OpenAI Misses Targets, Codex vs Claude, Elon vs Sam Trial, Big Hyperscaler Beats, Peptide Craze
2026-05-01 spoken.md · speaker-labeled ▶ watch ← E270 all episodes E272 →
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 67 ideas nobody mentioned gave up this week; it applies only when an episode is processed.
Tier crossings
conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green
Who moved the board
each voice's force on conviction this episode — supports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement
What got argued (6 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
Retatrutide's triple-agonist phase-3 data (37lb loss, liver fat -80%, A1C 7.9→6.0, muscle-sparing) makes it the premium upgrade above $50-Medicare tirzepatide — 'the Mercedes to the Honda... that's where they're going to start to make all this money' — with approval possibly late 2026, mid-2027 latest, and an 'avalanche of peptides' expanding the market beyond obesity into fitness and longevity.
Frontier models (Mythos, GPT-5.5 Cyber) can now chain vulnerabilities end-to-end, forcing every code base through a one-time pre-AI→post-AI hardening cycle over ~6 months — 'the hottest part of the market' — with incumbent security platforms (Palo Alto, CrowdStrike) selling the picks and shovels; line-out-the-door demand reported.
This leap from pre-AI cyber to post-AI cyber is going to be a big one-time upgrade cycle... you're going to find all these dormant bugs and vulnerabilities... the Chinese models are going to have these capabilities within approximately six months [so the window is now].
going to look like this big bulky industrial business in five years. I'm not sure that there's a good valuation case ... to just follow the dollars, like a trillion dollars a year going out of the hyperscalers. Where is it going? Just follow those dollars and buy those companies because those companies are already underpriced.
The bull thesis for AI just got validated in a single afternoon... $725 billion in CapEx guidance in 2026 from four companies... more than 2 percent of GDP... AI was 75 percent of GDP growth [last quarter]... AI is now synonymous with the growth of the American economy.
Opus 4.7, the latest Anthropic release, appears to be a bust... they're rolling back to 4.6... Opus 4.7 is rationing compute, reducing thinking time... Anthropic is token constrained... I'm seeing a lot of people saying that they are shifting their coding usage from Opus to GPT 5.5.
I think they're going to be fine. I think this is a multi-trillion dollar company. I think Anthropic is a multi-trillion dollar company. ... Everything in this market is power constrained. The reason that these folks may miss a number ... nothing to do with demand. It is entirely 100 percent due to the supply of the power necessary to generate the output token.
Episode digest
written during extraction and stored in data/extractions/ep271.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
OpenAI misses its 1B-WAU and revenue targets (WSJ) with $600B compute commitments equal to its entire secondary-market value and Friar-vs-Altman IPO tension (IPO odds down to 32%): Sacks' contrarian take — GPT-5.5/Spud is landing while Opus 4.7 'appears to be a bust' with compute rationing ('Sam may end up being right for the wrong reason'; developer mojo shifting to Codex) — a mild oppose on anthropic-ipo; Chamath counters that BOTH labs are multi-trillion companies and every miss is power-supply, not demand ('less than half of announced gigawatts actually being built'; turbine/transformer backlogs; hyperscalers gain leverage; 'Elon and Dario should do a deal tomorrow'). Friedberg's Rule-of-Three market mapping has Google first-or-fighting-first in both consumer and enterprise (attached), plus the MIT pruning paper (10x inference efficiency) as the efficiency frontier — digest. Cyber goes commercial: GPT-5.5 Cyber matches Mythos and is servable; Sacks' one-time hardening-cycle frame plus Jason's Kurtz line-out-the-door receipts spawn the new ai-cyber-upgrade-cycle idea (PANW/CRWD); Chamath teases that 'the best cybersecurity company in the world' (Nikesh/PANW) can penetrate and manipulate every model. Hyperscaler blowouts: $725B 2026 capex guidance from four companies, Google Cloud +63%, free cash flow sacrificed (AMZN FCF -97%) — Sacks calls the AI bull thesis 'validated in a single afternoon' (capex support) while Chamath flags the industrialization of the Mag7 ('not sure there's a good valuation case') and re-ups 'follow the dollars — buy the suppliers' (attached to his capex-equipment-boom-2026). Elon v Sam trial (Judge Rogers) — settlement expected, digest. Son-of-Anton agent deletes a production DB incl. backups — supervision, not scheming (Sacks). Peptide craze: retatrutide phase-3 data mints a new LLY idea (premium-tier upgrade path, approval late-2026/mid-2027). Labels: clean, core four.