Josh Brown: If This Is A Bubble, It’s Pathetic

Anthropic

Across the shows: 71 speakers in other episodes, 74 bullish and 57 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  Anthropic's revenue growth is remarkable, and its SEC-filed revenue figures are credible.

  Anthropic's revenue rose from about $1.2B to $16.3B in the first half of 2026 versus the same period a year earlier, a roughly 1300% increase. Off a low base, but still a stunning pace for a three-year-old company, and on track for more than $30B in its third year. Josh acknowledges customer concentration and questions how the ARR calculation works, but argues that because the figures are in an SEC filing, they should be treated as a reasonably accurate reflection of sales. Even bears would have to admit the growth is impressive.

  “So it's off a low base, but it's a 1300% increase for the first six months of this year versus the first six months of last year, the comparable period.”at 4:40

• Josh Brown · Ritholtz Wealth ManagementUncertain

  Anthropic's moat is unproven; durability of its revenue matters more than growth rate.

  Josh argues the real bear concern is not that headline growth will fall from 1300% but whether the platform is sticky. The key question is how easily a next model that is 10% more capable can take away customers and workflows. He says he cannot say definitively whether Anthropic has a moat, and that this is what gives him pause before agreeing it is worth $2T right away. He compares this with SpaceX, Nvidia and Apple, whose moats he calls unassailable. Against that, he notes Anthropic's enterprise business is further along and less price-sensitive than ChatGPT's, and that momentum favours Anthropic.

  “Is there a moat? And I wish I could tell you definitively one way or the other, but that's the thing that might give me pause in saying, yeah, you're right.”at 7:07

• Ed Elson · Prof G MediaBearish

  Anthropic's roughly $800B in compute commitments are not yet in its income statement, creating major cash and survival risk.

  Ed is more worried about profitability than revenue. The filing shows about $800B in compute purchase agreements, including over $160B with Broadcom, $120B with SpaceX and $110B with Amazon, that are not yet reflected in the income statement. He says the company will pay in cash and plans to build its own compute, which will change its cash flow profile. The income statement is therefore backward-looking. He also notes that Ramp data shows 80% of spending on OpenAI and Anthropic comes from 1% of companies, and 80% of those are venture-backed, so the model depends on VC money and a buoyant stock market. He acknowledges Anthropic's cost discipline, but says there is significant risk it fails to keep up with costs and could go bankrupt.

  “that they have signed up to spend $800,000,000,000 on compute purchasing agreements at some point. And we don't really know how long over what time frame that's gonna be spread out.”at 9:03

• Josh Brown · Ritholtz Wealth ManagementBullish

  Anthropic's $80B headline loss is accounting noise; operating loss is narrowing and spending is a land grab.

  The headline loss of about $80.7B reflects how financing and compute are accounted for, so it is not a real cash loss. The number that matters to investors is the operating loss, which fell to $693M from $2.1B a year earlier. Josh likens this stage to a land grab, as John Malone did with cable systems and Bezos did with e-commerce. In such periods, the priority is scale, with profits following. He says that if Dario projected 25% margins now, with OpenAI, Gemini and free open models competing, that would be the wrong move. He sees the battle as potentially a 30-year contest, with year three properly about losing money but making it count.

  “the operating loss is the thing that people in New York or people that are managing money are more focused on. And that number, $693,000,000, down from 2,100,000,000 a year earlier. That is the more important number.”at 10:16

• Josh Brown · Ritholtz Wealth ManagementBullish

  Anthropic will likely raise $100–140B in its IPO later this month.

  Josh says that if you do not think Anthropic will raise $100–140B later this month when it goes public, that would be a huge problem. He thinks it probably will, and he argues that the shareholders who have invested are not widows and orphans or a Scottish insurance company. They are investors like SoftBank, who expect the company to chase scale. His view is that the expected raise is consistent with what those investors want.

  “if you don't think that they're gonna raise a 100 to 140,000,000,000 later this month when they go public, then that's a huge problem. But if you think that they probably will, and I do”at 13:49

• Scott Galloway · NYU Stern School of BusinessBullish

  Anthropic could become the world's most valuable company, possibly reaching $4–5 trillion within four years.

  Anthropic's revenue rose from about $1.2B to $16B in the first half, roughly a 13x increase. Galloway says no company has ever grown revenue this fast at that dollar scale, and that Palantir's 85% growth looks anemic beside it. Anthropic trades at about 60x revenue, but if it reaches a $200B run rate in 2027 or 2028, as estimated, that implies a multiple of about 10x, which looks cheap or reasonable. He argues the market pays high multiples for durable, moat-protected infrastructure businesses, and that the product is unbelievable. He thinks there is a real shot the valuation reaches $4–5T at some point in the next 48 months, especially when Nvidia and Apple have a bad quarter. He would not feel comfortable saying it will be the most valuable company in eight to ten years.

  “I think in the next thirty six to forty eight months, there's a nonzero probability that Anthropic is the most valuable company in the world.”at 19:41

• Scott Galloway · NYU Stern School of BusinessBullish

  Anthropic's gross margins and free cash flow should rise as compute costs fall and scale grows.

  Morningstar estimates about $150B in revenue at a 30% free cash flow margin, implying roughly $45B a year in potential free cash flow within 24 months, or about 44x against a $2T valuation versus Palantir's roughly 150x. Anthropic already shows positive gross margins and is generating early free cash flow at a fraction of the scale it is expected to reach. Galloway says the per-unit cost of compute is falling sharply, so gross margins should rise over 24–36 months as the company scales. It already has the second-highest revenue per employee, behind only Nvidia. He cautions that if the company keeps growing at its current pace, the $500–800B in commitments, some of them firm and non-cancellable, could make it burn cash, and he dislikes adjusted EBITDA.

  “a 30% free cash flow, which it looks like it's charting towards, or 30% free cash flow margin, that's 45,000,000,000 a year in potentially free cash flow in the next twenty four months.”at 23:06

• Josh Brown · Ritholtz Wealth ManagementBullish

  Anthropic's cash position is strong, with most of its capital raised still on hand.

  Per the S-1, Anthropic held $112.4B in cash and equivalents as of June 30, about 80% of the $141B it has raised over six years. It also has about $19B in securities and short-term investments, some of which are money market funds. Josh argues the company has managed its capital well, combining the money it raised with cash generated by the business against the costs of running it. This supports the view that it is not about to run out of money despite heavy spending.

  “They said in the S1, they have 112,400,000,000 in cash and cash equivalents as of June 30. That's 80 percent of the 141,000,000,000 that they've raised over the course of six years”at 26:00

• Josh Brown · Ritholtz Wealth ManagementBearish

  Anthropic lacks its own compute despite planned $900B spending, a structural weakness versus integrated rivals.

  Brown contrasts Anthropic with Alphabet and Meta, which control both their model and compute. He says Anthropic is going to spend about $900B on compute but will not supply any of it unless it spends heavily to build its own. That leaves it dependent on others' infrastructure, which he presents as a disadvantage against vertically integrated rivals.

  “Unfortunately for Anthropic, they're gonna spend $900,000,000,000 on compute. They're not gonna supply any of it unless they spend a lot of money to build it.”at 55:57

Space Exploration Technologies

Across the shows: 35 speakers in other episodes, 30 bullish and 16 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  SpaceX has an unreplicable moat in rockets, launch and Starlink; no real competitor.

  SpaceX is the blueprint for going public at a valuation of $1–2 trillion because what Musk built cannot be replicated: the rockets, the payloads, the cost of getting satellites into space and the Starlink network, which now has about 200 million users. Rocket Labs, the only other public space company, is tiny by comparison, with two launch sites and a few rockets a year. Josh uses this to explain what makes a trillion-dollar company: a moat that is almost one-of-a-kind rather than merely hard to compete with.

  “SpaceX is sort of like the blueprint for how do you go public worth 1,000,000,000,000 or 2. People understand that what Elon has built is not replicable by any other company.”at 21:07

• Scott Galloway · NYU Stern School of BusinessBullish

  The company is priced below Palantir and near Nvidia while growing much faster; attractive on risk-adjusted basis.

  Galloway says the company posted positive adjusted operating income while scaling at 13x growth, which he calls rare, since most such companies burn cash. Adjusted figures exclude stock-based comp, a caveat he flags. He says it is priced cheaper than Palantir and roughly in line with Nvidia while growing much faster than both. Despite existential and political risks, if it sustains even a fraction of its growth it becomes a cash spigot at a multiple that makes Nvidia and Palantir look expensive. He worries more about political risk than financial risk, though Ed Elson counters with the financial commitments.

  “If you look at this thing in the pricing, it's pricing cheaper than Palantir and roughly in line with NVIDIA while growing much faster than both.”at 42:43

• Josh Brown · Ritholtz Wealth ManagementBullish

  The company's IPO is timed as enthusiasm peaks, which should help the first days of trading.

  Brown argues the company is going public just as excitement around it is peaking. He thinks the timing is good, since it bodes well for the first few days of trading. He contrasts this with the staggered-liquidation structures used in Cerebras and then SpaceX, which he says were learned and will apply to the Anthropic IPO. He concludes the company's bankers nailed the timing. He is focused on the near-term trading, not long-run returns, and his point that enthusiasm will be peaking implies more caution beyond that.

  “I think the enthusiasm around this company will be peaking right as it goes public... I actually think they nailed the timing.”at 45:23

Amazon.com

Across the shows: 26 speakers in other episodes, 29 bullish and 6 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  AWS is the winner if neoclouds fail; its backlog and growth are accelerating.

  Brown argues that if debt-funded neoclouds shake out, hyperscalers gain. AWS grew 28% in the April quarter and 37% in the July quarter. It reported $496B of backlog at its last report, up from about $200B three quarters earlier. If neoclouds fold, AWS is ready to serve the demand and its backlog will keep growting. Microsoft Azure and Google Cloud also benefit, and he asks what the stock market cares about more, Nebius and CoreWeave or Microsoft and Alphabet.

  “If you're AWS... AWS hit 28% growth in the April quarter, 37% growth in the July quarter, they reported $496,000,000,000 in backlog for AWS the last time they reported.”at 53:59

• Scott Galloway · NYU Stern School of BusinessBullish

  Amazon's retail unit could become a massive high-volume, reasonable-margin business via robots and AI.

  Galloway names Amazon his big tech pick for 2026 and says AWS will be a serious unlock. He thinks the retail unit is the real story, a collision between industrialized robots and AI. Jassy says the company can double retail by 2032 without one incremental hire, and retail is Amazon's largest business by dollar volume. That turns a large, low-margin top line into one with reasonable margins. Amazon has about a million robots versus 400,000 industrialized robots in America.

  “I think it's return or the empire strikes back around the retail unit because what I see is a collision between industrialized robots and AI.”at 1:00:20

• Josh Brown · Ritholtz Wealth ManagementBullish

  Robots-as-a-service could become a major Amazon business in about three years, like AWS.

  Brown likes Galloway's idea and thinks robots-as-a-service is about three years away. He draws a parallel with AWS, which began as internal infrastructure before being opened to others. He suggests Amazon could offer a fleet of millions of capable robots that customers rent rather than own. He says it is not impossible this becomes important.

  “I like that idea. I think it's three years from now, but robots as a service, the same way that AWS was initially conceived as being for internal use.”at 1:00:48

NVIDIA

Across the shows: 47 speakers in other episodes, 36 bullish and 15 bearish calls, past 30 days

• Scott Galloway · NYU Stern School of BusinessBullish

  Nvidia trades at a suppressed multiple because investors fear its earnings have peaked.

  Galloway argues Nvidia is trading at suppressed earnings, meaning a lower multiple, because investors are worried about where its earnings go after the peak. He uses this to contrast Anthropic's multiple with Nvidia's, implying Nvidia looks cheap relative to its earnings power. The argument is brief and not fully developed.

  “Nvidia, which I would argue is trading at suppressed earnings because [of the] PE [ratio], because everyone's worried about where the e[arnings] goes.”at 20:01

• Josh Brown · Ritholtz Wealth ManagementBullish

  Nvidia's forward P/E of 15 is the lowest valuation in history; it is not in a bubble.

  Brown says Nvidia trades at about 15 times forward earnings, which he believes is the lowest forward valuation he has seen for the stock. He is not claiming it should trade at 30 times, but asks whether 15 is right. If the stock is at 15, any bubble would have to sit in this year's earnings, not in the valuation. He argues the market is pricing Nvidia's growth skeptically rather than in a bubble.

  “NVIDIA's forward PE is 15 times earnings. This is the lowest valuation, I think, in history on a forward earnings basis, you have ever had the opportunity to buy Nvidia.”at 50:45

OpenAI

Across the shows: 28 speakers in other episodes, 13 bullish and 23 bearish calls, past 30 days

• Scott Galloway · NYU Stern School of BusinessBearish

  OpenAI's momentum is fading as Anthropic gains; Galloway is more worried about ChatGPT.

  When asked which of OpenAI or Anthropic has the stronger moat, Galloway says that in software moats are intellectual property, human capital and culture, and that he is more worried about ChatGPT. He argues the momentum is very much with Anthropic and slightly against ChatGPT. Josh agrees that Anthropic's enterprise business is further along and less price-sensitive.

  “Well, the momentum right now is very much for Anthropic and slightly against ChatGPT.”at 30:41

Sandisk

Across the shows: 2 speakers in other episodes, 2 bullish and 0 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  SanDisk is among the cheapest stocks in the market, with a PEG of 0.08; no bubble in memory names.

  Brown names SanDisk as one of the three cheapest stocks in the market, with a price-to-earnings-growth ratio of 0.08. He contrasts this with a bubble, where he says the figure would be around five. The valuation is his evidence that the semis rally is not a valuation bubble.

  “SanDisk has a price to earnings growth or peg ratio of 0.08. 0.08. In a bubble, it's like five.”at 50:17

Micron Technology

Across the shows: 15 speakers in other episodes, 10 bullish and 6 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  Micron is cheap at about 5x earnings with a PEG near 0.1; no bubble in memory stocks.

  Brown argues the semiconductor complex is not in a bubble and cites valuations as evidence. Micron's PEG ratio is 0.10 and it trades at about five times earnings. He uses these as the cheapest names in the market, pointing to how far these stocks are from bubble-level pricing. This sits within his broader argument that semis are driving tech earnings growth while the stocks still trade cheaply.

  “Micron, point one o... Micron is selling at five times earnings.”at 50:29

Alphabet

Across the shows: 36 speakers in other episodes, 26 bullish and 16 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  Alphabet, with Gemini and its own compute, may matter strategically as investors value end-to-end AI.

  Brown says only two publicly traded companies have a premier model and their own compute: Alphabet with Gemini and Meta with Muse. He thinks investors may come to care about a company with AI end to end, one that sells to consumers with its own product and supplies the compute behind it. He says that is not actionable today, but Gemini's position could become strategically important. He contrasts this with Anthropic, which will spend around $900B on compute without supplying any of it.

  “At some point, investors might actually care about a company that's got AI end to end.”at 55:36

Netflix

Across the shows: 17 speakers in other episodes, 9 bullish and 14 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementUncertain

  Netflix's stock slump reflects a time-spent comparison with YouTube it cannot control; the business is fixable.

  Brown bought Netflix at about $100 (down from $140), added in the 80s and 60s, sold half in the 70s before earnings, and says it is now around 63, calling it a slow bleed. He argues the problem is that Wall Street fixates on time spent, where YouTube has about twice Netflix's viewing time and has lapped it. He discounts how meaningful that is, since autoplay inflates YouTube's hours. Netflix is hugely profitable (around $3B a quarter) and expanding internationally, and its weak programming can be fixed with a $20B annual budget. Premium content that viewers actively choose should signal sticky subscriptions and future cash flows, so Netflix should double down on quality. Trying to be YouTube-light with podcasts and video games will not work. He cites East of Eden as the best thing they've made in five years.

  “I think the big problem for Netflix, they're being compared to something that they can't do anything about.”at 1:12:17

Paramount Skydance

Across the shows: 2 speakers in other episodes, 0 bullish and 3 bearish calls, past 30 days

• Josh Brown · Ritholtz Wealth ManagementBullish

  Paramount will double down on HBO premium, must-have programming that can support higher prices.

  Josh argues Paramount has gotten the memo and will double down on HBO. He says it will make premium-quality shows people cannot live without, which lets it raise prices. Podcasts that anyone can watch on YouTube cannot support that pricing power. He cites the hire of an executive from Disney with a long studio-system background and big sports deals. He says the programming must be something people get upset about missing when everyone is talking about it, pointing to Netflix's House of Cards in 2013 as the template for reinventing a platform. He thinks Paramount can do that again. He rejects the idea that a YouTube-style approach is the solution.

  “Paramount's gonna double down on HBO. You'll see premium quality stuff that you can actually raise the prices for because people can't live without it.”at 1:14:40

AI pricing & unit economics

• Josh Brown · Ritholtz Wealth Management

  Per-token AI costs fall, but usage growth means total compute spending keeps rising, showing sustainable demand.

  Josh agrees that per-token costs are plunging, as they have throughout the history of technology. But usage is growing so fast that total dollars spent on compute are not falling; they are rising. He argues this is a sign of how sustainable the AI demand boom is, because the services are so popular and in such demand that usage outpaces price declines.

  “The thing is the usage is going up by so much that the overall dollar amount being spent on compute does not fall, it goes up.”at 32:31

• Ed Elson · Prof G Media

  Falling per-token spend hurts AI labs because their customers' revenue will slow.

  Ed agrees that falling per-token compute costs are good for the labs, since they can get more compute for less. But he worries that per-token spend by the startups buying tokens or LLM usage is also falling, which is already happening. That would mean the labs' revenues start to slow, which is a problem for Anthropic and OpenAI. He says the labs are highly dependent on VC-funded startups.

  “if that per token spend is also going down, which it is, then that is also gonna be a problem because it means that their revenues are gonna start to slow.”at 33:27

AI bubble & tech valuations

• Josh Brown · Ritholtz Wealth Management

  AI and tech are not a valuation bubble; at most an earnings bubble, with semis driving growth.

  Brown argues the market is an anti-bubble because it prices in growth skepticism every day. The only credible claim is an earnings bubble, meaning tech profitability is unsustainable, not a valuation bubble. Tech earnings are estimated to grow 65.6% year over year this quarter, but only 24.9% excluding semiconductors and semi equipment, so semis are the biggest contributors. Western Digital and Oracle beat EPS estimates, yet the SMH semiconductor ETF has had no 52-week highs since July, with no above-average reading in new highs. He calls this inconsistent with a bubble.

  “It is an anti bubble. The only thing that you can credibly say is that it's an earnings bubble... You cannot say it's a valuation bubble.”at 48:35

• Ed Elson · Prof G Media

  Any AI bubble is very small with risks priced in; expects a correction but not a crash.

  Elson agrees with Brown that AI multiples have fallen sharply: Nvidia's has dropped almost 50% since last year, and Micron's has been cut roughly in half. Depending on the definition of bubble, he thinks it is very small, with all the risks of things going wrong already priced in. He says Wall Street is front-running the boom-bust cycle after the 2008 short, so he expects some correction or drawdown. He thinks even a crash would look like nothing compared with past crashes.

  “It's a very, very small one where it seems as though all of the risks of things going wrong are being priced in.”at 52:16

AI infrastructure & data centers

• Ed Elson · Prof G Media

  Neoclouds have precarious debt; many may run out of money, a likely shakeout in that segment.

  Elson says neoclouds, which were popular last year, are now being scrutinized for precarious debt obligations. He asks whether the money will run out and answers that for many of them it will. He expects the neocloud segment to see a correction or drawback as part of the broader AI cycle. He adds that a crash here would look small compared with past crashes.

  “We're beginning to see it with the neo clouds... they have some precarious debt obligations, and will the money run out? And I think for a lot of them, will.”at 52:57

• Josh Brown · Ritholtz Wealth Management

  Data center buildout creates millions of construction jobs; unions gain, so backlash is muted.

  Brown says Anthropic's spending of about $800–900B on compute has to be built, which means millions of jobs. He argues the construction unions have not spoken out against data centers because the buildout guarantees employment and could let them double membership. He thinks this ripple effect means even without stock options for employees, there is work for a million people, with the dignity and pride that come with it.

  “There's a ripple effect anytime a company like Anthropic comes along and spends money the way that they're spending... it's employment for a million people.”at 1:10:25

US stock market outlook

• Josh Brown · Ritholtz Wealth Management

  Large caps keep winning; hardship becomes moat for the biggest companies, which take share.

  Brown argues that the S&P 500 is not really corporate America, since the index holds the best companies in each category, with a minimum market cap near $25B and proven profitability. Compliance costs, energy costs, trade issues and international law are headaches for average firms but moats for the biggest, which can staff experts or circumvent them. Mag Seven companies are nation states with ticker symbols. Over 15 years, the Russell 2000 has lagged the S&P 500, and the NASDAQ 100 the Russell 3000, which reflects those advantages. He says the S&P 500 is driven by earnings from 50 or 60 companies, so the question for investors is whether something is good for the Mag Seven or the S&P 100.

  “The Mag-seven, these are basically nation states with ticker symbols.”at 1:03:22

• Josh Brown · Ritholtz Wealth Management

  S&P 500 likely peaks in December; year-end rally setup as bull market, midterms and Fed hike pass.

  Josh's base rate: when the S&P 500 is up double digits at the midpoint of the year, the year's high almost always falls in December or January, based on about a hundred years of data. He says this is not science but a lot of years of data, and he notes that 2026 is a bull market. October's seasonal volatility has already happened, with real corrections, semis down 29% and some big industrials down 30%. The Fed hike is done, bond yields are backing off and oil is not racing higher. He expects the midterms to pass and investors to be up about 15% by Thanksgiving, and he points to the biggest IPOs in history being priced as part of the setup. He would buy something. His main risk is an exogenous shock nobody sees coming.

  “When the S and P is up double digits through the midpoint of the year, you have an 89% chance that the high of the year will take place in December.”at 1:17:13

Enterprise AI adoption & AI software

• Josh Brown · Ritholtz Wealth Management

  Enterprises are rationing AI token spend and can swap providers, a real bear case for AI vendors.

  Josh describes the 'real bear case' for AI vendors. Companies pushed AI on employees with little control, and token use spiralled: agents ran for days, and Uber blew through its annual token budget in the first quarter. Companies will now impose company-wide AI budgets and rationalise token use, which Josh calls running a business rather than experiments. Open-weight and free models mean there are essentially no switching costs. When quarterly results miss, Fortune 500 finance teams will trace the shortfall back to the money paid to Anthropic and look for a cheaper harness.

  “Number one, there's gonna be some sort of budget on AI use company wide, and they're gonna have to make decisions like, okay, this department gets x dollars in budget for AI.”at 35:19

AI regulation & safety

• Josh Brown · Ritholtz Wealth Management

  Europe or Asia could force US frontier labs out or require EU-conforming versions; fault lines are unclear.

  Brown says the Europeans could decide they don't want US frontier labs in their country, which would require a different version of the product that conforms to EU rules. He says this could happen in Asia too. He thinks the industry is very early in this, and that the true political fault lines are not yet known. He uses this as a reason the market is not pricing in the existential and political risks.

  “the Europeans decide that they don't really want this in their country... There has to be a different version that conforms to the EU, and this could happen in Asia too.”at 44:13

Wealth, inequality & billionaires

• Scott Galloway · NYU Stern School of Business

  AI-driven wealth concentration is a dangerous trend, with gains going to a few employees and owners.

  Galloway says there is a serious concentration of wealth around AI companies. Anthropic has about 5,300 employees, versus 60,000 at Eli Lilly, 100,000 at P&G, and 150,000 at Apple. He cites rents up 24% quarter over quarter in San Francisco and a SpaceX IPO that created 12 billionaires in Texas. He notes that 40% of American households have medical or dental debt and that about 10% of Americans own all the stocks. He calls the spoils going to fewer and fewer talented or lucky employees something to watch closely and potentially very dangerous.

  “But there is a pretty serious concentration of wealth here... a disproportionate amount of the spoils are going to a fewer and fewer number of very talented and also very, very lucky employees.”at 1:08:14

AI & jobs

• Scott Galloway · NYU Stern School of Business

  AI will create more jobs than it destroys over time, though current job creation is uneven.

  Galloway says he is an AI optimist and believes AI over time will create more jobs than it destroys. He cites The Economist's estimate that AI has so far created about a million jobs and destroyed about 100,000. He also notes that the best earnings calls now show headcount cuts, such as Meta growing 23% with 20% fewer employees, which concerns him. He thinks AI will create more jobs than it steals even though the current pattern looks uneven.

  “And granted, I do think that AI over time will create more jobs than it destroys.”at 1:08:25