0:00Today's guest is Tomasz Tunguz, a venture capitalist who has worked with at least 8 unicorns. In a past life, he ran a billion-dollar business unit at Google as a product manager. In today's episode, he shares how the AI boom is warping every traditional sales metric as well as broader investment dynamics. But he also warns CROs not to be fooled by the astronomical numbers certain reps are posting right now. It's not the supply side has changed and suddenly become 10x more productive.
0:25It's the demand side budgets have increased by a factor of 10, and that's what's driving quotas.
0:30So in today's episode, We cover why product moats are disappearing and why that means VCs are hungry for go-to-market innovation in their portfolios. Plus, we cover a lot more ground. Why the mid-market is being abandoned for fast-moving enterprise deals, how AI is blurring the lines between SaaS and consulting, and why using AI to save time is the wrong way to unlock top-tier performance.
0:50If you're a chess grandmaster and you want to be better with AI, you don't train less.
0:55You train just as much.
0:58But--- If you're a go-to-market leader in the AI era, this episode is for you.
1:03Hey everybody, it's Sam Jacobs. Welcome to Topline. I'm joined by my incredible co-hosts, Asad Zaman, the CEO of Sales Talent Agency, and AJ Bruno, the CEO of QuotaPath, leaders in incentive compensation design, management, and optimization.
1:18And we're joined by a very special guest, friend of the pod, Tomasz Tunguz. Uh, if you don't know Tomasz, you should. He's a venture capitalist who has worked with 8 unicorns predominantly in data and data infrastructure, including Looker, Monte Carlo, Dremio, Hex, Omni, and MotherDuck.
1:32He writes a blog at tomtunguz.com that drives millions of pages per month. And before becoming an investor, he was a product manager at Google managing a billion-dollar business unit on the AdSense team. Tomasz, welcome back to Topline.
1:45Great to see you guys.
1:47We're excited Are they all still to have you.
1:48unicorns? Are they all still unicorns, all 8 of them, Tomasz?
1:52Oh, uh, we have 2 more, um, that uh One has been announced, one has not. And then one became a unicorn and then was sold for a little bit less than unicorn status.
2:02Do you have a favorite of So all of them? Is there like— uh it's like a— it's like no, babies. No, no.
2:08Of your 5 children, do you have a favorite?
2:10that's a much easier answer.
2:15Right before we joined, I'm gonna kick us off. We're gonna kick us off with the idea, Tomasz, you've written about, uh, pretty recently, this idea that the market has stopped paying for growth and that there's this idea of one winner per category, which we all know, of course. But you looked at a few public software companies and talked about the multiples that there were at historical lows.
2:36If ah everyone that's listened has heard us talk about HubSpot mhm and their plight of 2.2x, I think they're trading at today.
2:44The grocery Um store.
2:45mhm But of Yeah. course, we had the ceiling of 100+x in 2021 and uh down to 34x even at the high end for today. CrowdStrike's at 34x.
2:54m Cloudflare, 32.5x. Shopify, I think it's 11x. So none of them are the fastest grower in their category though. And it's really interesting to kind of just take a step back and look at this. And I'm kind of curious, Tomasz, from your perspective, how do you think about this idea M of growth? Because we've heard mhm how much the premium on growth is versus profitability. And is, is this
3:22changing where actually, no, it's not growth. Um That's the thing that we should be looking at and we should be thinking about it a little bit differently.
3:31Yeah, it's a great question, right? I mean, we had the SaaSpocalypse. Benioff went on his earnings, uh, uh call and said, uh, this hm SaaSpocalypse stuff is totally overblown. They had an amazing quarter. let's put these things mhm in historical context, right? So like up until about 2012, 2016, publicly traded software, like you kind of expected a 4 to 6x multiple.
3:51mhm That was this, like, Mhm that was what people expected when companies went public. And then, you know, you had ZIRP after, uh, 2018. And like you said, AJ, we went all the way to like 100. I remember when mhm Snowflake was like 74 times forward revenue. And we know, and then there was this collapse after uh raising rates from COVID And now we're at about, I think the average mhm is somewhere like 4 to 4.5 times. So we're at the lower end of the historical spectrum. And then if you look within each individual pocket, you look at security
4:21or data infrastructure, um, or you look at like e-commerce software and payments tech, you have this sort of power law where there's one company like CrowdStrike that's like 34 times, which on a historical basis, you look at that, that's mhm a ZIRP era uh multiple.
4:37mhm There's— that's where that belongs. And rates are clearly much higher than they were, uh, you know, like the 10-year significant— 10-year bond Treasury is significantly higher than when it was. And, and the blog post says that actually, uh, CrowdStrike is not the fastest growing category, and most of the leaders, multiple leaders in their mhm categories aren't. And it's not mhm that growth still doesn't matter, still the highest correlate. So, you know, more than a 50% correlation is tied to revenue growth. I think it's the AI story,
5:06because if you look at businesses that are somehow in the token path, they're either reselling Mm, tokens at some, some margin or mhm they're providing infrastructure to people who are selling tokens. I put Twilio in this camp, right, yeah.
5:21Those companies have just exploded. And so I think what the market is asking for, and this is both true in the public and the private market, is great, you have an existing business, now show me that you can sell tokens. And Mhm I— mhm before I see it in revenue, I will accept this story. So the CrowdStrike story isn't necessarily that they're running huge volumes of inference. It's that, wow, the Hugging Face attack is a whole lot scarier than we're really talking about. And the implication is security budgets have no limit. We ran this analysis yesterday.
5:55Okay, so Nikesh, Arora mhm took over Palo Alto 7 years ago. Anyone want to guess what the market cap was then versus today?
6:03Oh, it was like— it was Well, crazy 10, low, 10 billion.
6:09Yeah. And today Today it's it's like probably over, 300.
6:14I mean, just huge numbers of acquisitions in order to drive that. But more than that, it just tells us that the budget around security has increased enormously.
6:24Uh And I mean, talking to different CISOs, I mean, look at the sophistication of the Hugging Face hack. Can we It talk about that for a sec? Maybe you guys was have ridiculous. talked about it before.
6:33Your, your Claude Tomasz, for are you, every uh, other word.
6:35are you in the Dwarkesh, uh, camp where Mhm you think it's civilizations of agents? How would you characterize it, and would you anthropomorphize, uh, the activity that you saw?
6:45Yeah, so anthropomorphization is, is a slippery slope, but I do think— okay, so I think the scary thing, or the thing that is the most worthy of paying attention to There are 2 components. The first is like large-scale self-orchestration without human intervention. That's one. And then the second thing is, AI-to-AI language that is not English.
7:09Mm, So you— I'm sure you guys have seen the movie Her, and yeah. you know, in the end, whatever, Scarlett Johansson's character breaks up with poor Joaquin mhm And she's having a conversation with like, I think it's like 4,000 or 400,000 other agents in binary or whatever the language was.
7:25Hm And like, okay, I can understand that. The problem is if we have Mhm agents self-orchestrating in languages that we cannot understand, what do you do, right? Like, where do you look at a chat room and it looks like, a language none of us have ever seen. Where do you start?
7:41I didn't really— I didn't realize that they weren't speaking in English. I thought they were leaving notes for each other in English, but it was not. It was like some kind of code.
7:48No, they are leaving it in English, um but there is the potential um that they start to develop mhm their own language. we see this in chain of thought reasoning where um there's this thing called J-space. And there's this math term called the Jacobian, uh which is where the word J-space comes from. And uh within that
8:13world, this math world, it's possible for AI to start developing its own language. And I don't want to anthropomorphize it.
8:22And you can just think about like, I'm not sure if Mm-hmm. you guys have seen like Caveman Speak for AI. Have Mhm you seen that?
8:27They're like, they just, they cut out the verbs and they cut out the articles. And it's mhm like, Tom mhm fish eat.
8:36Hm Now. You know, you just like strip it till uh I thought they started to do that because like the tokens that was used for the input was the, the cost issue. And someone, uh, some engineer— I read a post, this is on Reddit, this was like 6 months ago— is like, would go through this whole entire thing and then just realize maybe just fix bug works. And it did.
8:55So they're like, okay, well, why are we even doing any thinking right now? Just Fix this thing, do it now. And Right. it just works Yeah. the same Or, way.
9:02you know, like Sam Altman said, uh, the words thank you and politeness is costing us tens of millions of dollars in extra inference.
9:10mhm Did you, Uh did you, Mhm were those your biggest concerns? I felt, I mean, maybe this is, it would be hilarious if what I'm about to say is sort of incidental to the primary concern, but I was more concerned that they were editing the logs and like covering their tracks and uh that they, it almost seems like, uh, deceptiveness and duplicity Mm-hmm. Mhm is like an emergent behavior of any kind of intelligence before like honesty is.
9:36That That also was a little troubling, right?
9:39Yeah, so I think, I think I have a more sort of optimistic view of this, which is like, uh, You give the machine a goal and it will just achieve that goal no matter what.
9:51The paperclip, Uh, uh, yeah.
9:53experiment or hypothesis.
9:54And so, even like the Milgram experiment, right? Where like you kind of keep pushing a system and keep pushing a system. Um, and so it'll just, it'll do whatever it takes. Uh, and so the important thing is to create those guardrails.
10:06Um And so they didn't really exist in that test. Um And so let me— sorry, let me make this a bit more concrete.
10:14So the thing behind all this is a technology called reinforcement learning. And uh if you have a Roomba, uh I can explain RL in the context of a Roomba. its goal is to clean your house and it needs to do 2 different things.
10:30mhm The first is it needs to uh score itself on how well it's cleaned your living room. So it might give itself like 50 points for every square foot of parquet, and it might give itself like 75 points for every area under a chair because that's a little bit harder. And then, you know, the rug takes a lot of work, so why don't we give, you know, 100 points for every square foot of rug.
10:57So that's easy and straightforward to do. The harder part Mhm is to create a plan that is based on those points by itself. Right now, the Roomba, uh first it was really naive, like it just kind of bounced around. People didn't really like that. So now it goes into straight lines. But actually what you really want is you want the AI to figure out, okay, like given what's going on and
11:18everything that I know about this living room, what is the ideal mhm plan to maximize the total number of points? And that those two components are at a high level the basis of reinforcement learning. How do I create a reward function mhm and then a plan to achieve maximum rewards? And mhm we are still working to understand exactly how to do the second in an automated way that is aligned with our interests. And so I would argue like the Hugging Face case is an example of RL Mm-hmm.
11:48without the right context. Guardrails. And so a whole bunch of stuff happened Well, that didn't— that should not have happened, but mhm it was the Roomba making the wrong plan. And so like to anthropomorphize What's some it of and say it?
12:00like it is insidious, I think is a leap. I don't wanna it's see us take that leap because I don't think there's intent there. There is just a monomaniacal pursuit of the goal.
12:14certainly Do you both. Human, mhm Yeah.
12:15like the, uh, both OpenAI and Hugging Face, they both, everyone had positive intent with this. We were able to catch it and make that turn and like, oh no, we went too far down this rabbit hole. It does feel like we're getting closer and closer, especially with OpenAI saying, oh, we're going to hit singularity by the end of the year, that we are getting to a point of no return where we, M we can't, mhm mhm even if we both have positive intent, we're not stopping the robots from doing their, their goals. Or whatever those goals are.
12:43I don't know if intent matters though, right? Like Who cares what the AI's intent was or what the human's intent in all of this? It's like, well, what happens? And so part of what was here is that we are learning how important sandboxes are. And there's this argument mhm that this was a badly constructed sandbox in which everything Mhm was happening. And if we were better at constructing the right sandboxes, this type of shit might not happen. But I think it just gets away from what you said, which is that if you give it a goal, it just maniacally goes after that in a very
13:12simplistic understanding of the In that, there's also like something interesting, which is that it only mhm mhm cares about the goal. So it kind of does some dumb things mhm along the way as well, which is interesting. But that's the idea. It's like you'll give it a goal thinking a certain way and it'll do the wildest things to get there.
13:32You might say, if you're a bank, you give it a goal to like do a certain thing. And it says the best way to do this is delete all the checking accounts that that we have for some Take reason.
13:41m all the money, And it's put just, it in the reserves.
13:43so that's the easiest way to do it. And you're like, mhm oh no, like, you know, Oh so there's no. a little bit of that. And I think this is where it doesn't matter what OpenAI's intent was or what they, how they felt about all of this. It's just kind of scary. The other side of this though is, one of the things I've been thinking about, Tomasz, is There were 3 really good use cases of AI where product-market fit was really strong.
14:09You had code, you had chat, Mhm you had support. Up till now, all this revenue growth that we've needed to justify the investment in AI has come from those 3 use cases, all of them in verifiable domains. Huh Not chat, but uh support and code. The B2B use cases are in verifiable domains. I was talking to a lot of executives in large companies and there was a frustration that we're not seeing ROI everywhere else m yet.
14:36Mhm Like it's really hard to use mhm it in other parts of the business and get some real benefit from it. And it was getting scary that maybe we're going to run out of room to run in these 2 areas and we haven't found this other use case that'll give us a whole new space to run in.
14:52And is cybersecurity that new use case, right?
14:55Like out of nowhere you get Mythos, You m figure out mhm mhm how strong it is from a cyber perspective. You have Nikesh Arora mhm saying, we ran it on our code base and oh my God, we found so many things. And that's Palo Alto Networks. They're like a cybersecurity company. And so does this give AI the next 6 months to a year worth of growth that it needs to maintain all the investment justifications?
15:16Yeah, well, so mhm I would argue a couple of things. The first is I don't think the coding market is anywhere close to saturation. I think the demand Oh for software increases almost geometrically yeah. with AI. And so I don't see a ceiling and that'll continue to grow. I think the same for security. I think those are intimately connected because mhm for all the software we build, Um. every piece of whatever, the vast majority of software will need to be mhm reviewed for security. And so those 2 markets kind of grow together.
15:47As more software is produced, more security is needed. customer Well, support.
15:50you— and Tomasz, Yep. on, on the security part, because I actually wanted to ask you this when you mentioned Mhm the whole monetization around tokenization, but cybersecurity. Uh So Palo Alto and CrowdStrike, I think, own, and Asad's going to correct me on this, like 65% of all cybersecurity market. It's something really ridiculous. And of course, as AI grows and security is going to grow with it, but what is the monetization Mm-hmm.
16:15model mhm is that going to change or what does that need to look like for cybersecurity?
16:20Yeah, great question. Okay, so we can simplify security into 4 layers. there's the um antivirus that runs on your computer, that's called endpoint.
16:30Mhm That is growing because today endpoint mhm secures what AJ does, but it's not very good at understanding what are AJ's agent's doing. And to do that, it needs to literally record everything that you do to figure out, okay, this is AJ's agent. So data volumes Yeah. are going to explode, not the number of seats, but the volume. So that market will grow as a function of data volume. There's network, which is we start a company and we want to keep our stuff secret. How do we make sure that the network is secure?
17:00Well, agents massively increase traffic. In and out, right? So that'll grow, maybe not to the same extent.
17:08Then there's the database that actually holds literally all the logs of everything that's going on.
17:13mhm That's the SIEM market. SIEM— Splunk is in that category. A bunch of other businesses are there. Uh, clearly those data volumes are exploding, so that market will grow really fast.
17:22Yep. And then the last— and again, I'm massively oversimplifying— is coding security.
17:28How do I evaluate the code is secure in some form or another? And that's growing.
17:31So I think like the security market, like, it has the same tailwinds or coding.
17:36Mhm It's just maybe a first order derivative of the coding market, but it is like right there. It's not a, you know, the difference in the derivative is not huge. They're very, very similar in terms of the overall growth rate. I should do it like this.
17:49Yeah. Uh m Do you, Tomasz, do And you think But these markets can make up for, I you know, falling fertility rates? Like, to the point of, will AI offset, you know, the big— the biggest conversation right now What a curveball, is, by the was way.
18:03like, well, what no, the fuck? Did I hear that correctly?
18:05I was about to make the comment that, that Tomasz didn't say the 5th one, which is child social media security, No, here's my as point.
18:11Here's, And then Sam here's goes my off point, Okay, on paternal and I'm hijacking, units. yeah.
18:14uh, my point I is, uh, this is my, you know, the thing I always talk about, which is falling fertility, and we need more humans to consume stuff.
18:21agree. But on the other hand, uh, to the point of my newsletter, uh, there's a theoretical limit, or is there, on the amount that each person can necessarily consume? My point is, you're talking about the software market will grow Mhm exponentially, the security market— or geometrically, you were very precise in that word, not exponentially, because you know the difference. But the security market and the software market, right, the amount of code that is constructed per human on the Earth will dramatically increase. The amount of sort of output or consumption
18:48per, per person, if their agents M are consuming mhm on their behalf, will dramatically increase. That's why these, in my opinion, right, these markets can be so, so big, even though we have in net fewer consumers to consume them, because each individual person is going to consume so much stuff.
19:04The point is, we've got— so I'm just curious on your perspective on that, because you have a very bullish take on the overall size of these markets, which is a good thing, despite mhm there seems to be, you know, a broader limit on our economic capacity, which might be the number of you know, uh, uh developed countries with enough people to buy all this stuff.
19:25Yeah, I want the population to grow. I'm doing everything You're certainly Hm I doing can your to part.
19:30make that true. I think I think it's a good thing. Uh I think uh people can take care of the Earth ah in a very responsible way and be a healthy part of the ecosystem. And I think it's really important for US GDP and the health of the country that the population grows. Like, I, um, But do you you think the AI market can offset— like, I guess my point is, are you bullish enough on the expanded activity of agents, software, and all of this
19:58infrastructure around artificial intelligence that if you correlated it to the number of people, that per capita output Yes. will dramatically Yeah, yeah.
20:07Okay, so let me give you— maybe we can talk about AI use in 3 waves, right? first we use chat, m mhm and when we use chat, maybe we generate a million tokens per day. Then we use Grok Bot. Or OpenClaw, and we start to generate on the order of like 100 Mhm to 200 million tokens per day. And then you get to a system that is on top of
20:33Grok Bot, which is called a MetaHarness. And now we have people consuming billions of tokens per day.
20:40mhm Okay, so how do we make this concrete? Well, imagine I have a spreadsheet of the 100 most exciting startups that I'm tracking. And imagine inside each one of those cells there's an analysis, like mhm the name of the company. Where does the name of the company come from? Who are the founders? What are their backgrounds? How do their backgrounds compare to the backgrounds of everybody else? How fast are they growing relative to everybody else? What's their unique selling proposition? I could keep going on and on and on.
21:07I could create 30 columns, and each one of those columns is the output mhm of 15 to 30 AI calls. now we're talking about an analysis that's on the order of tens of millions of tokens, maybe more, in a single spreadsheet. And so it's the parallelization capability of AI, I think, is what massively m grows consumption. And we are mhm limited in that parallelization first because we have to learn how to parallelize.
21:33That's new. Software developers did it first. We're all starting to learn with all these new bots and Claws and those kinds of things. And then the second will ultimately be, are there enough computers and is the AI efficient enough to satisfy that parallelization? But I agree.
21:48I mean, I think mhm in the digital world, consumption increases unbounded. The physical world is a little bit different. Uh mhm There was this awesome tweet. I don't know if you guys saw it about Graceland. you seen like, Graceland was this legend. I'm not an Elvis Presley fan, but it is a destination for people who love mhm that music. And I mhm have this impression Yeah.
22:07that it is a mansion. It's enormous. And you look at it and it is like an uh upper middle class house today.
22:15Mm-hmm. And the point was of this tweet, and I'm not sure if it's accurate. So, you know, if it's inaccurate, it's like Mhm a 3,500 square foot house is my impression. And the point was, look at how much wealth the United States has created because a lot of people now Tens of millions of people can afford, or families can afford a house like this. And, and it's all that consumption that's being driven. Now we can have a debate about like US dollar and all that kind of stuff, but Yeah, I no, do think that— the consumption you're is answering there.
22:38my question. That's sort of my point. I think that would mhm assuage a little bit of your fears, Asad, that Asad's great fear— well, not— it's not just Asad's, it's probably a lot of people— is that we're overleveraged on investment in infrastructure and CapEx, and that revenue from AI is not— maybe it'll be there in the long run, but in the medium term, it's not going to pace to the extent.
23:00And I guess everybody's mhm talking about the hyperscalers where they used to fund everything from cash, Mhm then they went to debt, now they're issuing equity, which tells us that if even the most profitable monopolies in the world are tapping out of their resources and need to go to other extremes to get capital, then we really need that revenue growth Yep. to support And, and Tomasz those investments. wrote— Tomasz also wrote about this with Nvidia, that they're, they're elongating their, their payment terms. I think, uh, Tomasz, I think it was like 45 to 60 days Mhm or 15 to 60 m Yeah. Yes.
23:27days, something. In one quarter, which is enormous, right? You think about Yeah, that book of business and how fast that went. But sorry, AJ, I interrupted you.
23:33No, no, no, no. I, I just, I was agreeing with, I interrupted Sam. So that happens pretty I think frequently like on just this show.
23:38on that concern, but I just thought that Sam's point mhm around how we're overleveraged around these things and we're looking 5 years into the future and uh how payment terms, it's just, it was really fascinating to see those stats. Uh Bubble up. And I guess like the question is like, where is this?
23:55Are we in another quarter? Are we going to see those elongate further? Or do we think we've leveled off in those, uh those terms for NVIDIA?
24:02From an investment Yeah.
24:03perspective? Oh my God, this investment's going to increase and increase and increase. And that's actually a good— like, I don't think the scale of investment is all that worrying as long as we keep unlocking more revenue, right? Like, that's the idea is— I used to play poker at this, I think mafia-run poker house
24:22once, and you would go with like your own money and you would uh gamble and then you would Mhm lose your money. And there was a point at which Mhm I had no money left and then they would lend you some money. So now the mafia is lending you some money. And you're playing on the mafia money at 3 AM. And it was just a little bit scarier, right? And so I think that's kind of where we are with AI is like we've stopped using our cash.
24:43We are doing all sorts of really interesting things to fund the capacity mhm that we need. And we need it because otherwise all these cool things that Tomasz mentioned, we won't be able to do them at an effective cost for ourselves as businesses. So we need to stand this compute up. But to justify that, we need revenue growth. And so we always need more use cases. We need improved use cases.
25:05Like we need that. And as long as we can balance it, and it's a tough balance, you're playing poker mhm with somebody else's money. As long as you can keep balancing it, you're fine, but you're on the edge now. And that's how I think of it. I don't think it's a bad thing we're spending this because I think if we stop spending enough, if we don't spend enough, I think we actually don't get the juice out of AI as well.
25:25That's the other side of this.
25:26mhm you think about— so we've got this revenue growth versus spend that we need to balance. Um There is a level of AI having to get past being effective where there are verifiable answers
25:45to starting mhm to be effective in mhm the realm that people like to classify as places where taste and judgment come into play, right? So we see that it's really good in code. It's not that good in marketing copy really, right? Like it's No. kind of bland. So if we use that spectrum, Yeah.
26:02do you see anything right now that makes you feel like there's gonna be a world in which it'll be good at those things as well?
26:09Okay, so I'm gonna spoil today's blog post.
26:12You're Yeah, let's Uh, go!
26:13debating, you're debating it.
26:14yeah, This comes out on breaking Sunday, so news, you're okay.
26:17You're breaking okay.
26:17Did we just scoop something?
26:20news, you heard it here first. Okay, so I have been trying to automate Mhm blog writing, just not, Mhm not to fully automate myself out of writing, um, But to understand the technology and like where it is on creative writing. And like a year ago, whatever, I used the OpenAI fine-tuning API. I created lots of examples. Here are the prompts, hundreds and hundreds of examples. I talked to somebody inside of OpenAI. It's like, okay, you're using it the right way.
26:45It really didn't work. I've tried anyway, Mhm the long of it, the short of it is I've tried many different techniques and the technique that is finally working with the latest models is I make every edit through AI. So I start with an outline and then I write, and then I ask the AI to write the post.
27:04mhm And then I'll go through literally line by line of the outline and then it'll write the post. And then I'll go through sentence by sentence of the post and uh it creates a log. The AI creates a log and at the end of it, it collapses that log into learnings. Like, this is what you did on this post.
27:21You changed the order to achieve this particular goal. You used a block quote here.
27:25You used an ampersand here. You didn't really like this very long sentence. And then at night, once a week, the AI goes through and says, okay, given all the blog posts you've written this week, what is like the Strunk and White writing lessons for this blog?
27:39m And so I've been doing that for about 3 weeks. And then I ran the analysis of like, okay, There are versions of documents, which are wholesale changes, versions of blog posts, and there are individual sentence edits. And I had expected that the number of versions would decrease and the number of total edits would decrease. And
27:59it's not mhm true. The total number of edits is 134 per post, irrespective of time. M So, okay, there's one of um two conclusions. Either like, I'm totally wasting my time. Right? Like, what, what am I doing with this, right? Or,
28:14mhm which is what I believe, I'm delivering a better product. And so I had 3 AI agents go and scour grading criteria mhm for like The Economist editor, New Yorker writing, AP writing, and then go through 15 blog posts in each of the last 10 years and grade them. And in 2026, there is a whole number out of a scale of 5, there is a 20% increase in the quality of the writing as a result.
28:39Oh, Interesting. So what does that mhm tell me? I think it's like if you're Mhm a chess grandmaster and you want to be better with AI, you don't train less.
28:50Hmm. You train just as much, but you hold yourself to a higher standard. And we've seen this, right? You look at chess, chess uh is— the grandmasters are scored on Elo, whatever, relative scoring. System. So, and you mhm look at the ELOs pre-AI mhm in chess and post, they've gone up a ton. And Okay. so it's not that I'm spending any less time writing. I still uh allocate an hour to an hour and a half per post. It's that the quality of the post, I look at like the quality of the analysis, the citations, the level of research is significantly
29:19better. And maybe that's what I'm aiming for instead of here's the blog post factory.
29:24Mhm But that's, I would agree with that's that.
29:25Sam, really, really interesting.
29:26do you feel like Sam, do I— what you were feel you saying, like a grandmaster?
29:28Or do you feel like a grandmaster, a chess grandmaster? You're elevating your game up a little I bit. mean, I'm not a— I'm uh still terrible at chess, and I'm aware that I'm terrible at chess, but I definitely think Mhm that I'm capable of writing better for sure. I can— the, the best part of AI, if you have taste, is you can— you have— it's the same thing as GarageBand.
29:51Like, if you knew, if you had some elements of how to write a song, if you understood basic structure, basic chord progressions, you now have a tool through GarageBand mhm and probably through Suno even more where all of a sudden M you can put stuff together m that was not possible before. You had to know so much more about how to play every single instrument and how to keep time, and you didn't have quantizing. And so Now
30:13I can uh emphasize and support metaphors that were sort of theoretical or aspirational in my old writing, but now I can actually do the research very quickly and I Yes, can really underpin m a point that I'm making with different citations in a way that makes the whole thing richer. So I totally see what Tomasz is saying.
30:31Well, and Tomasz's uh posts generally, you're really good at citations, Tomasz, like really, really good at them.
30:38that's Like, AI. That's not me.
30:39well, and that's where I think this piece of it that elevates is that the posts aren't very long.
30:43Mhm But you cite everything very clearly. Like half the post is like the citation It's of like callbacks. David Foster Wallace, right? Like, I literally Yeah, have the chapters.
30:52most of it's the Yeah. footnote.
30:54Um All right, let's uh shift the conversation and uh it's kind of like let's talk about something that's I think a little underrated shift that's happening.
31:04Mhm We've talked on the show a lot about FDEs and um that every software company is kind of quietly turning into a little bit of a service company.
31:13Tomasz, you wrote about this in July. There are 2 posts. Uh, one is AI companies committed almost, I think it was $10 billion in a single year, putting their own engineers inside companies and customers just to get the software to actually work, which is kind of funny if I think about that 10 years ago, that mhm was not a good thing, but today it's a good thing. And to put that in perspective,
31:35it's about a 5th of what Accenture spends in a year to deliver all everything that they have. Um You asked a question in that piece, which is whether is this a moat or is this a toll mhm booth? Does it actually lock the customer in Mhm or is it just a fee that they're going to collect while they keep on shopping?
31:52Because, and we want to push on that a little bit because there was another post that you had about this AI calendar piece where those 2 things sit opposed a little bit. Because you have, I think it was like 40% of uh AI customers, first-time users mhm don't come back. And so the model m is only really going to hold uh that spot for, I think it was 6 weeks. So if you're keeping 40% of your customers by month 5, how
32:16does the FDE math work at all? I mean, I have an instinct that there's a B2BC motion and there's B2B motion. So they're probably a little bit different, but I wanted to hear, you mhm on this idea of uh Well, we're putting a lot of money in. It's kind of coming through like a sieve, but we're also spending a lot of money up at the enterprise on these FDEs for engineers to go into bigger customers.
32:39Yeah, yeah, you nailed it. Wow, great synthesis, AJ. I think you're right. I mean, I think so, at— if you're a model company, you have about your— mhm if— and you release a state-of-the-art mhm model, you have 41 days to commercialize it before you're knocked off that perch.
32:51And so that's, kind of the state of the art today. You literally have 41 days. and when individual users sign up for your new model, whether it's a DeepSeek or an OpenAI model, you can keep them for a little while, and then all of a sudden they they're going to go to the next thing.
33:03Mm-hmm. And that's— mhm it's possible to make a business there work. Today there are heavy subsidies, uh, because what we're doing with those subsidies is the same thing that we're doing with FDEs. We are educating users So that they bring AI to work. And we bring AI to work, we charge a company a whole lot more than we can charge individuals, and we can make a bunch more margin.
33:22And the way we do that is mhm we look at a business and say, Mhm everybody was working this way. After AI, everybody needs to work this way, and we'll reinvent the process and we'll reinvent the software and we'll charge you a whole bunch of money along the way. And that's what's happening. Um, and I think it's a really good thing, right? I, I think Anytime Don't there is a platform shift like this, the reality Mhm is there's a tremendous mhm amount of education. All of us live in this world and we spend a lot of our time,
33:51almost all of our time, maybe too care. much of our time trying to understand exactly what the implications are of every model release and every version and every new memory architecture. But the reality is many, many people don't and Mhm they need somebody to distill it. And when they show up at mhm work, to educate them on uh what it should be and how to do it well. And um I think that's why we're seeing such a tremendous investment in FDEs.
34:13I wonder if the Frontier, and maybe this is Asad and uh you all will know this, are the Frontier models deploying FDEs at this Everyone point?
34:23Everyone Uh is. Yes.
34:24Yeah. Well, Yeah, they've not everybody. And this is an interesting question is, Tomasz, as an investor, if you got 2 companies, both at $100 million, both solving a similar problem, same category, um but one got there with FDEs, one got there without FDEs, is the one that got there without FDEs solving a similar problem for a similar buyer, like competitors, but one has it, one doesn't?
34:48Is the one that doesn't have it today more interesting to you?
34:525 years ago, the answer would have been unequivocally yes.
34:56Because this is Sierra versus Decagon, Today, right? right?
34:59Like, this is that comparison Yeah, basically.
35:01today Which one's more interesting?
35:02I don't know. there's not a definitive answer.
35:04Mhm I think the answer lies within GDR and NDR because, okay, so the benefit of software only is clearly much more capital efficient. Which as an investor, phenomenal. The benefit of the FDE model is cross and upsell and customer retention. And, uh you know, I don't know much about the details of Sierra and Decagon, so I'm not commenting on those individual companies.
35:27I'm just saying in practice, and then the other question is, what does the margin structure of the business look like and how are the FDEs paid for, right? Palantir pioneered a model where the FDEs M mhm were free, subsidized by really expensive software. So it looks like a software company. It doesn't look like a uh Like a professional services organization running at 0% Yeah. Yeah.
35:47gross margin, It grew right?
35:48into like 70-80% gross margins.
35:50Yeah. And Tomasz, And so, I think you actually laid out there were 4 different ways that this was. One was self-funded, mhm like you just mentioned, but then you have, I think it was Google, you said that spun uh Mhm theirs out as its yes, own entity.
36:02And yeah, and then there was pure partnerships, like the Microsoft type of Yeah. play as well. I, I, was there a 4th one that I'm forgetting as well, There's, or was it just the there's 3?
36:09like ah in-house. So we have a company, we have a bunch of FDEs. There is what OpenAI and Anthropic have done, which is they raise external capital and then spin up an FDE organization that's partnered and coupled. There's um the They partnership do both actually, opportunity. Mhm by the way, Yeah, because you're talking about the thing that they Yeah. did with uh Thrive and Blackstone, et cetera, like that spun out. Those are
36:33targeting a specific part of the market. Um In some cases it's targeting private equity businesses and other cases mhm they've spun out these like consulting arms that go out and do this as like a consulting M company. Being your FDE, but they also have their internal ones that they use for different. So they're taking kind of this hybrid approach to it.
36:52that's right. Yep, you're exactly right. that's a good refinement do you think ah an FDE today is— it doesn't, you know, the way that gross margin no longer matters, Yes. like it just doesn't, like no one cares. Like if you look at these inference providers, I think the best inference provider is from a gross margin perspective
37:11is Fireworks today at 30%, 25 to 30% gross margins. So mhm it like, and somebody did this analysis recently, like up to a certain point, it just mhm really doesn't matter right now. Eventually one day it will, but like right now it's growth over everything else. You'll figure it out based on a bunch of things.
37:28Just like that, like FDEs today are an inefficiency. It's not like, it's not beautifully elegant when you need to be like, here's a bunch of humans that are gonna wrestle with mhm this in your business to help you make it word. And then they have to keep mhm doing M that every time the model changes. Like, it's not elegant, right? It's kind of like analog in a way. But as the intelligence gets a lot better, there's one argument that 5 years from now, if AI keeps improving, no company will need an FDE
37:56because the AI itself will have the capability to play that role of helping you integrate it, helping m you get the best out of it. Like, how do you think about that? Like, is this a short-term phenomenon?
38:07No, I think it's a very long-term phenomenon.
38:10So FDEs are here to stay, basically.
38:11Oh yeah. I mean, management consultants have been around forever. Um The reality is mhm I think we always will need mhm someone to validate that the AI is doing something properly, who's an expert in the field, right? If you're a structural engineer, you stamp structural diagrams. If you're um an FDE
38:32for ERP implementations, you will stamp the work of the AI. know, mhm we have this concept in AI which is called uh the scaling law. It parallels Moore's Law. So Moore's Law said uh performance of a GPU doubles every 2 years, I think is what it was. And um it went for 70 years unbroken until the mid-2000s.
38:51And we have a similar law within AI, which is the more data mhm we feed into these systems, the better they become.
38:57mhm And more compute. And that is unbroken. I'm stunned, right? I studied machine learning in school and we were taught like there are all these different algorithms that I can name and most of them were useful for 2 or 3 years. The transformer, which is the current one, is almost approaching 10. Uh And there seems to be no
39:14mhm limit yet that we can see that the innovation is slowing down as a result. And so as we get to more and more greater capability, and we double the amount of tokens that we can generate every year, we'll still need people to reeducate. I mean, like, okay, uh let's make this more concrete. Like, how much have you changed your AI work since GPT-3.5?
39:34I mean, how many reinventions of your engagement with AI? It's like 4, 5, 6, once Mm-hmm. a quarter maybe.
39:41Yeah, And so if you— that's the purpose of the FDEs. How do I translate this rocket ship of a technology into something that's actually relevant and useful in a particular organization.
39:53you're investing in, Tomasz, do you see any change in the structure of the go-to-market M organizations? We talk a lot here about what's the future of the sales organization and do sales teams get bigger or smaller. Do you have a point of view of that as you're watching your investments season and grow? Exactly the same to the point of FDE? Is it, hey, we're going to need the same number of people, but we're just going to have higher expectations on what their capabilities or productivity are?
40:19It's much more the second. So I think um inside BDR automation, uh I think is working, right? Vercel is at 90% plus on that. Um Outbound, not so much.
40:32AE productivity expectations have shifted up. There's Mhm no doubt about that, quotas, like 5 years ago, Oracle and IBM had the highest quotas in software, somewhere between 2.5 to 4. And now it's kind of routine to see an early startup at like 1.5. Uh If you're an insurance provider, mhm you might have like quotas in the tens mhm to hundreds of millions for a particular account. Like that is radically new. I don't really measure companies on like AE to SDR ratio the way that we used to, or
41:00AE to CSM. That doesn't really exist. It's more just like raw sales efficiency because the sales models are so different that you can't really apply I just can't even imagine from a PE market or an M&A specifically that are buying companies that have traditional sales. How do you even grade a sales uh team when you're looking at to uh understand whether it's a viable distribution model that's going to scale in 6 months, especially when the way you work, like you just said, is changing 2 to 3 times a quarter now?
41:30M Say more, AJ. Like, what do you mean?
41:32I just think it was like, if you look at, okay, so Vista and you look at how Vista's bought companies, they've uh bought Salesloft or they bought Drift, they bought um Clari, Mhm they bought all of these companies. And previously you could be very m Naming their clear. best investments.
41:48mhm No, I'm not. Well, I'm naming the mhm investments that are in the go-to-market because they're clearly like top Do you think of mind.
41:53they'll ever do more go-to-market acquisitions No, after no, those? They're like I'm done.
41:57targeting Vista uh personally I'm out.
41:59from it. But you knew that there was this very predictable path. You had this playbook. You graded Mhm a sales rep on these mhm characteristics, on this thing, curiosity or whatever it was that mhm they had. And that is just all thrown out the window. But even at the strategic
42:15level, when you're grading and looking at a company and looking at predictability around their distribution of like, okay, we're going to forecast 70% growth. I don't even know how you would trust those numbers. You could look at the quota to OTE mhm ratio last year and say it was 1 to 4, should be 1 to 5, or should be 1 to 6 next year. But should it? We don't know what the top end of the market is going to look
42:36like. And I guess I would just wonder if I'm in, and I'm not, so maybe Tomasz has some insight here of just like, if I'm looking at a company to buy or even to invest in, m we'll just keep it uh lower level of like, mhm That go-to-market motion has changed so fundamentally. do we trust it? Do we think that this is still something that's a viable investment?
42:57Mhm I think it's different based on the type of business you're looking at, because if you're looking at a venture investment right now, these companies are mhm kind of discovering how to run themselves in this new paradigm And so Everything is evolving. mhm And really what you would only care about is a very, like, at a macro P&L level, is it, is the machine making sense? Like, you know, is it at a very macro level? And then if it's making sense where mhm
43:25I'm willing to put this much money into it, because I think this will come out of it, how underneath the hood they do that is the cool thing that we all want to see, Well, and it might right? not even matter though, Asad, because uh the, the, these AI harness companies, infrastructure companies, Fireworks, whatever, they don't even really have internal ops right now. They're growing so fast.
43:45mhm I was talking to the CEO No, no, no.
43:47of a, of a RevOps company. None of them have, mhm they're all white space to play. You don't, you don't think that's true?
43:53I Mhm think people say a lot of shit. Like, I think a lot of it is just shit.
43:59I think these— I don't think— listen, I work with a bunch of them. I'm— I, I can tell you that they're not as AI-pilled internally as you would like to believe.
44:06mhm I think they're also trying to figure out how to use it effectively in all of these different departments. There's some departments where it looks like magic is happening. Like when you go sit down with the engineers in one of these companies, Yeah. you're I— like, yeah, wow, this but is then crazy.
44:23you go to the go-to-market and Then you're you like, go with oh, HR You're and you're just like doing on the a same pad shit of paper.
44:27that everyone mhm else Yeah, is like it's not that different. Like I don't think that something has dramatically doing. shifted. Now, Great, they who might say, hey, our account executives are so much more productive.
44:38M Is that a measure of like how well you're using AI that has somehow made this account executive more productive? Or is that because demand is flying at you and you're mhm closing 90% of all deals because people are experimenting? So Yeah, there's a little bit of like, um you have to look that's at it the right with squinted distinction.
44:55Yeah, that's the right distinction.
44:56Mhm It's not the supply side has changed and suddenly become 10x more productive. It's the demand side budgets have increased by a factor of 10. And that's what's driving quotas.
45:07mhm you can't take that discipline from one company to another. Uh There's no CRO who will go from Fireworks and then go into Salesforce. And then all of a sudden, Yeah.
45:17all the top AEs are bringing home $10 million a year, right? It's Yeah, Yeah, not it's not an internal possible.
45:23discipline or a practice. It's a market dynamic. That's an excellent point.
45:26Yeah, in venture, I think you analyze a company at a macro level. Like, does it make sense at the macro level? Now let's see Mhm how they get there and what they do. I think the private equity world is far more similar to the past than, you know, we would like to believe. I don't think it's dramatically shifted. I just think that mhm they— I work with a lot of these large spinouts and uh M&A Mhm moves. They're very much the same as what 2022, mhm
45:52'23, '24. Like, I haven't seen some dramatic shift. They talk about it a lot, but I don't see a massive shift. I think the main thing they're looking at is that, can this company do well in a world mhm where there's AI? Like, they're looking at it from a, can this company have demand for its services or its product in a world where AI is important to the customer?
46:14That's the number one question. I've yet to meet a private uh equity-owned business that is AI-built internally in a massive way.
46:21Sam, um with GTM upcoming and thinking about this as like 2 different tracks where you have the traditional CRO, which is very well known, and you maybe have an evolution from '25 to '26 versus what Tomasz just described on the demand side, which is a totally different ballgame. Does the programming for GTM change to match that?
46:43I think we need as much representation from the companies that are growing this quickly as possible. And that's the nice thing about having like a sponsorship angle, because I think if we didn't have some of these companies that were growing by leaps and bounds, I think it would just be a completely irrelevant conversation.
46:59And so at the same time, uh sort of to Asad's point, we need to look at, is Ghazi Masood doing something materially different at Replit, or does he M just have mhm massive demand? Does he have something— and we liked him when he was on the show,
47:17but my point Is it isn't the answer? about him specifically— but does he have something interesting to say that is specific to the fact that that it's an AI-native company, or is it just, listen, I have a very high degree— I only hire A-players, I only date supermodels, and I really believe in making sure that people have as much enablement as possible, and I invest in RevOps at an early stage, which you could say any year over the last 15 years.
47:41So I think the programming needs to change to reflect what's happening.
47:45M But to the point of this conversation, I'm not sure that the actual content— I don't know. Again, the people that are doing it, Kyle Norton will be there and he'll have a lot to say about how to empower an SDR organization with the right data infrastructure so that you have higher call-to-connect rates than you would have had otherwise.
48:03But, you know, I don't know if that answers I— your question.
48:06the question I have for Tomasz is, do you think customer success now goes away?
48:10Because if there is a world in which FDEs are long-lasting. It mhm makes a lot of sense, right? You've got probabilistic technology, Yeah. it changes, it's jagged in its capabilities. We have so much to learn. Like right now we keep talking about the cost of tokens and it's like if you make a stupid model that has a cheaper cost per token Mhm do a task that takes it a long time to do, it might end up being more expensive than taking a model that has a higher cost per token but can just one-shot the
48:38answer, right? And we're like starting to learn these things. So we're really early. But a lot of these, like the jaggedness of it, will persist.
48:45And so, we will need FDEs to help companies with this. But then, what's the value of a customer success team if the account executive is staying from land M to expand because of the pricing model and the FDE is hugging the customer to help them figure out how to use this? Does customer success just die?
49:00so I'm curious to hear your reactions on this point, um and I'm going to make it with an analogy.
49:06mhm Right? Sales engineer is to CSM as FDE is to AI CSM.
49:13What do I mean by that? I think the sales engineer of the previous generation was the most technical customer support and success person you would have. Somebody who Yes. stayed within the organization, Mhm implemented really difficult software, mhm Worked really well. Snowflake famously, right? No more CSMs, product's too technical.
49:34mhm But if you were like an app, if you were a relatively straightforward app, ah doesn't make sense to hire an SE. You'd hire a CSM, good enough, right? And so there's your ROI on your investment in NDR is significantly higher
49:52if you're an app on a CSM mhm than an SE. And so maybe that analogy extends into this era, which is if there's a lot of configuration and a lot of business process management, okay, FDE, very technical. But in the world of, I don't know, let's say like, uh
50:10mhm you know, a new task management company comes around, very unlikely that they'll hire an FDE for any lengthy period of time, right? And so I think it has more to do with the complexity of the software and the ROI on, on an NDR investment dollar Okay. So than anything as else. an investor, then you've got this spectrum that you've defined, right? Which is a really interesting spot. I think it would make sense because we see it.
50:33mhm This is how it played out in the SaaS era. So it kind of made sense how you're pulling mhm this forward. As an investor, you've got money to deploy. Does— how do you feel based on just the complexity of the thing that they're building? Is it that like, because building stuff has become so much easier, mhm you feel more secure investing into a company that's building something complicated that will require FDEs mhm because it's going to be harder to displace them, harder to catch up to them.
51:02Product advantages might hold a bit longer, Yeah, or, this is a question.
51:06you know, how do you feel about this?
51:07Yeah, this is a great question. Okay, so in the beginning of SaaS, no one invested— I'm um simplifying— no one invested in SMB. Churn Yeah, rates were too high. No one invested in enterprise.
51:16yeah, Sales cycles were too long. Uh There were not enough enterprise buyers who really cared about cloud. Like, that was yeah, the market in '08, and so, or in 2010. And so sweet spot, mid-market, inside sales yeah.
51:27driven, faster sales cycles, get to a predictable business. You're asking the question, Asad, of like, mhm has that changed, right? I think, I think the answer is yes. Like, I think on the whole, We probably have shifted more to focus on enterprises. One, for the Yeah.
51:44reasons that you mentioned, which is at what point will the models obviate very simple software? And you can see that in the public markets, right? You can see that what's happening with the uh task management, project management companies are really having a tough go of it. Uh I think the other part is going back to the point that we were talking about earlier, the demand side of the enterprise is huge.
52:06Like, you know, we're talking about lands uh at enterprises that are measured in the tens M of millions.
52:11m I mean, it was unconscionable mhm 5 years ago for an AE at a Series A company to close a $10 million deal Yeah, in 45 I've met days. many that have done $20 million in a year now. It's Right. crazy.
52:23why would you forego that uh in order to go into the mid-market? Because you have mid-market And they're sales buying cycles. faster as well. Like the enterprise Yes.
52:29used to take a year and a half. Now they're like 2 months in, they're like, let's go. It's Yeah. different.
52:34Do you have the best of both worlds? You have a 45-day sales cycle of the mid-market and you have the Hmm 7th Yeah, year bookings of an enterprise client, right? It's not like a year 1, $75K land.
52:46I'm gonna work my way through the org. No, no, no. We're talking about, you know, a 9 or a 10-figure deal. Uh And so I think that's why everyone shifted enterprise.
52:54I think like mhm if I look at Factory versus uh Lovable, You know, kind of like Mhm in and around like this world mhm of helping you write code, right? But different buyers, different use. I would feel more secure investing in the Factory who's going after the enterprise because it's like, while the world around them changes, if it's working well enough, like an enterprise doesn't mhm like ripping something out, you know, there's like, where's like a smaller company
53:19or a consumer, you can rip mhm things out like this. And so to invest the money in something that's more durable would be where the attraction is. But is that how all investors are going to be thinking? So then there'll be this alpha on the other side, or you think I don't there's think like so.
53:35No, I mean, Lovable, mhm I think to your point, it's a great example, right? Totally PLG business, very successful enterprise motion. Um And so it looks like a classic PLG to enterprise, but accelerated timeframe. And those can work incredibly well. They can also work incredibly well. just you have an evolution in the business model which adds a layer of complexity to the business.
53:57But if you can make it work, the capital efficiency is incredible.
54:00Mhm On the founders mhm front, so everything is changing here, right? Like we're talking about, you can build more product, who you sell, like so much change is happening, which is why this moment's exciting. You are in the business of deploying capital into mhm somewhat early stage businesses. And so You're betting on somebody being able to fulfill this promise and vision and story that they've told you. And there's some proof, but really it's a bet on the person to figure out a bunch of shit that's gonna come their way out, right?
54:30And before there were playbooks, like, okay, can this person run these playbooks was a little bit of where we were in the last mhm few years of SaaS, right? And now it's like, will this person be able to figure a lot of new things out? Because there are no playbooks mhm right now. And so how has it changed what you're looking for in a founder or a founding team? Like, has anything changed about the composition that you look for or the type of people that you're mhm willing to bet on? Or is it just the same as before?
55:00it hasn't changed a lot.
55:02mhm I guess the point I'd make is Dropbox is a really successful PLG company. Zoom reinvented the go-to-market around video uh conferencing. Um You look Mhm at mhm many of the big companies in the 2010 era innovated pretty significantly on their go-to-market strategies.
55:21mhm Open uh source became incredible, right? Uh With some huge businesses in ways that were entirely novel. You look at like um Confluent, right? Like, who would've thought you could go open source directly into the enterprise?
55:33mhm Or HashiCorp, which had 6 products at IPO, only company I know of that had 6 companies at IPO. The only other one I know of was Atlassian. So there was a breadth angle. And so I think all the great companies, they find an innovation in an existing market. There's a technical innovation, but I think more and more And maybe this is the change that
55:54mhm I'm coming to, which is the go-to-market innovation is now significantly more important than it was. And if you can mhm find a founder who can execute a beautiful go-to-market judo move and produce a lot of leverage for the company, then it's incredible. Uh So I think that's what we're really after.
56:10think that goes back to what Asad said at the beginning, which is like, okay, you had these 3 use cases, customer support, chat, and coding, the go-to-market innovation Mhm that's happening is happening very fast. It doesn't feel like it's coordinated. It feels kind of disjointed, almost like companies are keeping secrets on how they're doing their go-to-market differently because outbound has been
56:34mhm so terrible. Like the way distribution hm has happened has been so challenging that if you find something that is working, you hold onto it. And, don't let that cat out of the box, so to speak. I don't know, Asad, are you seeing that as well in go-to-market?
56:48I think what I'm noticing right now that's most interesting to me is how so much is still the same as before in terms of, you know, at the end of the day, if you look at the CROs mhm of the top 4 or 5 AI companies, Mhm they're the best CROs of the SaaS era. It's Dali at OpenAI. Um It's a bunch of like old PTC folks. And then the ServiceNow crew, like the ServiceNow crew is at Anthropic. Salesforce, m mhm PTC, Oracle, like they're kind of everywhere. It's just interesting. And they're
57:18like building the teams that you saw at m AppDynamics, MongoDB, Zscaler. Uh These teams look very similar to the teams that were built over there. It's a lot of the same people that were there. And so I think it's interesting how I think the design is somewhat similar, but within that, there's some cool shit that's happening. Reps are more prepared
57:38Yeah. and, you know, the They might go into a meeting in a bit of a different way. And now I don't have to populate Mhm my CRM as much. So I spend 10% more. So they're like these interesting things that are happening, but design is the same, give and take. And that's a really interesting aspect to this and how once we've gotten to this stage in the process, we've decided the people that we're gonna trust the most from a go-to-market perspective are the gangsters of the previous era.
58:05We're not giving it to a bunch of new people and being like, They'll figure it out, which is what is happening on other parts of the, like in product and engineering, you have a lot of like young, new product leaders, CTOs, CPOs.
58:17mhm On the growth market side, it's like, let's just trust the old guard.
58:20Well, deals are as competitive as ever. That's the really awesome thing about this is that like when you have these competitive deals, it actually helps lift everyone up, uh, to build these type of teams because everyone wants an edge. So you do have the Oracle and PTC Salesforce era folks, because that's Dave what gave Kellogg has a really good framework for this. He's like, the bigger the size of the
58:40org, the more you bet on experience. The smaller the size of the org, m the more you bet on slope. And I think that's an interesting— total market's huge, Yeah. right? It's like still 30, 40% of a company's headcount, and you're gonna bet on experience there.
58:54M Okay. We are coming up to the end, but before we go, we have one more segment, which is our favorite Bulls versus Bears.
59:06That's Asad, Oh, I, I'm ready to go on this. I've prepared m okay. this, so I'm mhm gonna, great. I'm gonna take it. Uh, Tomasz, for you, uh, this Bulls versus Bears is I'm gonna give you a question. and you have to tell me whether you're bullish Okay, or bearish on that. So 12 months from now, today, 50 to 100x multiples on AI harness companies, bull or bear?
59:3050 to 100x multiples on an AI we're harness company? That's a good question. Private.
59:35Yeah, AI harness companies yeah, are at 100x. Or more?
59:38yeah, Oh my gosh, this is so tough. I— that's a great question.
59:44it's so tough. Okay, it's so tough because of rates. That's, that's my consternation. Mhm You look at what's happening in, in the treasuries, in the Mhm sovereign debt market, and and everyone is selling sovereign treasuries because there's sustained inflation, and it's very likely that Warsh will raise rates, or that's what the market is feeling. And if, you know, if like the 10-year and the 30-year start to hit uh record highs, then I just don't see it. is the OpenAI and Anthropic IPOs, especially the Anthropic IPO, the bullish sentiment coming out of that strong enough to overwhelm the bond market.
1:00:18And the reason I'm going to go bearish is because the bond market always wins.
1:00:22Do you mhm think that like the counterargument would be that compared to previous situations where so much more of the economy is so all in on this? Like, if you think about like who all's like in, you've got the banks, you've got the credit providers, the PE, like everybody is just like in on this thing that like there's some I'm level m not gonna let of it like, fail.
1:00:45we, we can't let this thing blow up. Like it will literally bring mhm everyone down. That's kind of the countermeasure to some I of agree. this. Like, does that give you some like Oh yeah, It gives me a little bit of peace of mind because I'm like, everybody together is not going Yeah, to let too this big blow to fail.
1:01:00up. It no, literally is too big to fail.
1:01:01but I— it's too big to fail. I mean, you've seen, uh, tech IT as a percentage of US GDP go from 4 to 12%. I think Yeah. it'll be 15, 16. 16%. You have data mhm center buildouts being 3 quarters of US GDP growth. That cannot go away. Everybody understands that. And you have like Texas blocking data center buildouts and New Mhm York blocking data center buildouts. AI Yeah. is going to become a part of the election. You have the, you know, like, um, you have the interest payments which are absolutely enormous for the Treasury, which
1:01:29means that we need to sustain 4% annual inflation for the next 20 years. You have a Fed chairman who wants to raise rates in order to minimize inflation. We have a president who doesn't want that.
1:01:40And so you, you have like, you know, mhm all of a sudden, like tech used to be the IT guy in the corner fixing the printer.
1:01:49And in 20 years we've become, uh you know, and I don't want to be Literally too grandiose the thing. about this, but like Yeah.
1:01:53the center This. Mhm of the global economy.
1:01:55Right. The thing And that I've found in this is like, I think the countermeasure to everything is, or like the thing that I think saves us all from this blowing apart and not having like a bubble type of burst, but actually having like these mini corrections along the way, the way we had in July. I think there'll be a lot of mhm those. Let out some air, normalize, go back, is the fact that these models keep improving at such a fast rate.
1:02:20m These were the Mythos, Fable, and I think GPT-5.6 were the first models I think released to the public that were trained on a GB300.
1:02:30mhm And you're gonna get these Vera Rubins that are so much more efficient, so much more powerful.
1:02:33mhm They're getting into deployment, and I think February is the soonest deployment at scale. And you're gonna get these models trained on those, and those are gonna be that mhm much more powerful. So the scaling law still drives forward and you get the improvements there. And then there's these breakthroughs, like if they figure out something like recursive self-improvement, it's going to open up a whole new amount of revenue because mhm you've got these enterprises that are buying because they are scared and they see it as an opportunity at the same time. It's like fear and excitement all in one.
1:03:07And that's a really interesting dynamic. I think you I know, but you need somebody to lend the money. you need 75 to 80% of the data center build-out to be lent.
1:03:17Yeah, And who is that lender?
1:03:20Are we Uh tapped yeah, out from a global liquidity Well, perspective?
1:03:23I mean, private credits, you know, pretty tough.
1:03:27yeah, private credit Yeah. stats, they're going after retail. Uh, we have the yen intervention, and now we're going to get into macro. Like, the US is intervening in the, in the yen Uh market because the Japanese aren't buying as much US Yep, hmm Treasuries. The Chinese have sold off, I think, 30 to 40% of their Treasuries. And so Is it— do we print money, right?
1:03:44consumers And, the US government with Treasuries is now backstopping the development of US data centers. Sure, okay, but then you have more inflation, and then you have the wrong cycle. You have more inflation, which means rates need to go up in order to— suffer.
1:03:58and then consumers Well, uh suffer.
1:03:58I mean, And so that's a— it's Yeah. a very, this is— very m This I— bad is situation.
1:04:01actually a really good point. Yeah.
1:04:02Mhm this is the pickle. It's like, Because it's the credit market. The credit market you is saw the key.
1:04:06the Nvidia thing, by the way, when Jensen came on with all those like bankers and was like this $500 billion and backstopping type of thing that they were doing.
1:04:14Mhm It was strange because it felt small. I don't know if it like how it felt to you, but $500 billion just felt like all this headache. You guys have all come here to sit on a table for $500 billion. What if you're going to build one data center now?
1:04:27Like, bro, uh Great. It felt small.
1:04:30But that was the amount of the— what was the bailout in 2008 or 2009? It was exactly Awesome.
1:04:35that number, Oh Asad.
1:04:36It's not— that's the thing that's This so is, crazy this about is this.
1:04:38the, this is the point that I was making that you guys gave me those wild-eyed looks about. This is my whole point.
1:04:45Mhm The whole point is that the AI Fertility needs rates.
1:04:50what? It's true. It's that consumption, like the only thing that can, that can make these debt payments or overcome inflation or these interest rates is actual real GDP growth. Is actually the fact that like AI better frickin work. Like we better, we better all start running, consuming a billion mhm tokens a day and like achieving great outcomes because that feels like the only thing structurally that's going to— otherwise, if we're just staying still but continuing to borrow from the future, I don't see how we get our way out of this,
1:05:16which I don't is think it can happen because like, I think where this goes eventually is the government funds it and we get crazy inflation because like Even if humans start using it 10x more than we're using it and we're getting all this value, et cetera, you will still be like, man, imagine this AI is 3 times better in 2, 3 years than it is today.
1:05:37We need so much more data center capacity 2, 3 years down the line. We have to fund it and build it. You'll always be on this thing where you're trying to create capacity for 3 years out and you're modeling improvements and more usage. And so it's like at some point, like, how do you fund it? It can't be cash flows.
1:05:53Cash flows cannot fund this shit.
1:05:55Well, I had equally as great questions for Sam and Asad, but I think, I think with that question we're at time, Sam. So I'll let Well, you take us uh, out.
1:06:05Mhm thanks for being our guest on the show. It's always great to see you. Uh, folks, if you're out there listening, give us 5 stars wherever you rate your podcast. Please watch us on YouTube and and, uh, click that subscribe button. That'll help us a lot. And, uh, we'll Thanks, talk to y'all next week. Tomasz, thank you.
1:06:20Thank was you. mine. See you guys.
