The State of Data & AI with Tom Tunguz, VC at Theory Ventures

is teeth at Google on the AdSense program before moving to Redpoint as a partner.

Richie Cotton

On top of that, he's a repeat data frame guest. So let's welcome Tom back and find out what's going to happen in 

Richie Cotton

So, Hi, Tom. Welcome to the show. Or should I say welcome back.

Tom Tunguz

Thrilled to be back. Thanks for having me on, Richie.

Richie Cotton

Yeah, yeah. Good to see you again. So I want to kick off with, The biggest question at the moment is, are we in an AI bubble? Different. This is true.

Tom Tunguz

Yes. I would say we are in a period of incredible, rapid investment in AI. And I mean, just to kind of. So I think it's we ran this analysis of comparing AI to other boom cycles or the bubbles and the the most relevant one, I think only, for scale is, railroads in America. So at the peak of railroads we were spending, the US was spending about a little bit more than % of our GDP on railroad, construction.

Tom Tunguz

And it took about years to get there. So the railroads started like and the peak decade was And we were spending the equivalent of what would be today of about five, trillion per year, right? I mean, it tells you like, wow. I mean, can you imagine how many people were chopping trees down?

Tom Tunguz

And so today, AI is about % of US GDP. Next year, probably closer to two and a half to three. So it will be half of the scale of railroads.

Richie Cotton

Okay. I had no idea railroads were so expensive. I mean, I suppose it makes sense is like, yeah, it's a lot of infrastructure. Then I guess that's the same thing with AI, right? Is it requires a lot of infrastructure in order to make it work.

Tom Tunguz

Right? Huge amount of infrastructure. And so, anyway, so it's not, our spending is profligate, but it's not as bananas as, hey, it was, years ago. The other dynamic here is we thought there were scaling limits. We thought that pre-training was dead. And all of the techniques about bringing in the total of human knowledge and compressing it into an LLM.

Tom Tunguz

And we had solved that problem. But the launch of Gemini three through basically shattered that illusion. There was a tweet two days ago, I think, when they launched Gemini three, the, the, the improvement from two and a half to three was the single largest improvement in performance of any model in Google's history. Oh, wow. Right. Which is I mean, that'll make you sit up and, and b evaluate.

Tom Tunguz

And what does that tell us? Well, you know, within the last six months, the team at Google have figured out how to completely break the perceived law that we thought of the scaling limits associated with pre-training. And that means that when the Blackwell chips ship and they're to times more powerful than the current hopper chips that we should see improved performance.

Tom Tunguz

So you have two components contributing to performance. You have an ability to improve algorithms both on pre training and post training. And post training was the theme of the summer. And then you have basically, you know, significant movement on on the chip side. And so yes, are we spending a whole amount of money on all this. No doubt.

Tom Tunguz

Are we seeing dramatic improvements in performance? Yes. And then the last thing is, I think as a result of those performance improvements, the productivity gains that we'll see should be real and significant. And then will we overspend, no doubt. Like we will we will build way more data centers than we need. Nobody knows what that point is. We will.

Richie Cotton

Overshoot. Okay. Yeah. That that's really, like, optimistic or at least very good news that, we haven't we're not hitting these limits, around scaling just yet. There are still innovations being made. And, do you have any sense of, like, where all these performance gains came from in Gemini, like you mentioned? But post post-training was a big theme recently.

Richie Cotton

Is that where we're getting all the gains from now?

Tom Tunguz

Well, yeah, the theme of this summer was around post training. So reinforcement learning environments and the way I think about that, and this is not totally true, but the way I think about it is if you have a robotic vacuum that enters in a room, it needs to do two different things. It needs to make a plan for how to clean, around tables and the carpet.

Tom Tunguz

And then it needs to reward itself. So just the way that you and I might take a vacation after a long stint of work, whether or grade ourselves at our work, where our objective and keep yourself to a class a Roomba has to create OKRs for cleaning a room. So a ten points for every square foot or meter, and then, you know, points for every chair and then points for a full bin over the summer.

Tom Tunguz

We have been focused on the automation of both of those. The first one is nearly fully automated. The second one is still research. And so that was a big push. And the theme coming out of Qwas, yeah, all the pre-training, all the compression of human knowledge is done. My guess. And I don't have any knowledge here, but my guess is there's been some pretty significant use of synthetic data to drive the pre-training gains.

Tom Tunguz

And that's really exciting because if we can have AI generate data that is then used to improve the AI systems, we have this virtuous cycle that will last for some period of time. The naysayers will say, and I think with some reasonable caution, that at some point, if you just keep generating synthetic data, you have what's called model collapse, or if you keep building fake data, infected and fake data, all of a sudden your foundation is not very strong.

Tom Tunguz

But Gemini three, and again, I don't know, the details suggest that there meaning I mean huge AI biggest gains ever to be had from intelligent applications of synthetic data.

Richie Cotton

That's very cool. I have to say, I like, your idea about, okay, it's basically like a reinforcement learning system for humans. Then I the dogs.

Tom Tunguz

Yeah. And if you can create right now, we're at a place where if you can create those loops and we could do it in coding, where you can spin up a cloud agent and tell it, keep iterating until you satisfy these conditions and satisfy these tests, then, it works. And now we're trying to figure out what are those reinforcement learning loops look like for non-coding applications.

Richie Cotton

Okay. Yeah. So certainly seems like, coding applications have been one of the big, areas, of AI, over the last year or two years, really about this is maybe the most important use case for generative AI. But, you think actually some of the techniques applied from that domain, they're going to be used, in other, you know, the fields.

Richie Cotton

Is that your suggestion?

Tom Tunguz

We've been deploying them inside of theory. So, managing most of my, you know, with an agent managed most of my tasks, that customer relationship management, system that we have is all through AI. We have agents developing, tech, algorithms to extract information from podcasts and, and lots of other applications. The challenge historically has been, what's called tool calling.

Tom Tunguz

So, read an email and then find the company URLs and there and then look them up in the CRM and then enrich them. There has been a significant amount of error at each step. When you have a multi-step process, that error explodes. The longer the number of, conditional probabilities. And what's been amazing and I have been using Gemini three year for the last three days is, the tool calling ability.

Tom Tunguz

I had looked at the benchmark. There's a benchmark that was created by Sierra called Tau, and the benchmarks would lead you to to believe that the improvement in tool calling ability of Gemini three compared to, sonnet or are relatively modest, but in practice it's unbelievable the difference. And so we're we're getting there. We're getting very, very close.

Richie Cotton

We are certainly I think, I mean, all year I feel like it's been talking about AI agents nonstop and at the start of years, like agents are coming at then it's gradual at the point. Well, they're kind of here. And but because every company's now selling an AI agent, some of them are very, very simple things that I guess a genetic is probably the maybe the politest way of saying it.

Richie Cotton

And some of them are incredibly like powerful things. Do you have a sense of where the sweet spot is in terms of like, powerful agents that actually do what they promise to do?

Tom Tunguz

Well, yeah, summarization agents, I think, were the first ones to work very, very well to large volumes of information. We're investors in a company called dropzone that summarizes security information for people who analyze the incidences, inside of secret operations centers, inside of enterprises. And they're the agents will query a database and they'll go and read online about some particular vulnerability.

Tom Tunguz

They might read previous documentation and summarize it. That works. This brilliant application of it is sort of the same thing, right? Go and research. Within a particular domain. I think we'll get to a place in early where you will start to see some much more sophisticated, applications. I mean, and we're started. There's some evidence of it within the world of customer support.

Tom Tunguz

So a typical customer support call might have to agent tool calls. That's not to say it's a single call with trains, pulled by a locomotive. It's a it's short, base pairs combinations. I'm mixing metaphors here, but. So we're starting to see some of these workflows with, with pretty significant tool calls. I hope, if anything should give us hope.

Tom Tunguz

OpenAI released the updated version of Codex this week where they said they had broken their record of like two months ago of a seven hour continuous job, and now it's closer to eight and a half, nine hours, right. So if we can do it in coding, well, there's no reason that we can't do it with marketing or customer support or sales.

Richie Cotton

Okay. That's interesting. So, yeah. I remember someone recently there saying, okay, if you're trying to build an agent and you want five tool calls max, and then otherwise it doesn't work, and you're saying like is now the sort of limit of what's good. So I guess this means like, it's like moving a file or looking up something from a database or it's it's opening or changing something on an, on a ticket somewhere.

Richie Cotton

Is that the idea or.

Tom Tunguz

Yeah. That's right. Yeah. Exactly. Right. Exactly. Right.

Richie Cotton

Okay. And so, the idea with, the long running Codex thing, then this is like what constitutes like, a good thing you can do with agents. So I get the feeling for AI coding, if you got a very simple task, you get the you get the bot to do it for you. You don't need to touch it yourself.

Richie Cotton

If it's like, I want to build an enterprise application, something takes days or something, it doesn't work. And so eight hours is like the sweet spot for what can do this run.

Tom Tunguz

Even though I'd say eight hours is the upper upper.

Richie Cotton

Bound up about.

Tom Tunguz

Yeah, up or bat. I think we consistently see agents operating for to minutes with success. And that will be the sweet spot today.

Richie Cotton

minutes. Okay. So.

Tom Tunguz

Encoding task encoding does.

Richie Cotton

Does that mean you need to change your workflows in terms of like trying to break things down into smaller chunks. So then you can hand it off to an agent? Well, I think.

Tom Tunguz

It's the opposite. I think it's how do you start to create bigger and more complex documents that describe an entire project in a way an agent can understand and work on while you are hosting a podcast, while you are attending a meeting, while you're having lunch. I think the the status symbol now will be how many agents are working for me while I'm having tea.

Richie Cotton

Okay, I that yeah. How much going to get done while I'm busy doing the stuff that I need to do as a yeah.

Tom Tunguz

How much passive work right. There is a wave I think, in cognitive like side hustles. And now it's actually no, I want all these agents working for me. I can I have ten of them simultaneously. Can I fill up a queue hopper that where all of the tasks are sufficiently well-defined that the agent can go and and just grind through it?

Richie Cotton

Okay, I do love the idea of just like, outsourcing a lot of my work in this way. Do you have a sense of how to get started on this? Like how do you get your first agents going?

Tom Tunguz

Oh, I think the okay, so a very simple agent is, upcoming meeting. Summarizer. That's a very simple one for a lot of people. You know, podcast listener, it was one of the first ones that we built a newsletter, Summarizer. So I have an agent that goes through my inbox and looks at any email that has the word unsubscribe in IT archives, all of those emails, and then summarizes and tries to find anything that could be interesting and then sends me a compendium of all that.

Tom Tunguz

Those are the really simple agents or actually the. Okay, so I have this I have this question in my mind, which is what is the difference between an agent and the cron job?

Richie Cotton

Oh, go on, talk through. I presume the difference is like some kind of, smart, thinking or a.

Tom Tunguz

Yeah, I think, I think the main difference is that the, the agent should be self-improving, but that's not necessarily always the case. Like a cutting agent does not necessarily improve. A coding agent is just constantly checking to see is there work for me to code, and if so, I code and if not I don't then a cron job.

Tom Tunguz

Well, it's kind of the same thing, right? Check to see if I have new emails, or I should be checking to see if I have new emails. And if I do, I process it and if I don't. So maybe the only difference is that there's some form of non-deterministic programing within, an agent that as I've developed all these different tools internally, my crontab file is enormously long because there are all these agents running at different intervals.

Tom Tunguz

Check every one minute, check every three minutes. And in fact, I tip for anybody else building out there. Don't use cron on Mac. You should lose use launch D. It's far more robust. Com or restart after you put your Mac to sleep.

Richie Cotton

Okay. That I feel like this, horror story. What are you supposed to do? Trying to figure something out?

Tom Tunguz

No, no, no.

Richie Cotton

It's just it's just like I.

Tom Tunguz

Just killed all these agents, and they. They're so lazy every time I put the computer sleep down on startup. And so. And it was just, it was cloud code that that taught me that the cron is historically and has been for a really long time, incredibly brittle.

Richie Cotton

Okay. No. No chrome on Mac. That feels like you've a useful thing to know about. Or so, I think, I mean, we're talking about, outsourcing a lot of your work to agents here. I know, recently the job market's been pretty tough, particularly for recent graduates. Do you have any career advice, like, particularly for people like you just coming out of college?

Richie Cotton

Or are you just trying to get start on your career? What do you do?

Tom Tunguz

Yeah, unemployment rates are very college graduates are approaching %, which is quite high. The I think the most important thing is there's a tremendous amount of demand for people who, irrespective of their age or seniority, can reinvent workflows with AI. Nobody knows that. There's no one in the Valley who has a secret for the best way of working with a coding agent, there's no one who has discovered the new optimal way of designing a marketing campaign, or the best way of structuring a customer support organization.

Tom Tunguz

It's all new, and in addition to it being all new, it seems to be all new every three months because the capabilities of the underlying models advance to such an extent, the previous assumptions no longer hold. And consequently, if you are a person with ambition and curiosity demonstrating a reinvention of a workflow in a domain you want to pursue is very likely to help you get a job.

Tom Tunguz

Because a manager hiring manager will look at that implementation and say, this may or may not be perfect, but it is a meaningful step forward. But more importantly, I'm looking at someone who has both the curiosity and the ambition to iterate to what could be a meaningful differentiator for me internally. And as a hiring manager, I want to be promoted and then ultimately for the business as a differentiator.

Richie Cotton

I love that just being curious and trying new things because it's kind of Wild West. Everyone's, equal in this sense.

Tom Tunguz

Yeah. That's right. We reached I mean, with SAS, we had reached. Yeah. This is what the modern data stack look like. And this is the Salesforce Marketo stack. We don't have any of that. It's all wide open.

Richie Cotton

However. And so actually, I've heard it, both ways from hiring managers. So I think, some companies are very much like, we only want to hire very senior people because they've got domain knowledge, and we don't care about junior people at all. And other companies are we want young people because they're much more likely to be content with a fresh mind about how can I use AI to improve my productivity here?

Richie Cotton

So I don't want to have a take on this. It sounds like you're leaning a bit towards the latter.

Tom Tunguz

Yeah, so you definitely see, I think you're right. We used to we used to see engineering organizations structured as pyramids, where you'd have one senior person and then two architects and a bunch of tech leads, and then a lot of software engineers and and now we see them structured more like rocket ships, where the wide part of the pyramid is smaller.

Tom Tunguz

I, I bet that is a temporary phenomenon. And I think it's a temporary phenomenon, because the rate of code generation by those senior people will very quickly outstrip their ability to review the code. There's just no way. There's just no way. And we still need human in the loop, to be able to review the code. And that will initially be some of the work of people coming out of, school will be evaluating the code and effectively QA on top of the AI quality assurance, on top of the AI generated software.

Tom Tunguz

And the other dynamic here is I think we're in this little nadir of people generating a tremendous amount of productivity, while the rest of the market catches up. But as the rest of the market catches up and we'll go back to higher levels of burn, higher level of investment because the overall productivity to remain competitive will rise. Just the way the sea level rise is.

Tom Tunguz

Everyone needs to respond.

Richie Cotton

Okay. All right. So you think, it might be, a temporary phenomenon then the even more senior people can, but at some point, it's going to change again. So actually, you mentioned code reviews is going to be one thing, this kind of, I guess, bottleneck in the near future. I have heard, an idea that actually for it, teams need to change because engineers are more productive.

Richie Cotton

You need to hire more product managers in order to manage the extra stuff that the engineers, are creating. Do you believe that story?

Tom Tunguz

I think it's true. Yeah. I think at some point, well, so within EPD, within the engineering product and design, an engineer can now code, design a product and then design the UX themselves. And the same is true for a designer. A designer can design the product, create a pod, and then ultimately code it. So you have a skill set expansion that blurs the lines across the EPD departments.

Tom Tunguz

And so some people take I mean, I've heard some perspectives where and even clear will say this, AI is the death of product management.

Richie Cotton

Oh, it's going the opposite direction then.

Tom Tunguz

Yeah. Why? If you're an engineering manager, why would you need a p m? P m is just orchestrating communication across multiple stakeholders, both internal and external. What is the customer one. How do we build it? When do we prioritize it? Can we validated? Who are the initial design partners? Well, you could argue I can do all of that.

Tom Tunguz

A product manager can look at an engineer's role and say, particularly for front end work, well, an engineer is just implementing a spec. Well, what is the coding agent do? Oh yeah. Well, we need far fewer agents. I think we'll see. Short term aberrations, but I don't think we'll see pretty. I think the AI engineer teams of the future will look very similar to the ones that we have today.

Tom Tunguz

The ratios might change the productivity will be hugely, increased. But I don't know if any one role increases in scope or is obviated, I think that remain relatively similar.

Richie Cotton

Okay. All right. Yeah. I guess you can't get rid of, like, one role entirely, but there's going to be some sort of blur between the different roles. I mean, do you think any other, teams, going to be restructured or they're going to be affected by, the, AI productivity changes?

Tom Tunguz

Well, customer success is probably the one that has been the most acutely and immediately impacted with the advent of the forward deployed engineer.

Richie Cotton

Oh, tell me about this role. I've heard about it. What it involve?

Tom Tunguz

Yeah. For deployed engineers is exactly what it sounds like. It's a salesperson who is an engineer, and they typically report to the VP of engineering that they are deployed or they are sent to the customer. And the primary responsibility is taking an AI product and making sure it works really well at a customer. And so you need somebody who, as typically has an engineering degree to do that.

Tom Tunguz

Historically, customer success has been not so technical. And if you think about the boom in the SAS ecosystem, if you're configuring HubSpot or if you're configuring Salesforce, you don't need an engineer to be able to do that. In fact, you probably want somebody who is more comfortable or looks more like the buyer. So the buyer has a certain level of comfort.

Tom Tunguz

Many of these AI systems are bought by very sophisticated engineering teams, and as a result, you want engineers on both sides. And when the buyer is not a technical person, there's, another need, which is the configuration of the software to drive the intended outcome, particularly because many of these pieces of software charge on a resolution basis actually do the work.

Tom Tunguz

That's really important. And so the combination of those two things has basically reinvented customer success for many companies to be these forward deployed engineers, which, they're typically much more expensive. And the contract sizes have to go up and they have. But so that's, that's I think the team that has changed the most.

Richie Cotton

That's fascinating. I have to say, the combination of sales skills and, engineering skills, that seems to be pretty or certainly like, I guess post-sales and an engineering know many people with that. So I guess, is a is that like a good high salary role then? If it's such a rare thing.

Tom Tunguz

Yeah, yeah. I mean, I think the number of FTE job requisitions the beginning of the year up ten x.

Richie Cotton

Oh, wow. Okay. All right. So yeah, maybe a career to account for some of the audience. And if you've got they're both those sorts of, the skills, I'd love to, think about, like, how you, go about dealing with all these, like, rapid changes. So, I mean, because there's so much going on. I know you talked a bit about how product market fit for a lot of products just changes.

Richie Cotton

So rapidly. I can joke with you like, why is that happening and how do you keep up?

Tom Tunguz

Yeah. So I think, so Steve, I wrote this book called Four Steps of the epiphany, which the design philosophy for companies around establishing product market fit. I remember reading it in It was a brilliant book. And then it kind of created this movement called the Lean Startup Movement. The idea was remain as small as possible until you've achieved PMF, and at that point you can scale.

Tom Tunguz

And then there was a, counter movement, countercultural movement, which is called the fat startup. Raise as much money as possible, hire a lot of sales people, and then after you've sold, figure the product out, which actually is the way Salesforce was built. If you read Benioff book, a tremendous amount of salespeople, it was burning a huge amount of money very early on.

Tom Tunguz

And during that era, it was kind of taken for granted that once you had crossed the PMF threshold, once you were able to, a startup was able to grow from founder led sales to account, executive led sales. The founder could train a salesperson to sell the product, and that they could demonstrate repeatability, that, it was all about scaling.

Tom Tunguz

And that's because the underlying infrastructure really didn't change that much, and fire demands really didn't change that much. And today, as we've talked about Richie, the capabilities of the models are changing so fast, the buyers don't really know what they want because nobody knows what's possible. The definition of PMF is no longer constant. It's this like you have to constantly reestablish PMF.

Tom Tunguz

So you might have a foundation model company moving up and competing with, one of the clients, one of its customers. And so all of a sudden, you might have had PMF as a business, an agenda company building on top of, say, anthropic or OpenAI. But if OpenAI or anthropic move into your market and they have the underlying model where you've lost PMF or there have been categories of software where initially the startup was developing a content generation, prompts that were very, very good at generating prompts.

Tom Tunguz

And then a new model iteration comes around and the that model has intuited, its prompts. And so there's no value in those additional prompts. And you've lost product market fit. So I think that's that's the dynamic where we're seeing. And it's happening so fast. Faster than I've ever seen it.

Richie Cotton

Wow. Yeah. I mean, to me, when the technology is changing so fast and like, everyone's kind of jostling for position, yeah, there's this chaos going on. So, Jay, do you think, like, do you have a sense of who is going to be the big winners from all this change and who are going to be the losers?

Richie Cotton

Like, how do you make sure that, you're going to be the winner, I guess.

Tom Tunguz

Well, I think customers will be the winner because they have so much choice and the models are becoming that much better. And, the price for performance is significantly better. The that's the nimbleness companies, the companies that are relentlessly customer focused will be the ones who win because they need to anticipate and hear from their customers how those buyer preferences are changing.

Richie Cotton

Okay. Oh, well, maybe there is still the use of product managers and it's like speaker. That's right. Deciding what you want. Okay. Yeah. So, the idea of like, being customer focused is the way to win. So, you mentioned, that in general, customers are winning because there's so much, competition. And I have to say, like, any time, like, I think of, like, what's an AI use case, you know, you do a search for that and there's already like at least half a dozen companies trying to solve a problem with AI.

Richie Cotton

Do you think this sort of overlap is going to persist? Like the fact that, like, almost any possible I use case for most startups at the moment?

Tom Tunguz

Yeah, but the price is so big. Yes. And then the amount of venture capital dollars coming in are just ramping, I think %, % of venture dollars this year are already in AI and asymptote to %, a total dollars coming into the venture capital asset class are growing pretty enormously. I think there's a very reasonable likelihood just to kind of give everybody a sense.

Tom Tunguz

The US venture capital industry was about billion when I started in And, it's currently around let's say to I think by the end of the decade you might see a profit, approach billion. And there's lots of different reasons for that. But if that's the case, then you can think about the number of startups that would be funded.

Tom Tunguz

And that's, twice as many today.

Richie Cotton

As impressive numbers. Like, I'm, I'm always bugged with the these things I have to start with like, well GDP I think is about $trillion. So you go like, okay, trillion. That's % of world GDP. So, billion as like as a good chunk of like the amount of money in the world.

Tom Tunguz

Yeah. Well, I mean, yeah. So half $trillion in the US is about it's about one third of %. I know it's about %. It's about % of US GDP.

Richie Cotton

Okay. All right. Yeah.

Tom Tunguz

Yeah, I was a little, %. Yeah.

Richie Cotton

There's probably an I can, I'll. That was about.

Tom Tunguz

You know, just trying to do the math. Yeah. It's about % of US GDP. That would be then going into startups, would you? We'd think about. I mean, I remember looking at the National Venture Capital Association analysis of publicly traded companies, something like a quarter to a third of publicly traded companies at the time where venture backed and so and the total number of jobs created as a result of them was something similar.

Tom Tunguz

So it's actually an unbelievable a an unbelievably efficient machine. Yeah.

Richie Cotton

Incredibly. So, so I'm curious, so since you've got quite a lot of overlap with, with, I startup and I guess even bigger companies. Like what, what they're selling. Do you think we're going to see, a round of, like, mergers or acquisitions going on next year?

Tom Tunguz

I do, I the FTC just lost their lawsuit against, meta Facebook, for monopoly within, social media and I really hope that that opens up M&A in a very big way. One of my hopes is that we stop seeing these Frankenstein acquisitions where, the cap tables are renegotiated at acquisition time, where only a fraction of the team is going at the totality of the team.

Richie Cotton

Oh, can you give an example of this?

Tom Tunguz

Yeah. I mean, I think, the scale I met acquisition, where the business was valued at billion, meta paid billion, and a small number of the scale team went over to meta. But but scale is continuing to operate as a separate entity. It's not really M&A.

Richie Cotton

This is the like I guess, skirt around all the, the rules on mergers and acquisitions.

Tom Tunguz

Yeah, yeah. And the hard part about all that is that it breaks the social contract. It breaks the social contract that early employees make when they join a startup. And they say, I'm willing to take below market compensation because I believe in this business. In the case of an extraordinary outcome, I will be rewarded commensurately. And if that doesn't happen and that loss of trust exists, well, then Silicon Valley is a concept.

Tom Tunguz

We cannot cannot thrive because there's distrust. And so anyway, I think we had this and I hope it's a temporary contortion around, legislation. And I hope that the FTC or the lawsuit that that meta one really opens up the M&A environment, in a very meaningful way. So I think we'll see that in 

Richie Cotton

Okay. Yeah. So, I guess some interesting, corporate, transitions, going to happen. I see the biggest sort of recent example of this. It was it was on the dataset, on the AI, with, Fivetran buying DBT labs. I know you have a take on is this going to be good for the the data industry is, well, I think.

Tom Tunguz

I'm the stock is in a wave of consolidation, no doubt about that. The DBT merger very strategically important. It's a way for, I guess the way I phrase it was, you have two unicorns competing with two deck corns with Snowflake and Databricks. And by joining hands, they now are a viable third alternative. And there are three just simplifying for a second.

Tom Tunguz

There are three components that we're talking about. We're talking about the ETL. So bringing data from many different places into our database. There's the database itself or the cloud data warehouse. And then there's the transformation of the data within that within that database. And so DBT plus Fivetran represent two of those three. And Snowflake and Databricks also represent two of those three.

Tom Tunguz

Although they're not identical. And so I would expect to see DBT fivetran push into compute and develop their own cloud data warehouse, because that's where all the margin is to be able to compete. And and then maybe they can create a third viable alternative to Snowflake and Databricks. And that would be the ambition, I suspect, behind the merger.

Richie Cotton

They'll be fascinating. I'm been, really for the last few years, it has been snowflake and Databricks have been the two main players in the sort of cloud data warehouse. So, yeah, having having a third option rather than that being a duopoly. B very interesting.

Tom Tunguz

It would be very interesting. Yeah. Okay.

Richie Cotton

And I guess where does that leave, business intelligence then? Because that's, that's the part that those companies don't really touch.

Tom Tunguz

Yeah. Well, we are investors in a company called Omni, that's growing incredibly quickly, also riding the wave of consolidation. And they, it's an X looker team. And, they're having a tremendous amount of success, consolidating existing buy spend and rationalizing on to a single product, by I think will remain a separate category even though it is adjacent.

Tom Tunguz

Many times customers by API plus data warehouses at the same time to have an end to end solution that both products are sufficient. They're just different, right? Or your database system has its own level of complexity. Much more infrastructure focused. And by system has a not obvious level of complexity underneath that. A lot of people think it's just charts, but it's actually far more complicated than that.

Tom Tunguz

And so I think they'll remain relatively separate here for forever.

Richie Cotton

That's interesting. And actually, Colin Zimmer, the, the CEO, I mean, he's, actually a former DataFrame guest as well. So. Yeah. Yeah, a very interesting take on like how BI is changing a lot of things about just getting the data right. And how do you think about what's going to be what's going to feed into my dashboard rather than just like, like anyone can create a dashboard at this point, but it's like, how do you get the data right first?

Tom Tunguz

Yeah, that's right. There's this pendulum within the world of BI that swings between governance and freedom or, anarchy. And so I guess I characterize it is the during the first wave, you've had Cognos Business Objects, Hyperion and MicroStrategy sold it. It defined all of the reports, took a really long time to develop those reports. There was a rebellion in the form of Tableau in that allowed anybody to download visualization software and make really beautiful charts, and that created a sort of chaos.

Tom Tunguz

And then Cloud Wave came and looker, developed a system of governance, on top of cloud. And then on the is trying to and I think doing very well, being able to straddle both the individual creator freedom as well as the governance associated with standardizing metrics through an organization.

Richie Cotton

Absolutely. Like this very much, a different stream, like, okay, I'm creating a dashboard to analyze, like podcast lessons or something like that. And this is a dashboard that is like the official metric of like, how much money does the company have? Like they need very, very different levels of governance and freedom.

Tom Tunguz

Exactly, exactly. Right.

Richie Cotton

Okay. Wonderful. Or are there any other trends that you're particularly, interested in at the moment or you're keen to talk about? Well, I think we're.

Tom Tunguz

Seeing a lot in the world of AI, data engineering and the automation of AI, of data engineering, the dynamic assembly of pipelines at runtime. One of the questions we've been asking ourselves is there's been a wave of companies developing AI specific languages, programing languages, and we can debate whether or not that has, merit. But it is interesting to ask ourselves the question, what if SQL were rewritten without the need for humans to understand what it was doing differently?

Richie Cotton

That's fascinating. I'm SQL has been around more than years now. It's not changing often. Lot. Actually, I quite like some of the changes. In DB, like, they've cleaned up SQL a lot.

Tom Tunguz

Yeah. And then the line calls where you can call an LLM as a column.

Richie Cotton

Yeah, yeah. But tell me about these are the, these new programing languages specifically for AI.

Tom Tunguz

There are many of them are still in stealth. And the idea is many of the, the programing languages have affordances for humans so that they're legible. Right. We went from assembly to Python. Python looks like a poem. And assembly looks like somebody banging their hands on a keyboard. And so as we move higher and higher in the world of interpreted languages, it becomes more and more prose like, but now we can go in the opposite direction, which is the why can't I write assembly directly and then only transpile to an English, interpretable language during debug?

Richie Cotton

Okay, I guess this was kind of the idea behind, Microsoft. Net right. You had these sort of, you got C-sharp and whatever, which is vaguely human readable, but then it compiles. Oh, no, but you had this intermediate, dot net language underneath. And then so there was, I guess, human readable, sorry, machine readable, but not human readable.

Richie Cotton

So is is this a similar idea?

Tom Tunguz

It's a similar idea. Yeah. But instead of, the idea is that a machine would just, in, intake English and then output assembly or machine code by by code. And I don't know if it would work. I mean, the reality is for a lot of these, models, the reasons are so phenomenal. Coding is just because of GitHub, right?

Tom Tunguz

The amount of open source code and the examples and the design patterns associated with these languages, it's always fascinating thought experiment to say, let us remove the the initial parameters that that really determined the path that this particular technology has progressed over the last years. Right. It's total first principles thinking. And that's that's always exciting to people who want ideas.

Richie Cotton

Absolutely. Yeah. I mean, so I remember there was a thing last year. So I mean, most, I code seems to be I mean, there's all this stuff here in PyTorch, which is like this sort of, nice layer of, this is how I build models. But then Andrej Karpathy was like, okay, let's rewrite, I think Gptand C, and it's like it runs much faster and it's like a much smaller code basically, you know, all these like Python dependencies.

Richie Cotton

So this seems like a similar sort of idea is like, can we go a bit, closer to the metal?

Tom Tunguz

But that's right. Exactly. Very well summarized.

Richie Cotton

All right. Wonderful. Okay, so, before we wrap up, I need to talk about, one of the, one of the things you built. So I know you've created your own podcast processor. So, tell me all about this. Like, what did you do and how did you build it?

Tom Tunguz

Yeah, I don't have ham. I mean, there are so many wonderful podcasts, and, and there's a lot of phenomenal information within those podcasts. And so I have time for a few per week, but, wouldn't it be great if Robert could listen to the podcast on my behalf and identify the key quotes? So we built an internal system that collects all the different podcasts, listens to all of them, and then identifies interesting statistics or companies or people that we should know about.

Tom Tunguz

And it runs in the background. And and it's been working really well, I think, for the last or months. We're able to process. I think my list is now around or different podcasts, and it helps us with diligence and figuring out memos. And, what are the data points might be interesting to a particular company.

Tom Tunguz

So becoming a, compounding data set for us.

Richie Cotton

Is, very interesting. I'm always thinking of like, the de trend audience being humans, but maybe like at the future, like some of my audience is bots as well.

Tom Tunguz

And yeah, I mean, it's it's, I wonder if the speech sound will be different. Right? We're we're starting to talk about generating websites that are optimized for robots and certain websites. More than % of their traffic is for AI. Will it be different for speech, I wonder?

Richie Cotton

Absolutely. I mean, yeah, I see, someone's trying to pitch me the idea that you need to, redesign your retail experience for bots because shopping agents are going to be the future at some point in the. Well, I guess maybe maybe 

Tom Tunguz

Yeah. It's right. I mean, so Nike, when they drop a new pair of Jordans, more than % of the traffic that day is robot. There are BB robots, wholesalers who are looking to buy that shoe and then resell it. There are, customers that Nike would love for them to buy the shoe. And then there are customers that Nike would prefer that they don't.

Tom Tunguz

And so, how how does Nike.com tell them apart?

Richie Cotton

Yeah. That's interesting. I have to say, Sneakerhead culture is a bit alien to me.

Tom Tunguz

Yeah. But. Right. But do you understand the dynamic, right? Or let's say.

Richie Cotton

No, it's the same with with concert tickets, right? Is that, it's a very popular, is is having a show and then there's a limited number of space in the, auditorium, so. Yeah, there's a lot of bots trying to buy tickets immediately.

Tom Tunguz

Great parallel. That's probably a much more tangible payoff for the audience.

Richie Cotton

We, Are you into sneakers by chance? I'm just wondering whether that's where, you know.

Tom Tunguz

No, no, no, I, I, I have the ones that come in the store, the ones that then I need an agent to bid for. Okay.

Richie Cotton

Yeah, but, that does actually sound like quite a fun project. If you want to build an agent. Okay, go and build your own shopping agent. It's maybe high risk because, you know, you give me your credit card details, but, if you're really into sneakers or or cons.

Tom Tunguz

Yeah, yeah, well, you can create a virtual card with a with a limit on it so it can only spend a certain amount of money. And so you can. But yeah, I think this is the way I mean, I wonder if this is the way. I don't know, we were researching pickleball rackets yesterday, and, I just asked for the four top Amazon links and then clicked on it, pressed by.

Tom Tunguz

And you read the reviews I just asked, I don't read them for me. And so we're close. We're really close. Yeah.

Richie Cotton

As so it depends on like the risk of purchase pickleball racket. It's like, okay, I guess you can't go too far wrong. And you've only wasted like tens, maybe, maybe hundreds of dollars. But like, would you do that for like, a car or would you like, how big do you want to go on this, I guess?

Tom Tunguz

Well, okay, this is the awesome because we're not far from this. But I think if you're in Texas, you could actually have a flow where you ask I, which is the best Tesla to buy for you, and then you can go to the Tesla website and press buy. And then a few hours later, the car drives itself to your house.

Tom Tunguz

That's really living in the future.

Richie Cotton

That is very rich in the future. Yeah. Nice. Okay. I guess, for the brave souls, like, putting a lot of trust in AI, then. Yeah, go for it. I want to test drive it first. Like, I hope it an option to, like, go back on your word at that point. Yeah. Yeah. Okay.

Richie Cotton

Cool. All right, so, just to wrap up, guys always need new ideas for people to follow whose work are you most interested in at the moment?

Tom Tunguz

Oh, there's this wonderful guy, Gavin Baker. Okay. Gavin runs a, hedge fund called As. And he is, wonderful analyst of AI and the public markets. So I had a wonderful tweet, I think, earlier this week just about, what's happening within the world of AI. And so he's very much on point. So I, I will recommend that.

Richie Cotton

Gavin make it. All right. Okay. One to watch out for that. Super. All right. Thank you so much for your time, Tom.

Tom Tunguz

Ritzy pleasure was mine. Always good to see you.