Andrej Karpathy Joins Anthropic | SpaceX Files S1: How Does it Trade | Cerebras Smashes Day 1

Anthropic's Valuation and Andrej Karpathy's Role

• Andrej Karpathy has joined Anthropic, a company that is in talks for a $30 billion funding round at a valuation of over $900 billion, nearly tripling from $380 billion in February, with investors such as Greenoaks, Sequoia, and Altimeter, and this development is seen as a significant move in the AI trend 10s.
• The discussion around Anthropic's valuation and funding round leads to a broader conversation about how companies are valued, with some questioning whether AR multiples still matter, and others arguing that they do, especially when considering the company's improving margins 2m6s.
• The conversation also touches on the topic of IPOs, with SpaceX setting June 12th for the largest IPO in history, aiming for a $1.75 trillion market cap and raising $75 billion, and Cerebras smashing expectations with its IPO and breaking the $300 mark 42s.
• The market trends are also discussed, with Data Dog up 31% and Figma up 12%, and the question of what happens from here, with no signs of a short-term crash coming, and the need to reflate and hire thousands of people to avoid social unrest 10s.

Valuation Metrics and IPO Discussions

• The topic of transparency and the potential consequences of a lack of transparency in the tech industry is also raised, with concerns about the impact on democracy and the need for tech leaders to be more open and honest 10s.
• The concept of ARR multiples is discussed as a metric for value, with the idea that if ARR multiples are the right metric, then the company with an ARR multiple of 18x and growing 10x per year would be the best value in the venture universe, which is why smart capital allocators like Green Oaks and Altimeter are investing 10s.
• The reason why Dario and Anthropic would do a deal at a certain price, even if it seems like a good deal for investors, is because it allows them to derisk their business for another year of high burn, with the goal of committing to large investments in compute, such as 6-5 gawatt, which has a total cost of $40 or $50 billion 2m6s.
• The idea of raising capital is discussed, with the notion that companies like Anthropic should keep raising capital well in advance of their needs, because the needs are so great, and it's the highest ROI on equity dollars, and that they will likely raise again before going public 4m30s.
• The difference in approach between Sam, the CEO of OpenAI, and Dario, the CEO of Anthropic, is highlighted, with Sam pushing the valuation to the absolute max, while Dario is more focused on getting a fair deal done quickly, and is willing to trade at a discount to avoid ripping people off and get it done with low drama 6m40s.
• The last two rounds of funding for Anthropic are compared, with the idea that the company intentionally traded at a discount in order to get the deal done quickly and with low drama, with a simple process of sending an email confirming investment and then collecting the money 10m30s.

Anthropic's Funding and Strategic Decisions

• The Anthropic team is discussing funding options, with Amazon offering $50 billion, $20 billion upfront and $30 billion contingent on going public, and Soft Bank offering $40 billion, but requiring borrowing $30 billion, with the team closing on $110 billion, $20 billion of which is clearing now and $30 billion in 6 months' time, depending on the lending market 10s.
• The team is also discussing Salesforce's spending on Anthropic tokens, with Ben Off stating that Salesforce spent $300 million on Anthropic tokens this year, almost entirely on coding, and Jason is asked to share his thoughts on whether this amount is reasonable and how it may change over time 2m6s.
• The cost per engineer is estimated to be around $15,000 to $20,000 per year, which is considered "table stakes" and a relatively small additional cost, with the fully burdened cost for a developer at Salesforce likely being around $500,000 per year, making the $20,000 per year a 4% additional cost 4m30s.
• The team did a survey of 40 portfolio companies and external companies to determine the average spending per year per month per developer, which was around $1,200 to $1,300 per month, with the median being lower, and this amount is considered to be in the "strike zone of normal" 6m40s.
• The discussion also touches on the fact that Anthropic and OpenAI are amazing businesses, with 90% of the products sold to Salesforce as a vendor being unique, and the spending on Anthropic tokens being one of the largest single external spends for a software company, with the amount being astonishing and explaining Anthropic's mediocre rise 10m0s.

Token Economics and Market Potential

• To determine the valuation of companies like Anthropic, it's essential to estimate how much of every knowledge worker's wage and coding wage will be translated into tokens, with a rough estimate suggesting that 5-7% of every knowledge worker salary and 20% of every engineering salary may be required to achieve a trillion dollars in token revenue across Anthropic and Open AI 10s.
• The total addressable market (TAM) for token businesses like Open AI may have been overestimated if companies like Salesforce only spend $300 million on tokens, or they may continue to increase their token spend, with potential consequences for the industry 1m20s.
• To justify the valuations of companies like Anthropic and Open AI, they would need to capture a significant portion of the total R&D spend, which is roughly $240 billion worldwide, with 20% of R&D spend being a key target 2m30s.
• The trend of spending on tokens is likely to continue, with companies like Salesforce expected to increase their spend, and the current numbers may be just the start, with the potential for token spend to reach 20% of R&D salaries 4m10s.
• The cost of using tokens is currently relatively low, with an example given of a company spending only $2,000 per month on direct token costs, but this is expected to increase as the technology improves and more companies adopt it 6m40s.

Claio's AI Integration and Cost Efficiency

• The recent hiring of Andrej Karpathy by Anthropic is also mentioned, highlighting the company's growing presence in the industry, although the details of his role and responsibilities are not discussed in this section 8m10s.
• Andrej Karpathy requires every employee at Claio who is close to product to commit code and use AI or agents to do their job, with the company having built a custom framework to support this, and this approach is seen as effective 10s.
• The cost of running autonomous agents is often overestimated, with the actual cost being significantly lower, such as $257 to run an AI VP of marketing, which is a point made by Andrew, who has experience with implementing such systems 2m6s.
• The use of agentic frameworks and managing models can help reduce costs and make the implementation of AI more efficient, with the most junior product person being able to use these tools, and this approach can help companies like Claio to manage their costs and improve their operations 4m14s.
• The adoption of AI and agents can have a significant impact on companies, with the potential to reduce the number of tokens needed and lower costs, and this can be a key factor in the success of companies like Claio, which is seen as a beaten-down public company software stock 8m14s.

Market Trends and Company Performance

• The growth of companies like Atlassian, Figma, and Twilio, which have seen significant bounces in their stock prices, suggests that there may be more time for companies to adapt to the changing market and implement AI and agents, and this could be a positive sign for companies like Claio 10m30s.
• The comparison between Claio and other companies like Shopify, which has a much higher revenue multiple, suggests that Claio may be undervalued and could be a potential investment opportunity, with the company's stock trading at a significantly lower multiple than its peers 14m30s.
• The discussion revolves around the growth and valuation of companies, particularly in the tech industry, with a focus on how they are perceived and valued by investors, and the importance of showing growth and progress, especially for companies that have had significant funding and resources, such as $300 million a year and 2,000 engineers 10s.
• The conversation touches on the idea that companies, once they lose their initial hype and valuation based on prospects and futures, will be valued on revenue growth and cash flow, making it difficult for them to return to their previous high valuations, similar to how people's perceptions and expectations change as they mature, much like the analogy of being 21 years old 2m6s.
• Data Dog is mentioned as an example of a company that has performed well, with its first billion-dollar revenue quarter, 32% up from its all-time high, and a market capitalization of over $40 billion, demonstrating a more realistic pricing after its initial exuberant valuation 4m42s.
• Figma is also discussed, with its accelerating growth for the second straight quarter, NDR of 139%, and a 2-year high, despite some initial skepticism and criticism, and the conversation highlights the importance of context and perspective when evaluating companies' performance 6m15s.
• The conversation also mentions the difference in valuation multiples between companies, with good performance companies like Data Dog valued at around 17 or 18 times sales, and poorer performers valued at around three times sales, and how this affects their growth prospects and investor expectations 8m10s.
• The importance of being a "good normal company" is emphasized, with the potential for companies to still be worth billions of dollars and achieve significant revenue growth, even if they are no longer the center of attention, and the example of Atlassian is given as a company that has successfully navigated this transition 10m40s.

Figma's AI Integration and Competitive Landscape

• Figma is building software and is a beneficiary of the software explosion, despite losing the vibe coding race, and it has internal AI tools that can improve Figma designs, which is a significant development 10s.
• The company is rolling out the ability to internally make vibe improvements to Figma designs, allowing the agent to look at a Figma design and suggest updates to the workflow or journey, making the building of software more efficient for its audience 2m6s.
• Figma's new ability to genetically improve design is likely to be a big deal for its customers, and the company can potentially get another 50% or more of revenue out of its base, but it needs to ensure that its customers do not leak out to a design flow that bypasses its product 4m30s.
• Lovable is a product that allows users to create working prototypes, and Figma needs to make sure that it does not lose customers to this design flow, as Lovable's design capabilities are limited but it has a native insertion point for Figma designs 6m40s.
• Figma missed the opportunity to create a seamless transition from design to production prototype, which is a significant loss, estimated to be around $500 million, and Replet and Lovable have this feature as a native insertion point 8m20s.

Stock Buybacks and Financial Strategy

• On a separate note, Wick's stock is down 45% since the stock repurchase, and the company now has a $2.2 billion market cap, which is a significant development in the market 10m40s.
• Base 44 has announced that it has reached $150 million in annual recurring revenue, which is considered impressive but is seen as substitution revenue at a high level, as it hasn't materially grown the revenue, and the core business is potentially terminal 10s.
• The same terminal business assessment can be applied to Squarespace, which, along with Wix, is being terminated by two vectors: the ability to vibe code your own website, which is already better in many cases, and the fact that Shopify has destroyed them as low-end competitors, making it not worth using these alternatives to save money 1m42s.
• Wix and Squarespace were previously low-end Shopify competitors, with growth coming from merchant services, payments, and e-commerce, but Shopify has successfully gone upmarket and downmarket, eliminating the need for these alternatives, especially for the low-end market 2m6s.
• Big Commerce, another competitor, has also been destroyed by Shopify, and the low-end market has evaporated, leaving no reason to use these sub-Shopifies, with even Big Commerce being hit hard despite not being low-end 3m15s.
• The core acquisition engine for these businesses, especially for non-e-commerce sites, needs to be customer-focused and low-end SMB, and if the core acquisition engine works, it should be possible to convert customers to new products and make the math work over time, potentially compounding revenue growth 4m30s.
• The stock buyback strategy is seen as a classic investor banker move, buying back stock at a low price, but it is not considered effective, as it may not necessarily be good for the stock in the long run 6m40s.
• When a company's stock is going down, it may be beneficial to keep money in reserve rather than using it to buy back stock, as the stock price can drop even further, and having a large amount of cash on hand can be more valuable for making strategic acquisitions in the future 10s.
• The strategy of buying back stock to offset dilution or stock-based compensation is not a convincing reason, as the only valid reason to buy back stock is if it is undervalued, and buying back stock at high prices is not a clever move 2m6s.
• Companies may use stock buybacks as a way to hold off shareholder activists, by using their cash to buy back shares and taking away the activists' ability to push for change, especially if the company is also undergoing layoffs 4m30s.
• If a company has no better use for its cash and the stock is cheap, then buying it back can be a good strategy, but there is always a risk that management will waste the cash on a bad acquisition, and companies must decide whether to stick with their current business model or adapt to changing times 6m40s.
• The goal is to optimize for 30% operating margins and 10% growth, and if excess capital is available, it should be used to buy back stock or give it back, as this is a common and relatively non-disruptive move, but it may not always be well-received if the stock price subsequently drops, 10s.

Market Uncertainty and Growth Projections

• The stock is currently trading at 1x revenue, which is considered low, and it is uncertain whether the company's stock price will be higher or lower in a year's time, with one opinion being that it will be higher due to the company's potential to operate at 20% operating margin and create more value, 2m6s.
• The company's ability to create value is dependent on its cash reserves and its ability to last for several years without significant revenue growth, and buying back shares at high prices can be regrettable if the stock price drops further, 4m30s.
• The idea that the stock price "can't get any worse" is not the same as it will get significantly better, and the range of possible outcomes has compressed, making it challenging to come up with ideas to achieve 30% growth and a valuation of 5x revenues, 6m40s.
• One possible scenario is that a founder-led company will use its installed base to build a better version of an AI app and sell it to its existing customers, which could potentially lead to a turnaround, with companies like Canva attempting to do something similar, 10m30s.
• The growth of companies like HubSpot, with 300,000 customers, presents an opportunity for startups to develop and sell AI-powered solutions, such as AISDRs, which could potentially lead to significant sales, like 150,000 Breeze AI agents, but it is challenging to predict which company will succeed in this space 10s.

Nebius and Compute Scarcity

• Nebius is growing at an accelerating rate of 684%, raising questions about whether this growth is justified or a sign of a bubble, with concerns about over market exuberance, and the impact of compute starvation on the industry 2m6s.
• The scarcity of compute resources is a key factor in the success of companies like Nebius and Kore, and if compute becomes plentiful, these companies may become commodity businesses, with the potential for some to go bust, while others like Gavin Baker suggest that the slowness of permitting and data center construction may save the industry from overexpansion 4m30s.
• The relationship between the growth of companies like Nebius and data center capacity is crucial, with the question of whether compute remains relatively scarce, and companies that have access to compute, like Nebius and Kore, will continue to do well, while others may struggle 6m40s.
• The industry is experiencing significant growth, with many categories, such as optical connects, on fire, except for traditional software, which is not inflating, and companies like Marll semiconductor and SanDisk are experiencing a resurgence, with the potential for new opportunities and challenges in the market 10m10s.

Market Volatility and Investment Caution

• The possibility of shorting companies like Nebius is discussed, with the argument that there may not be enough capacity in the market, but the challenge of shorting companies three years into the future is highlighted, with the example of Micron, which has experienced many booms and busts, and the difficulty of predicting the future of these companies 12m20s.
• The current state of the technology industry is characterized by significant growth, with every software and technology company, except traditional software, performing well, and this trend is expected to continue due to the large investments being made by hyperscalers and model companies, approximately 3/4 to a trillion dollars a year, with a substantial portion going to companies like Nvidia, 10s.
• The growth of companies like Open AI and Anthropic is dependent on the growth of their customers, primarily software companies, and if these companies do not grow, then Open AI and Anthropic will also stop growing at some point, 2m6s.
• The industry's rapid expansion is leading to increased demand for capacity, with companies like Corweave, SanDisk, and Nebius benefiting from this trend, but it is expected that at some point, the growth will level out and these companies will experience a crash, 4m30s.
• Investor Leo Ashen Brunner has recently shown signs of caution, adding puts across his portfolio and showing restraint in his latest filings, which may indicate that he believes the industry's growth is not sustainable in the long term, 6m40s.
• The decision to invest in companies like Corweave or SanDisk at current levels is a challenging one, and even experienced investors are hesitant to make investments in this space due to the uncertainty and risk involved, 10m20s.
• Some investors, like the one mentioned, are considering investments in seed rounds, but are cautious about the potential risks and are carefully evaluating the prospects of these companies before making a decision, 12m10s.

Deal Evaluation and Investor Feedback

• The decision to not invest in a company can be a difficult one, with the consideration that it might have worked out if given the chance, and the capital markets were accessible for the next three years, with a significant investment of half a billion to a billion dollars required to build capacity 10s.
• Providing detailed explanations to founders can be beneficial, but it depends on the amount of time spent with them, and it's not always necessary to give detailed feedback, especially if it's just a minor meeting, but for deals that have been thoroughly considered, writing out conclusions can be helpful for internal use and to test one's thinking 2m6s.
• Writing out conclusions for deals that have been spent a lot of time on can provide additional feedback and help to not lose insights gained from the 20 or 30 deals that were nearly done, and sometimes this information is shared with the team if it's of interest 4m30s.
• If a deal has only had zero to one meeting with a founder, there's no upside in providing feedback, but if the deal has gone deep, sharing the real reasons why it didn't work out can be appreciated and helpful to the founder 6m20s.

IPO Success and Market Sentiment

• The Cerebrus IPO, which was the biggest US tech IPO since Snowflake, priced at 185 and popped 68% on day one, is seen as a positive sign for companies like SpaceX, and it shows that there is infinite demand for companies of that size and quality 8m40s.
• The current market conditions and the company's technological differentiation make it a good time for an IPO, with the goal of being better than Figma, a company with a significant backlog, but likely not as large as $24 billion 10s.
• The company's product is an extraordinary complex technological product that has exploded in demand, and they have lined up a marquee customer, Open AI, making it a good time to IPO, especially since semiconductors and inference are hot categories 1m6s.
• The IPO is seen as a way to bet on Open AI and Anthropic, with limited options available, such as Coreweave and Nvidia, making it an attractive opportunity, and the company's ability to raise the range and go beyond it is a positive sign 2m6s.
• The success of the IPO is also seen as a positive tell for SpaceX, as people are willing to take risks on companies with high upside potential, and the fact that the company was unable to get a deal done two years ago shows how fickle the market can be 3m42s.
• When considering adding the company to a public book at $300, it's essential to look at the base rate return on IPOs, which is typically negative, and to evaluate whether the company is a long-term enduring company with technological differentiation, rather than just blindly buying on the pop 5m10s.
• SpaceX is set to have the largest IPO in history, with a suspected $1.75 trillion valuation and a $75 billion raise, which is expected to be epic, but the outcome is uncertain, and the S1 filing will provide information about the company's state in December, without including recent developments such as X.AI, the cursor deal, and the Anthropic deal 7m30s.
• SpaceX and Starlink's financials from last year show $18 billion in revenue with a 20-30% growth rate, but the recent acquisition of X.AI, which has a significant burn rate, will impact the company's profitability, and this information will be partially reflected in the S1 filing 10s.

SpaceX IPO and Market Expectations

• The S1 filing will not include forward projections, and the financial information will only cover the company's state as of December 31st of last year, without including the impact of recent acquisitions such as X.AI and the Entropic deal, which will have to be explained by the company's bankers during the road show 1m15s.
• The market's current excitement and appetite for growth stories, particularly with Elon Musk's involvement, may drive interest in the stock, but if the market's focus shifts to cash flow, the story could change, and the company's valuation might be affected 3m30s.
• The potential for retail investors, such as those on Robin Hood, to drive up the stock price is significant, but it's unlikely to result in a 10x increase in valuation, given the company's already large market capitalization, and the mathematical feasibility of such a move is questionable 6m40s.
• The upcoming IPO and the market's reaction to it will be an interesting story to watch, with many investors likely to be drawn to the company's exciting growth prospects and Elon Musk's involvement, but the actual impact on the stock price will depend on various factors, including the company's financial performance and the market's overall sentiment 8m50s.
• The discussion revolves around the potential initial public offering (IPO) of SpaceX, with a valuation of $75 billion and a price-to-revenue ratio of 100 times, which is considered healthily priced, and it will be interesting to see how the stock performs on its first day of trading 10s.

Institutional Behavior and IPO Dynamics

• When institutions buy shares in an IPO, they typically have a price target in mind, and if that target is achieved on the first day, it can lead to additional trading, with the example given being BlackRock potentially buying $10 billion worth of shares with the expectation of making a 40% return over 12 months 42s.
• The IPO of SpaceX could experience a phenomenon where institutions sell their shares after achieving their price target, similar to what happens during IPO pops, and this could lead to a significant amount of trading activity, with the possibility of the stock price increasing or decreasing 1m30s.
• The performance of the IPO could be compared to that of Facebook, which was considered a dismal failure initially but later became a highly successful company, and it is noted that price matters and can have a significant impact on the success of the IPO 2m6s.
• There are varying predictions about the potential stock price of SpaceX, with some expecting it to trade up to $3 trillion or $5 trillion, while others think it could go down, and it is acknowledged that the demand for the stock will play a significant role in determining its price 3m20s.
• The brand loyalty and recognition of SpaceX are also discussed, with some people loving the brand and others hating it, and it is suggested that this could influence the stock price, with some people potentially buying or selling the stock based on their feelings about the company 4m10s.
• The conversation also touches on the topic of Bitcoin and the potential for SpaceX to have a higher valuation than Bitcoin, with one person claiming to have made billions of dollars from investing in Bitcoin and expecting SpaceX to perform even better 5m0s.

Bitcoin and Public Market Comparisons

• The initial trades for a company went the other way due to concerns about its ability to manage a mobile transition, and its stock price traded down over the next six months, highlighting the risks of assuming a guaranteed outcome 10s.
• The company's fundamentals are being questioned, with its stock trading at 100 times revenues, and its decision to sell compute to Anthropic rather than being a Nebius or Entropic, which may impact its growth engine and AI story 42s.
• A bet is made about the company's valuation, with one person predicting it will be below $3 trillion, and another person agreeing that the company's valuation will not reach $3 trillion or more, citing the casinoization of public markets 2m6s.
• Sam Altman has offered $2 million in OpenAI tokens to every Y Combinator startup in the current batch in exchange for equity, a move that is seen as smart and an attempt to earn back mind share and respect 4m30s.

YC Startups and OpenAI Token Incentives

• The offer of OpenAI tokens may impact the valuations of Y Combinator startups, particularly those that are compute-intensive, but it is unlikely to replace the need for human capital and may only take a little bit of the edge off the need for funding 6m0s.
• The investment of $2 million in OpenAI tokens is seen as a real investment that can be used to add value to deals and derisk investments, and may increase the valuations of Y Combinator startups 8m0s.
• The typical post-money valuation might go up to 60, and there is some correlation, but it's hard to predict, with GameStop potentially being higher since they're investing at 100, and the valuation can vary depending on when the deal is made, with earlier deals having lower valuations, such as 20 if done a month before demo day, and 100 if done after, 10s.
• The valuation and investment strategy may cause an anchoring effect, where investors are less likely to take additional funding at a lower valuation, and it may also shrink the size of rounds, making it harder for VCs to invest in YC rounds, potentially slicing the average ownership to 2 or 3%, 2m6s.
• The use of tokens can replace some of the engineering spend and can be used to serve customers, with the question being how much leverage tokens can provide, and how much of a typical YC company's spend can be replaced by tokens, with some estimates suggesting that token intensity could be around 20% of revenue, 4m30s.
• The token spend can be used for marketing, such as giving away tokens to customers, and this can be a significant advantage for startups, especially in the early stages, where token spend can be equivalent to marketing spend, and having a large amount of tokens can allow startups to focus on shipping the best product, 6m40s.
• The availability of tokens can be disruptive to the startup ecosystem, with the potential to change the game, especially if the amount of tokens available increases, and this can have a significant impact on the ability of startups to build and market their products, with some startups potentially being able to burn through 2 million tokens in 12 months, 10m20s.

Anthropic's Capacity and Financial Implications

• Anthropic's capacity constraints could result in a significant loss of revenue, with an estimated $1.2 billion to $2 billion per year, which translates to a $30 billion hit to valuation at an 18 times valuation, 10s
• Having spare compute can alleviate this issue, and if Anthropic has surplus tokens, they can use them to make investments without incurring additional costs, whereas OpenAI's investments will be valued at 1x, 42s

Jason Rory and Public Perception

• The discussion touches on Jason Rory's celebrity status at Zaster, where he was well-received by the audience, with a large crowd gathering to hear him speak, 2m6s
• Rory's skills and insights are acknowledged, and it is suggested that they justify a 6x fund or at least premium carry, with some even proposing a $199 fund, 4m10s

Elon Musk and OpenAI Legal Disputes

• The conversation also mentions the OpenAI and Musk lawsuit, which was dismissed on a technicality, with the jury returning a verdict quickly, and the discussion touches on the potential implications of the statute of limitations in the case, 6m30s
• The case involving Elon Musk and the conversion of a company to a for-profit entity was likely fraudulent, and the claim that Elon was unaware of the conversion is considered ludicrous, given that he was discussing it in the past 10s.
• The statute of limitations for a fraud claim is relevant in this case, and if Elon had only found out about the conversion in 2023 or 2024, his claim could have proceeded, but it is clear that he knew about it earlier 2m6s.
• Elon's decision to pursue the case was likely motivated by a desire to damage the other side, rather than a probability of winning, and he is willing to spend millions of dollars on legal fees to achieve this goal 4m30s.
• The outcome of the case is seen as justice being served, with Elon getting his "pound of flesh" and the other side being damaged, and it is expected that the case will not proceed further on appeal 6m20s.
• The case has also led to other investigations into Sam Altman's finances, creating more problems for him, and it has been noted that Sam has not taken any equity in Open AI, but may have indirect ownership interests through other companies 8m10s.
• Sam's decision not to take equity in Open AI and his statement that he has no economic interest in the company may be seen as factually incorrect, given his indirect ownership interests, and this is being used against him in congressional testimony 10m40s.

Sam Altman's Governance and Legal Issues

• The complex nature of Open AI's structure has led to issues, and if its founder had taken a simpler approach, such as owning 10% from the start, it would not have attracted attention 10s.
• Sam Altman, the founder of Open AI, set up a venture fund on the side, which allowed him to keep all the carry, and this move is seen as evidence that he did not have the best intentions for the company 2m6s.
• The idea that Sam Altman took no consideration from Open AI is disputed, as he established a venture fund without telling the board, which likely contributed to his firing, and he kept all the carry from this fund 2m6s.
• The setup of the venture fund suggests that Sam Altman did not believe Open AI would be worth anything as a nonprofit, so he wanted to monetize it by keeping all the carry from the investments made by the fund 2m6s.
• The complex arrangements made by Sam Altman have ultimately led to his downfall, as they have attracted the attention of Elon Musk, who is willing to take action against him, and have also led to a negative perception of the AI industry as a whole 4m30s.

Legal Outcomes and Industry Repercussions

• The jury in a related case did not find in favor of Open AI because they found the company more sympathetic than Elon Musk, but rather because they followed the law, and it is likely that the jury had a negative view of all parties involved 6m40s.
• The AI industry is facing a backlash, with leaders such as Eric Schmidt being booed, and this is likely due to the fact that the industry has spent years warning about the dangers of AI, which has led to a negative perception of the industry and its leaders 8m10s.
• The combination of brilliant scientists who are poor at politics and the increasing backlash against the AI industry is likely to lead to significant challenges for the industry in the next three years 10m0s.

AI's Societal Impact and Public Sentiment

• The current state of AI and its impact on society is a topic of concern, with some people expressing regret over the lack of transparency and the potential negative consequences of AI development, such as job losses, 10s.
• There is a shift in the message being conveyed by AI representatives, with some emphasizing the positive aspects of AI, while others are more honest about the potential job losses, like the CEO of Standard Charter Bank, who referred to job losses as "job reductions in favor of the machines", 2m6s.
• Several major companies, including Meta, Cisco, and LinkedIn, have recently announced significant job cuts, with Meta laying off 8,000 people, which may be related to the increasing use of AI and automation, 4m30s.
• The trend of job losses due to AI and automation is expected to continue, and there is a need for tech companies to create policies to rehire people who have lost their jobs and to address the social obligation to mitigate the negative impacts of AI, 6m40s.
• Some experts believe that AI may not be capable of replacing as many jobs as predicted, and that the efficiency created by AI might be less than expected, which could lead to a scenario where the impact of AI on employment is not as significant as thought, 10m20s.
• The conversation around AI and its impact on society is ongoing, and there is a need for tech leaders to consider the social implications of their actions and to work towards creating a more positive and equitable future for all, 12m10s.

Tech Layoffs and Political Implications

• The tech industry is experiencing significant layoffs, with companies citing AI as the reason, and this could lead to massive social and political implications if 20 to 50% of white-collar jobs are automated in the next 5 years 10s.
• The impact of these layoffs could be severe, with former employees potentially voting against the interests of tech companies and their founders, such as Mark Zuckerberg, in future elections, including a potential wealth tax 2m6s.
• The politics surrounding layoffs and wealth distribution become more emotional and less rational when people are directly affected, and the current layoffs could lead to a significant backlash against the tech industry 4m30s.
• The layoffs in the tech industry are expected to be particularly severe, with no other companies likely to hire the laid-off employees, leaving them feeling angry and disenfranchised 6m20s.
• There is breaking news that OpenAI might file for an initial public offering (IPO) as soon as Friday, which could indicate that the company is preparing for a significant change in its financial situation 8m40s.

OpenAI's IPO and Market Developments

• The potential IPO filing by OpenAI is seen as a sign that the company is moving forward with its plans, and it may be followed by other developments, including a potential filing by SpaceX 10m10s.