:33
growing growing their business and so and we're seeing it in the data right we're seeing it in census data of like how many business applications the government is receiving like more and more people are seeing this as a viable path for them to pursue. >> I mean I think it's like a great time to
31:46
>> I mean I think it's like a great time to just on the flip side also just be a consumer like we're just going to have more goods and services to choose from than ever so it's just kind of like a pretty amazing time to be alive. >> Yeah, totally.
31:56
And I love small business because like it's just like I don't know like you walk around Dogpatch here and just walking down like 3rd Street and seeing all the like random and quirky businesses that like just I don't know it just makes it so much more interesting to live in your communities. >> Cool.
32:10
Okay, well Andy thanks so much really excited to see where Gusto co-founder goes.
32:14
That's all we have time for today.
32:16
So >> Thanks for having me. >> [music] [music]