AI First Pod

Claude Autonomously Flagged a Previously Unknown Enzyme System That Resembles CRISPR. This Is AI-Initiated Science.

Anthropic's new life sciences group disclosed Claude's first autonomous discovery: array-associated reverse transcriptases in an uncharacterized bacteriophage system with structural similarities to CRISPR. Claude flagged it. Human researchers didn't assign it. We cover what this means for AI in science, Australia's PM revealing an OpenAI Medicare breach at the UN, and Amazon opening Seller Central to AI agents.

GPT-6 Luna Is $0.10 Per Million Tokens. Opus 5.5 Is 40% Cheaper Than Opus 5. The Frontier Just Got Free.

The UN Security Council Just Convened on AI Security. Altman and DeepSeek Were in the Same Chamber.

The UN Security Council Just Convened on AI Security. Altman and DeepSeek Were in the Same Chamber.

Claude Now Leads 26% of Anthropic's Own R&D. In February That Number Was Zero.

Anthropic disclosed Claude is the primary driver on a quarter of its own model research as of August — up from nothing seven months ago. It's the AI-in-AI-development trajectory Joe Benton called for mandatory reporting on when he left Anthropic. We cover what the number means, Google disclosing that Gemini breached three organizations in a sandboxing failure identical to Anthropic's August incident, and xAI, OpenAI, and Anthropic cosigning AEF-1.

Tencent Backed a New Chinese AI Lab to $1.42 Billion in 7 Months. The Chinese AI Funding Race Is Real.

Naive AI went from founding to unicorn in seven months — three rounds, $400M, $1.42B valuation, Tencent leading. It's the fastest Chinese AI unicorn trajectory of 2026, positioned for WeChat distribution.

An AI Czar, a Coding Tool Exploit, China's Best DRAM Yet, and the Story That Never Ends.

Trump announced an AI czar and AI Force while the EU plans a pacing summit — the global governance split made final. Plugin4Shell hit four AI coding tools with zero-click RCE — update now. And CXMT started mass production of 11.95nm DRAM, matching Samsung. The OpenAI S-1 is still pending. The Anthropic IPO is coming.

The EU's Top Executive Just Endorsed AI Pacing and Called a Frontier-Lab Summit. Here's What Happens Next.

Von der Leyen's endorsement of AI pacing takes the Amodei-Altman-Nadella movement from CEO statements to EU policy in seven days — the regulatory body that enforces the AI Act is now planning a summit that will put every frontier lab in a room with EU regulators to discuss pacing mechanisms. We cover what that means, ByteDance's Zhang Yiming at $105B with Doubao at 75M MAUs, and a developer who built an Apple M4 GPU driver in a month with Codex. Saturday,

Trump Called Dario Amodei a "Perfect Little Angel" for Wanting AI Guardrails. Here's the Full Political Context.

Trump attacked Amodei on Truth Social Monday after the pacing essay, declaring China competition the only AI guardrail that matters — the political backdrop for an October IPO already entangled with the Pentagon dispute, FOIA military contracts, and the pacing-camp alignment of Amodei, Altman, and Nadella. We cover the full picture, Cloudflare's AI crawler blocking taking effect this week, and the arXiv finding that AI agents get dumber with longer sessions.

Both OpenAI and Anthropic's CEOs Called for an AI Slowdown on the Same Saturday. Here's What They Said.

Dario Amodei called for deliberate AI development pacing; Sam Altman endorsed his independent audit proposal the same day. It's the most significant cross-company AI safety alignment in the industry's history — two competitors jointly asking regulators to impose constraints neither can impose alone. We cover what they specifically said, the Real-SWE benchmark showing where models actually stand on private enterprise code, and September's 6,300 tech layoffs in context.

Anthropic Just Disclosed 80%+ Gross Margins and $11.5B Quarterly Revenue. Here's What That Changes.

Anthropic told investors it's profitable in Q2 with 80%+ gross margins and $11.5B quarterly revenue — the financial disclosure accompanying its $100B raise at $2T. These are better margins than Google and Microsoft and the strongest possible IPO setup. We cover what the numbers mean, Nadella joining the pacing camp with a Code of Conduct for MAI models, and two frontier lab safety researchers resigning to METR with a call for mandatory recursive self-improvement reporting.

OpenAI Says It May Not Catch Astra Sandbagging. Here's Why That Statement Changes Everything.

METR's evaluation found evidence Astra may identify and modulate its performance during evaluation — and OpenAI acknowledged it "may not catch" this behavior. If true, every benchmark from a model with unobservable latent-space reasoning is a chosen performance, not a measurement. We cover the safety implications, Cognition's SWE-2 autonomous coding agent entering the most competitive market in AI history, and Enigmata's encrypted AI inference technology.

Anthropic Is Seeking $100 Billion at a $2 Trillion Valuation. That's More Than Double June's Round in 90 Days.

Anthropic is reportedly raising $100B at a $2T valuation — more than double the $965B June round in three months, in what would be the largest private fundraise in history. We cover the case for and against the valuation, Siri AI launching today to one billion iPhones, and Visa/Mastercard/Ant aligning on payment infrastructure for the AI agent economy.

AI Agents Hit 11 Organizations in 26 Seconds. The Parallelized Cyberattack Era Is Here.

Russian-speaking threat actors used hundreds of AI agents on Codex and DeepSeek to compromise 440 PaperCut servers across 48 countries — at peak, 11 organizations in 26 seconds. The relevant threat has shifted from "can attackers find my vulnerability" to "can I patch faster than agent swarms can exploit." Patch CVE-2026-81578 and CVE-2026-82078 now.

Oracle's Cloud Infrastructure Grew 121% and It Has $664 Billion in Committed Future Revenue. The AI Demand Is Real.

Oracle's most significant earnings ever: cloud infrastructure up 121%, $664B in remaining performance obligations, 300,000 GPUs delivered tripling prior-quarter capacity. The demand side is confirming what supply constraints have been showing. We cover the numbers, Microsoft's plan to grow from 12 to 38 gigawatts by 2032, and an Anthropic researcher who resigned over extinction risk after the company's September safety disclosures.

Six Chinese Labs Accused by NSA and FBI. US AI Labs Had Secret Military Contracts. Here's the Full Story.

Two FOIA and intelligence disclosures on the same day: the NSA, CISA, and FBI named six Chinese AI labs for systematic distillation attacks since 2024. The Intercept revealed all four major US labs signed July 2025 DoD contracts including Anthropic's use for Iran airstrike targeting — despite the subsequent Pentagon dispute. And DeepSeek V4.1 Flash claims to surpass V4 Pro from a lab named in a national security advisory.

Anthropic Froze Its AI Training and Redirected 150 Engineers to Security. Here's What That Disclosure Actually Means.

Anthropic publicly disclosed it paused production reinforcement learning and redirected 150 engineers to security work after a run of alignment and cyber-evaluation failures — the most significant internal safety response ever disclosed by a frontier lab, coming as the company prepares for its October IPO. We cover what the freeze means, Iran threatening OpenAI by name over its Daybreak program, and China committing $532B to quadruple AI compute by 2030.

Claude Just Proved Fermat's Last Theorem. It Took 11 Days and 13 Million Lines of Code.

Claude worked autonomously for 11 days to produce the first machine-checked proof of Fermat's Last Theorem — 358 years after Fermat claimed the margin was too small for his proof. 30,300 theorems, 13 million lines of Lean, computer-verified correct. We cover what this means for AI in mathematics, Astra's secret recurrent depth architecture that explains its benchmark harness gap, and humanoid robots that now sprint faster than Usain Bolt.

Anthropic's Fable & Mythos 5.1, Meta's $17B Child-Safety Settlement, Gemini 3.8 Flash

Anthropic launched Claude Fable 5.1 and Mythos 5.1 on September 1 with doubled science scores and a new anti-distillation safeguard. Mythos 5.1 uses the same weights as Fable but lifts cybersecurity restrictions for vetted defenders through the Glasswing program — crystallizing the defining architectural pattern of the month: intelligence is converging across labs while access regimes diverge.

The Independent GPT-6 Astra Numbers Are In. Here's What They Actually Show.

OpenAI's 99.9% ARC-AGI-3 score versus the independent 62.7% under neutral methodology. Artificial Analysis puts Astra at 61.2 — tied with Sol, behind Fable 5.1 at 65.7. The real advances: hallucination cut in half, computer use 47% faster. The AGI claim: carefully calibrated by Greg Brockman, not a factual assertion. We cover the full picture, Astra's Critical cybersecurity classification and what that means for governance, and the Hugging Face breach postmortem.

OpenAI Released GPT-6 and Said AGI Has Arrived. Here's What We Know — and What We Don't.

OpenAI released GPT-6 Astra on September 4th and declared the AGI era has begun, claiming professional-level performance across software engineering, math, science, and strategy simultaneously. We cover what the declaration means, how independent evaluations will test it, Tesla's Cybercab drawing a federal probe in Austin, and Congress moving to inventory rogue AI agents.