The Twenty Minute VCAnthropic's Fable Banned by US Government | Wix & Adobe Hit All-Time Lows | Mistral Raising at $20BN
CHAPTERS
- 0:00 – 1:15
SpaceX IPO pop, tiny float, and why early price action can mislead
The episode opens with reactions to SpaceX’s record-setting IPO and rapid post-IPO price increase, including discussion of how much of the move is “real” versus technical. The group stresses that with only a small portion of shares trading, the stock can be pushed around until lockups expire, so today’s mark may not reflect long-term value.
- 1:15 – 3:57
Gamma squeezes explained: options-driven forced buying
Everett breaks down what a gamma squeeze is and why SpaceX is especially exposed now that options trading has begun. With heavy call buying, market makers hedge by buying stock, potentially creating a self-reinforcing loop in a thinly traded name.
- 3:57 – 9:12
Can you profit from being bearish? The brutal math of puts and volatility
Rory explores what it would take to bet against SpaceX and why it’s so hard when implied volatility is high. Even if you believe the stock will be lower in six months, the cost of put options can require a large drawdown just to break even.
- 9:12 – 12:01
Elon’s "long-dated call option" fundraising playbook
Everett reframes Musk’s career as repeatedly selling the market on ambitious, long-horizon upside narratives that function like long-dated call options. The thesis: fundamentals matter, but Musk’s ability to set and then deliver on massive future bets lowers cost of capital and sustains premium valuations.
- 12:01 – 20:27
Solving the compute problem twice: acquisitions + mega contracts
The discussion highlights Musk’s speed in addressing compute and revenue gaps through both acquisition/retention mechanics and large customer contracts. They argue the AI-related revenue run rate could rival or exceed SpaceX/Starlink’s, underscoring how the company is being re-framed as an AI story.
- 20:27 – 29:11
Claude Fable banned: what happened and why trust broke down
Anthropic launches Claude Fable to strong reviews, then faces a rapid US government ban under an export restriction framework. Rory and Everett emphasize conflicting accounts, poor communication, and the political reality that Anthropic’s prior “danger” messaging made it vulnerable once a cyber-use incident appeared.
- 29:11 – 33:12
Regulating models by capability: a "Rubicon" moment for AI access
Everett argues the key issue isn’t Anthropic specifically but whether the US has crossed into regulating AI models based on capabilities and restricting access to “intelligence.” The group explores geopolitical implications if superintelligence access becomes gated by country or citizenship.
- 33:12 – 38:09
Benchmarks are misleading: test-time compute and the coming inference explosion
The conversation shifts to why leaderboard-style benchmark cards may be the wrong mental model. With enough test-time compute (inference), models can improve significantly, implying demand for tokens and inference infrastructure could surge again—even without brand-new “better” models.
- 38:09 – 56:00
Anthropic IPO odds after the ban + practical internal fallout
They debate whether Anthropic will still IPO this year, concluding probability remains above 50% given market conditions, but the ban is a material wrinkle. Rory also notes the operational/legal disruption if non-US staff must step away from restricted systems under export rules.
- 56:00 – 58:24
Mistral at $20B and the reality of sovereign AI
With Mistral rumored to raise at a $20B valuation, they examine whether sovereign models will proliferate. The conclusion: sovereignty demand is real, but building truly competitive models is exceptionally hard; many regions may default to fine-tuning open weights rather than training frontier models.
- 58:24 – 1:08:21
Salesforce buys Fin (Intercom): the blueprint for pre-AI SaaS survival
Salesforce’s acquisition of Fin is framed as a landmark example of a legacy SaaS company successfully reinventing around AI and outcome-based pricing. Rory and Everett describe it as a ‘golden path’ that turned a stagnating asset into a major liquidity event and a playbook for other boards.
- 1:08:21 – 1:11:20
Wix guidance cut and the SaaS sorting mechanism: good vs bad AI exposure
Wix’s reset becomes a case study in how public markets now separate SaaS winners from losers based on AI defensibility. Everett proposes a framework of positive vs negative attributes (usage-based, AI tailwinds, share-gain potential vs replicability, seat pricing, incumbent risk), explaining why Wix is being punished despite efforts like Base44.
- 1:11:20 – 1:16:47
Adobe crash despite beats: cheap multiples, CFO exit, and why money flows to semis
Adobe is discussed as a value trap candidate: trading at strikingly low cash-flow multiples but facing multiple “bad attribute” flags (incumbency, seat model, replicability fears). The group adds a capital-allocation lens: hedge funds are dumping SaaS to buy semiconductors because the AI capex wave is easier to ride than SaaS turnarounds.
- 1:16:47 – 1:28:09
Standard Bots $200M and the robotics debate: humanoids vs pragmatic arms
Closing discussion turns to robotics, using Standard Bots as an example of a pragmatic middle path between brittle legacy arms and expensive humanoids. They emphasize that real-world deployment is full of edge cases (‘reality has detail’), but AI makes robots more adaptable, potentially accelerating adoption over the long run.
