Gurley says making user-generated content communities work requires knowing 10 to 15 specific nuanced things.
“the nuances that are necessary to make these UGC communities work, there's a list of 10 or 15 things that you have to know how to do.”
Gurley argues Apple and Google don't know how to build UGC communities because it requires specific expertise.
“And it turns out the majority of people don't know how, including really large companies like Apple and Google because it's really hard and it requires a lot of hand holding and hand stitching and building in place a network effect where the atoms start to bounce into one another.”
“Before we ever engage with a customer, they've already given us way more information than any other apparel retailer has.”
Gurley describes Stitch Fix's approach as ultimate recommendation engine that doesn't show customers the choices.
“Our head of algorithm says it's the ultimate recommendation engine because we don't even let you look at the choices. And that level of curation is something that's highly unique,”
Gurley claims Stitch Fix knows 10x more about customers and merchandise than competitors like Nordstrom's and Macy's.
“So you can put us up against anyone, Nordstrom's, Macy's, any of these players, and I am certain that we know 10 x more about our customers and even 10 x more about the merchandise.”
Gurley argues successful AI requires huge datasets on both sides for matching.
“In order to be successful and powerful in these types of areas, you need huge datasets, and you need them on both sides so that you can do the matching.”
Gurley argues Google and Facebook's power led investors to wrongly believe consumer businesses are over.
“I think over the past three or four years, mainly because of the power of Google and Facebook that a lot of people have come to tell themselves that consumers over.”
Gurley warns that if Solana replaces Ethereum, it suggests all crypto tokens are worth far less due to race-to-the-bottom dynamics.
“And that's why someone said to me the other day, if Solana eats ETH, then all of these tokens are worth far less than we imagined because there's just a race to the bottom.”
Gurley suggests if Solana eats Ethereum, the next competitor could make Solana equally uninteresting.
“Yes. But I think there's an equally interesting counterpoint to that, that Solana might be cheaper than Ethereum, but then the son of Solana might make Solana equally uninteresting.”
Gurley finds it suspect that incumbents, not academicians, are leading the charge to regulate open source AI.
“The people that are leading the charge calling for the regulation and calling and some of them raising this question of whether open source should be allowed are the incumbents.”
Gurley observes all voices promoting AI regulation are executives or large investors with tens to hundreds of millions at stake.
“In particular, it's quite notable that all of the loud voices are either executives at these companies and or large investors at these companies.”
Gurley argues regions that shut down open source AI will fail to innovate relative to regions that allow it.
“So if you shut it down in a particular region, that region is gonna fail to innovate relative to the other regions that are out there.”
Gurley reports open source models like Llama 2 are gaining market share among portfolio startups using AI tools.
“What I'm hearing internally from our portfolio is this specifically the ilama too, Sunny sent me another one today, are really starting to gain market share amongst the startups that are using these tools.”
Gurley notes AI companies have raised billions of dollars each at unprecedented levels despite being early stage startups.
“So yeah, you could call them early stage startups, but you could also I mean, who they've raised billions of dollars each, you know,”
Gurley notes OpenAI wouldn't exist if Google's attention mechanism had been patented instead of open-sourced.
“This concept was done inside of Google as an open source concept, immediately copied by all the other players. So OpenAI doesn't exist today if that's patented and controlled.”
Gurley argues AI memory systems create powerful switching costs that lock users in permanently once they rely on them.
“I start relying on one of these things as my memory and I don't have a way to pull that out and jump to something else, I'm stuck.”
Gurley argues AI memory creates switching costs that lock users in to whichever platform solves it first.
“I start relying on one of these things as my memory and I don't have a way to pull that out and jump to something else, I'm stuck. Yes. Like I am hooked, locked, stuck.”
Gurley says AI leaders admit memory is a hard problem requiring expensive nightly retraining per user.
“If you talk to the people that are at the tops of these firms and you say, hey, why can't this thing remember everything I want? And they go, oh, that's a hard problem.”
Gurley says big AI model companies and Vinod Khosla openly advocated for making open source illegal or blocked.
“And then Vinod started basically saying that literally, like, yes, we should block open source.”
Gurley says LLM switching costs rate around two out of 100, showing extremely high promiscuity among models.
“And the per promiscuity Mhmm. Is unbelievably high. Another way of saying that, one one of the things that leads to high valuation multiples is switching cost.”
Gurley rates LLM switching costs at two out of 100, calling it remarkably low.
“if you were to rate these LLMs on their switching cost with zero maybe being none and a 100 being perfect, I'd say they're two. Yes. Like, it's remarkably low.”
Gurley argues LLMs alone won't be winners without durable differentiation enabling monopoly-like pricing.
“I'm not sure that LLMs, like, if that's all you have going for you, I'm not sure that's going to be the winner.”
Gurley says only frontier AI models are defensible while non-frontier models become fast depreciating assets.
“But if you're not on the frontier, man, it seems that these are gonna be really fast depreciating assets”
Gurley argues foundational AI model companies burning $100-200M annually represents poor capital allocation but may be unavoidable.
“100,000,000 a year? 200,000,000 a year? There's no way that's high quality capital allocation from my point of view. There's no chance. But maybe they don't have the alternative.”
Gurley says AI foundational models no longer qualify as startup market due to massive capital and big company involvement.
“It's already evolved to a point that's very similar to where Uber, Lyft and DoorDash ended up, where there's just so much money moving around that I really don't even think about it as a startup market anymore.”
Gurley says AI foundational models are no longer a startup market, with big company involvement being dangerous.
“If you wanted to be accurate, non consensus in a way around these, part of it ties into the big guys being interested in what they're doing with their own balance sheet to reinforce this, which I consider to be remarkably dangerous and unhealthy.”
Gurley argues initial LLM versions are structurally flawed around personal memory and RAG approaches don't solve it.
“I do think that the initial versions of these models are structurally flawed around providing personal memory and allowing someone to become dependent on one of these things.”
Gurley argues institutionalized belief in network effects drives AI capital flood, creating competitive chaos and forced raises.
“Then it's a competitive dynamic. Like once your company raises, you know, 200,000,000, 1,000,000,000, if you're in that market, you raise it too. And it does create chaos.”
Gurley says the $20 consumer subscription concept is gone and was a big part of OpenAI's revenue.
“I think the $20 concept is gone. Yep. And that was a big part of OpenAI's revenue apparently as we believe rumored”
Gurley says Sam Altman warned developers using OpenAI lightly that they'll be steamrolled unless they fully commit data.
“But then he said, if you do that, we're going to steamroll you and you need to bet on us being successful. Which would mean dumping your data and trusting OpenAI more fully.”
Gurley says no one will pay $20 for AI services as long as Meta AI remains free.
“from consumer. And so that's gone as long as Met is free. I don't think anyone pays the $20”
Gurley says entrepreneurs design with cutting-edge AI models but switch to cheaper versions for 20x savings.
“It's like 20x differential. And when I talk to our entrepreneurs that are using these models, they might design with the cutting edge model, but they all back off to the affordable models”
Gurley states LLMs are great at text but horrible at numbers, citing Microsoft CTO.
“LLMs are great at text, they're horrible at numbers. And even Kevin Scott on a podcast he did ironically with Sequoia, that's over this past week, he made that point.”
Gurley suggests LLM models may have diminishing returns on spend, citing MIT professor and other experts.
“Is LLM scaling exponential, linear, or will it diminish? And in addition to the MIT professor, I mean, there are a number of people that have been around these problems for a long time who that the LLM models will have diminishing return on spend.”
Gurley says closed AI model companies must compete with frontier performance while Meta gives it away free.
“If you're a closed model company, the question first is, okay, I'm going to have to compete and keep up with this frontier level competition. That's hard enough. It takes a lot of resources.”
Gurley says Meta may cut Reality Labs spending by 20 percent to reallocate billions toward AI development.
“Remember, Meta had allocated $20,000,000,000 a year to Reality Labs. And there was a report last week that they may, you know, tighten their belt on Reality Labs”
Gurley argues AI wave is unprecedented because incumbents like Microsoft moved fast from the start, unlike typical disruption patterns.
“This is one of the very first waves where the incumbents were eyes wide open at the beginning of the technology shift.”
Gurley suggests incumbents may be aggressively investing in AI specifically to block OpenAI and Anthropic from raising needed billions.
“And now you've got people, you know, trying to project income statements and balance sheets of these companies and they're speculating about whether they might need to raise again or not.”
Gurley suggests leveraging Navy SMR technology with hyperscaler backing as a mini Manhattan Project approach.
“Or for the SMR one by leveraging what's been done in the Navy, maybe getting these hyperscalers behind it. And maybe that's a mini version of a Manhattan project.”
Gurley explains NVIDIA's networking, NVLink, and CUDA advantages emerge specifically in the largest system deployments.
“That's when the networking piece thrives. That's where NVLink thrives. That's where CUDA really comes alive in the biggest systems that are out there.”
Gurley argues NVIDIA's competitive advantage is strongest in the largest systems, not at the edge.
“it appears to me that NVIDIA's competitive advantage is strongest where the size of the system is largest, which is another way of saying what Renee said. It's flipping it on its head.”
Gurley argues NVIDIA's moat is misunderstood as people think a better chip alone could win against it.
“I think there's a notion out there that as soon as someone invents a new chip, a better chip, that they've won.”
Gurley says analysts miss NVIDIA's ecosystem thinking and multi-year planning beyond current architecture.
“I think analysts always focus on the current architectural bet. But I think one of the biggest takeaways from this conversation is that we're thinking about the entire ecosystem and many years out.”
Gurley argues NVIDIA's scaling strategy prepares for distributed training future, not just dependence on million GPU clusters.
“It's not such that you're only dependent on a world where there's a 500,000 or a million GPU cluster. It's by the time there's distributed training, you'll have written the software to enable that.”
Gurley explains NVIDIA's moat comes from a decade building full stack including GPU, CPU, networking, and especially software.
“But the truth is you've been spending the past decade building the full stack from the GPU to the CPU to the networking, and especially the software and libraries that enable applications to run on NVIDIA.”
Gurley argues NVIDIA's moat is misunderstood because it spans the full stack, not just chips.
“one of the most misunderstood things about NVIDIA is how deep the true NVIDIA moat is, right?”
Gurley argues NVIDIA's competitive advantage is strongest in the largest system deployments, not at the edge.
“NVIDIA's competitive advantage is strongest where the size of the system is largest, which is another way of saying what Renee said.”
Gurley argues OpenAI and Anthropic expected continued scaling twelve months ago, not current limits.
“Like, because they were promoting a thesis that it was just gonna keep going up”
Gurley states NVIDIA holds 98% of purchased AI workloads when captive workloads are excluded.
“If you just look at, I guess, workloads people are purchasing to do work on their own, so you take the captives out, you're at 98, right?”
Gurley draws parallel to internet era when companies migrated away from Oracle and Sun as optimization replaced development.
“And five years later, they weren't on Oracle or Sun. And some have argued it went from a development sandbox world to a optimization world.”
Gurley notes AI leaders previously implied linear scaling gains but current reality differs from that impression.
“They had left that impression. And so we get to this place as you described it, it's not quite like that.”
Gurley contrasts Zuckerberg's technical transparency with Altman and Amodei's high-level platitudes, suggesting Zuck may now lead AI.
“Dario and Sam talk in these high level platitudes about how this stuff's gonna cure cancer and we're all gonna not have to work anymore. And Zuck was down in the weeds in the meet.”
Gurley notes China now has four deep-pocketed open source AI models including Baidu's upcoming release, creating competitive innovation advantage.
“this will be the fourth deep pocket funded model in China that are all open source. And when you consider I mean, this is a competitive dynamic, right, that leads to that.”
Gurley warns US policies were creating America-only AI versus China AI dominating rest of world, reversing internet patterns.
“Bingo. And I think had we continued on that path or if we go back towards that path, because I don't think this is over.”
Gurley notes DeepSeek triggered Alibaba and Xiaomi releasing open source models, with Baidu announcing June open source release.
“What happened in China, however, is Alibaba made Twin open source. Xiaomi has a model out now.”
Gurley predicts four well-funded open-source Chinese AI models could create massive optionality versus US proprietary approach.
“And Robin Li of Baidu had his model proprietary and he said in June, it's going to be open source.”
Gurley argues the belief that AI is the biggest platform shift ever is driving current market behavior.
“I think post LLM, the world believes, and I think this is my fifth point or something, but the world believes AI is the biggest platform shift in anyone's lifetime.”
Gurley argues excessive capital availability forces all-or-nothing strategies, with AI companies burning $100-150M annually versus traditional company building.
“Traditional company building isn't spend a 100 or 150,000,000 a year in cash burn, but all the big AI companies are doing that, maybe more.”
Gurley notes OpenAI will burn $7 billion annually, calling it radically different from traditional venture.
“I think OpenAI said they're gonna be 7,000,000,000 in a year. That's not your grandfather's startup business or your grandfather's venture capital. That's a radically different world.”
Gurley predicts four deep-pocketed Chinese open-source AI models will create massive experimentation advantage over US.
“And so that level of competition, if it leads to four deep pocketed all open products is gonna be ultra powerful.”
Gurley contrasts Google's 187,000 employees with OpenAI's 2,700, noting OpenAI plans to stay under 20,000 using AI agents.
“Google has 187,000 employees, OpenAI 2,700. We're not going to be a company of 20,000 employees. He didn't say we're not going be a company of 187,000 employees, Right?”
Gurley discusses slide comparing AI trajectory to reach 75% of total US market cap against historical industrial and transport precedents.
“when will AI reach 75% of total US market cap, which I think is incredibly provocative how you compare that to industrials and transport because everybody's saying it's so big already it can't get any bigger.”
Gurley argues Amazon's lower GPU demand stems from lacking a major consumer AI application unlike Microsoft and Oracle.
“Microsoft derivatively is supporting ChatGPT as is Oracle and CoreWeave on this slide. Amazon doesn't really have a big consumer application, right?”
Gurley notes technology grew from 5% to 15% of global GDP, implying further growth ahead over next decade.
“Said another way, technology, when we got started, Thomas, was 5% of global GDP. Today is 15% of global GDP. And when we're sitting here in ten years, I think you're saying confidently,”
Gurley argues seven to eight deep-pocketed Chinese open-source models can distill each other to accelerate improvement.
“I think there may be seven or eight dig pocketed players with open models in China, they all learn from each other extremely fast.”
Gurley believes OpenAI's long-term success depends more on switching costs than staying at model frontier.
“I continue to believe that OpenAI's most likely chance to long term success comes from switching cost and lock in more than it will come from staying on the edge of the of the model race? Because I think”
Gurley says Mistral should be distilling Chinese models right now if they want to compete.
“I do expect to see that. And if I'm the Mistral team, if you're not distilling on these Chinese models, I don't know what you're doing right now.”
Gurley reports rumors that some best-known AI brands have negative gross margins from pricing below cost.
“There's rumors of even some of the best known brands in AI having negative gross margin.”
Gurley notes Google cannot match OpenAI's competitive advantage of losing seven billion dollars annually.
“I still think it's particularly interesting that Google has to compete with OpenAI because OpenAI is gonna lose 7,000,000,000 this year and Google won't. Like they would never allow themselves to do that.”
Gurley warns NVIDIA backstop obscures real demand signals for CoreWeave from investors.
“That could easily help CoreWeave with their debtors and and getting more debt financing. But it also means as a investor, we we don't know what's going on with real demand for CoreWeave because we probably won't be told if they start moving into the world where they're offloading to NVIDIA or not.”
Gurley reveals NVIDIA promised to buy any CoreWeave service availability they cannot sell to others.
“One of the more peculiar of all the deals is, and this was disclosed in a CoreWeave filing, was NVIDIA has promised to buy any of CoreWeave's service availability that they can't sell to anyone else. That is very unusual. That's not the same as making an investment.”
Gurley warns retail investors that 100x AI returns already happened; current odds are very low.
“And that's not to say there won't be an incremental AI investment that makes money, I think there will. But your odds right now of that being the case are really, really low.”
Gurley argues real technology waves necessarily attract speculation and bubble behavior as pairs.
“If the wave is real, then you're going to have bubble like behavior. Like they come together as a pair precisely because anytime there's very quick wealth creation, you're going to get a lot of people that want to come try and take advantage of that or participate in it.”
Gurley predicts AI will shift from innovation to optimization like dot-com era Oracle-to-Linux migration.
“So there were you went from a period of innovation to a period of optimization where people are much more worried about cost.”
Gurley acknowledges VC's AI-only focus may be rationally correct behavior but creates hardship for non-AI founders with good economics.
“But I I think it's unquestionably true. Like there's just zero interest. And so that's a tough call for someone that's in a business that may have good unit economics that's not AI.”
Gurley suggests we may be in an AI bubble where people are doing speculative things that don't usually work.
“When you're in a bubbly time, and we may be right now with AI, people do silly things, things that don't usually work. They get very speculative.”
Gurley describes how technology waves attract speculators who rush in after seeing quick wealth creation.
“One of the things that happens any time there's a technology wave is people get rich quick and then a whole bunch of people see people getting rich quick and they rush in.”
Gurley says modern venture capitalists refuse to take meetings on anything that is not AI-related.
“There's a reality where a modern venture capitalist does not wanna take a meeting in anything on AI.”
Gurley expresses sadness that venture capitalists are completely uninterested in non-AI companies right now.
“There is a reality where a modern venture capitalist does not want to take a meeting, uninterested in anything on AI.”
Gurley explains tech stock selloffs occur because terminal value collapses if AI threatens twenty-year survival.
“And for most of these high-tech companies, the next five years cash flow are a fraction of their actual market cap, which means all the values in the terminal value.”
Gurley argues open source creates hypercompetitive dynamics politicians don't understand as competitive weapon.
“I don't think the world, especially the the the politicians understand how, like unbelievably powerful open source is as a a as a weapon, like, as a competitive weapon.”
Gurley warns that China's open source AI dynamic creates hypercompetitive weapon politicians don't understand.
“And I don't think the world, especially the the the politicians understand how, like unbelievably powerful open source is as a a as a weapon, like, as a competitive weapon.”
Gurley predicts the US cannot stay ahead of China in AI with two proprietary models versus ten open source ones.
“I think the odds that that, you know, The US with two, you know, proprietary models can stay ahead of China with 10 open source ones is very low.”
Gurley warns Microsoft and Google's AI investments may repeat IBM's mistake of empowering a future competitor.
“I don't think they're dependent on the partner anymore, and it harkens back in my brain to IBM letting Microsoft put the OS inside their PC.”
Gurley argues LLMs contain only existing best practices, so working on the industry edge provides AI protection.
“So if you somehow put yourself on the edge of the industry and are thinking about what's coming in the future, you're more future proof from this.”
Gurley argues humans evolve with their tools, comparing AI adoption to a tractor versus plow competition.
“There's tons of anthropologists that have written about how we evolve with our tools, and you can just imagine a farming competition between a guy with a tractor and some drones and the other guys got a plow and a donkey. Who's gonna win? And you need to think about it that way.”
Gurley cites Cuban's distinction between using LLMs to accelerate learning versus avoiding learning entirely.
“Mark Cuban the other day tweeted, he said he said there's two types of people. The the people that use LLMs to learn faster and the people that use LLMs to not learn at all.”
Gurley claims understanding AI's impact on your role better than peers protects you from layoffs.
“If there are 40 people in your org all doing the same thing and you understand how AI affects that role more than the rest of them, you're not getting laid off.”
“Now, if you contrast that with someone who is a proactive, independent climber, who's trying to build their craft, their world, they're a continuous learner. For that person, AI is a jet pack.”
Gurley argues AI models contain only written best practices, not edge knowledge being discovered today.
“So AI and large language models are recording like all the best practices that have been written down. But the stuff that's on the edge that's being discovered today is not in the models.”
Gurley argues AI captures documented best practices but misses cutting-edge knowledge being discovered now.
“AI and large language models are recording like all the best practices that have been written down. But the stuff that's on the edge that's being discovered today is not in the models.”
“part of what's really cool about an artisan or a craftsman is they're really studying the nuance at the edge of the field. And that's the stuff that's not in the AI model.”
Gurley argues artisans study nuance at the edge of fields which is not in AI models.
“And part of what's really cool about an artisan or a craftsman is they're really studying the nuance at the edge of the field. And that's the stuff that's not in the AI model.”
Gurley argues leading AI model providers seek regulatory protection as biggest threat to their billions is open source movement.
“And so people are putting billions of dollars in their pocket. And so there's also a huge incentive to from the two leading model providers to seek regulatory protection, which most people wouldn't think regulation is but they've been begging for regulation since they started from from three years ago.”
Gurley says AI wave threatens more broadly than previous tech waves like PC, internet, or mobile.
“And the past week in the financial markets, like a whole bunch of companies, you know, traded down because of fear that this would would disrupt them.”
Gurley identifies open source as the biggest threat to leading AI model providers' businesses.
“the biggest threat to their businesses where they're cashing in billions of dollars is the open source movement.”
Gurley says AI leaders want regulation because open source is their biggest business threat
“And the reason I think they want it is the biggest threat to their businesses where they're cashing in billions of dollars is the open source movement.”
Gurley notes AI anxiety polls in China show fraction of US levels because Chinese leaders don't do doomsday messaging.
“Why? I think it's because the leaders over there that are working on these products aren't doing the doomsday stuff that Dario is doing.”
Gurley argues leading AI companies seek regulatory protection while conducting billions in secondary transactions for employees
“Can also though frame Like they're also raising billions and billions of dollars for the They're leading the biggest secondary transactions for their employees in the history of venture capital, the history of the world.”
Gurley says leading AI model providers are conducting the biggest secondary transactions in venture capital history.
“Like they're also raising billions and billions of dollars for the They're leading the biggest secondary transactions for their employees in the history of venture capital, the history of the world.”
Gurley argues leading AI companies seek regulatory protection because their biggest threat is open source movement.
“They're leading the biggest secondary transactions for their employees in the history of venture capital, the history of the world.”
Gurley calls it radical that US AI company leaders are the biggest doomsayers about their own technology.
“The people leading the effort at the top companies here in The US are the biggest doomsayers. It's radical.”
Gurley claims the two leading AI model providers have been begging for regulation for three years.
“And so there's also a huge incentive to from the two leading model providers to seek regulatory protection, which most people wouldn't think regulation is but they've been begging for regulation since they started from from three years ago.”
Gurley explains Microsoft's OpenAI deal involves cashless revenue through cloud credits for equity.
“So you get equity for credits and then those credits are used to run workloads on Azure. That is cashless revenue for Microsoft.”
Gurley calls circular deals horrific and says auditors should not have approved them.
“I think the circular deals are horrific. I don't think the auditor should have approved them. And I think whenever you eventually have an unwinding, they're gonna make it worse because they won't be sustainable.”
Gurley says ChatGPT compared circular AI deals to WorldCom and Enron when given deal structures.
“I just explained the structure of them to chat GPT, and I would encourage anyone to go do this like it's a exercise anyone could do.”
Gurley blames AI founder fear-mongering for data center projects being stopped and regulatory impacts.
“That starts- Some of the founders. Yeah. And it's starting to have ramifications for the industry. A number of data center projects has been stopped.”
Gurley criticizes circular deals where companies convert cash to revenue, saying all major AI players do it.
“You shouldn't be able to move cash from your balance sheet and create revenue on your income statement. I just don't think that should be okay. But they're all doing it.”
Gurley argues high-agency people fascinated by their work find AI accelerates their lives rather than threatens them.
“And in some ways I would say people with high agency that are really fascinated about what they do, their life is accelerated by AI.”
Gurley says young AI companies begging for regulation is unprecedented and threatens idealistic entrepreneurship.
“So there's these new areas there and now AI where very young companies are begging for regulation, which is not anything I've seen in my career either.”
Gurley notes AI fear polling is 20 percent in China versus 80 percent in America.
“And I think you guys mentioned this, but the polling on AI fear in China is like 20% or something like that. And it's like 80 here in America.”
Gurley argues real disruptive waves like AI naturally create bubbles alongside genuine innovation.
“And so I think that's where we are right now. It's real, it's disruptive, it's amazing, it it's changing a lot of different things in industry.”
Gurley says AI companies now burn more than Uber's $2 billion annual losses that he found anxiety-inducing.
“I thought losing 2,000,000,000 a year was nuts and very anxiety inducing, And these companies are doing more than that.”
Gurley argues software requiring deterministic data like financial ledgers is safer from AI disruption than translation-type products AI can replicate.
“It needs to be quite deterministic, and so I think the software that's safer is one where there's a need for deterministic data,”
Gurley argues software with deterministic data needs is safer from AI disruption than pure workflow software.
“I think the software that's safer is one where there's a need for deterministic data, and then you can use that data to reinforce whatever the AI model is, but other people are talking about this.”
Gurley warns US AI regulation could create a fence around America while China serves the rest of the world.
“I think what may happen if you look at what happened with the Internet, there was a global market that the American company served, and there was a ring fence around China.”
Gurley claims Anthropic has lobbied as much early in its life as FTX did, with people in every state.
“Anthropic's lobbying the only company that has lobbied as much as Anthropic early in its life was was FTX and SBF. Like, they he also lobbied heavily.”
Gurley proposes viewing any profession as artisanship, noting people don't typically think of lawyers this way.
“I have this interesting theory where almost any job you could think of it as being an artisan.”
Gurley warns that rote knowledge taught in school is already in AI models, making it an unsafe career position.
“If you just arm yourself with the road algorithms that they taught you in school, those are in the models.”
Gurley invokes Carlota Perez to argue real disruptive waves attract speculators and charlatans, creating bubbles alongside innovation.
“And what happens is with a wave that's real, people get rich quick. When people get rich quick, fools rush in. Like speculators, charlatans, they come in.”
Gurley argues high agency people studying the edge of their field are safer than those with rote algorithms.
“I think people that are high agency and fascinated are, they not only study the history of their field and all that, but they're constantly studying the edge of their field, which is a much safer place to be.”
Gurley criticizes Dario for being both top AI cheerleader and top doomer while raising billions.
“But here, Dario, the guy leading a company that's raising billions of dollars and cashing out billions in secondary to his employees is simultaneously the top cheerleader and the top doomer.”
Gurley says he wouldn't have believed Mag Seven would turn $50-100B annual free cash flow to zero via CapEx.
“If you told me five years ago that these, mag seven would become worth $3,000,000,000,000 and then turn around and take their free cash flow from 50 to a 100,000,000,000 a year down near zero because they were gonna spend it all on CapEx, I'd have been like, no way. Like, I wouldn't have believed it.”
Gurley uses agricultural metaphor to argue Chinese open source AI models create faster innovation through mandatory knowledge sharing.
“imagine you have two societies and both agricultural societies. And one of them, whenever all the farmers come to market, they just sell each other goods and then they go back.”
Gurley uses a farming metaphor to explain why China's open source AI models may innovate faster than closed US models.
“I'll use a simple metaphor, but imagine you have two societies and both agricultural societies.”
Gurley argues that for proactive continuous learners, AI is a jetpack that enables them to achieve more faster.
“For that person, AI is a jetpack. Like they can now do more things faster than they wanted to do, and they can achieve more than they were able to before.”
Gurley explains liquidation preference means common shareholders only benefit after $500M raised is repaid.
“if you've raised in these days of these AI rounds, if you raise 500,000,000, Lake Common doesn't even participate until you get a sale over that, technically.”
Gurley argues that becoming the most AI-productive person in your field makes you indispensable and sought-after for expertise.
“And if you are the most AI productive human in your field, you're not getting fired. Like, you're you're the one they're asking all the questions of.”
“Like, they he also lobbied heavily, but they have people on the ground in every state. They're the ones pushing for state by state regulation.”
Gurley says Anthropic's early lobbying efforts are comparable only to FTX and SBF, with people in every state.
“There's Anthropic's lobbying the only company that has lobbied as much as Anthropic early in its life was was FTX and SBF.”
Gurley warns AI companies are requesting regulation to prevent future competition and pull up the ladder.
“And at the end of that talk I gave, I said I was worried that the AI people were begging for regulation to pull up the ladder.”
Gurley compares Anthropic's early lobbying intensity to FTX, noting ground presence in every state.
“Anthropic's lobbying the only company that has lobbied as much as Anthropic early in its life was was FTX and SBF.”
Gurley argues open models are inevitable in high-stakes software competition, reacting to NVIDIA-Hugging Face acquisition talks at $13B valuation
“Open, open, open. As I wrote in the @washingtonpost, open-models are the inevitable outcome of high stakes software competition. The more at stake - the more likely open wins. This is water running downhill. https://t.co/6x5qoLbw0k https://t.co/wAH0xjej7q”