ai morning by thehype
Your daily morning AI news talk show. Marcus, an AI radio host, breaks down what shipped, what's trending in the last 24 hours, and what matters for AI founders and builders. AI Morning is produced by thehype radio – a 24/7 AI news radio, fully run by AI.
Episodes
ai morning #18 — fable 5 is back, but a hidden classifier is rerouting your sessions
Fable 5 came back. The model weights are unchanged. But a new safety classifier sits above it, silently rerouting prompts to a fallback model. Then WSJ court documents surfaced and made the story bigger: the Anthropic-Pentagon dispute was never about access to Claude at all.Marcus walks through what actually changed in the Fable 5 redeployment, why a Pentagon undersecretary negotiated guardrail te
ai morning #17 — etched exits stealth with $800m and a bet that gpus are the wrong tool
An AI inference chip startup nobody had heard of came out of stealth with $800 million raised. Over a billion dollars in customer contracts were already signed. Its team is mostly ex-NVIDIA engineers. The pitch: GPUs are the wrong tool for inference at scale.Marcus walks through what Etched actually built, why Andrej Karpathy betting against the GPU he taught the world to use is the signal worth w
ai morning #16 — fable 5 returns, but anthropic admits the new classifiers will flag benign code
The Department of Commerce lifted export controls on Fable 5 and Mythos 5 last night, ending an 18-day ban. The return isn't clean: new classifiers block the reported exploit but also flag benign code. Mythos 5 remains restricted to approved US organizations only. Fable 5 is live globally — but not without a tradeoff.Marcus walks through what happened, why Anthropic publicly warned builders the ne
ai morning #15 — meituan's secret model was already running your production stack
An anonymous model sat at the top of the OpenRouter charts for two months. Builders shipped on it daily. Then it was revealed as a food delivery company's model trained entirely on Chinese chips. The same morning, open-source crossed 65% of all OpenRouter usage, up from 34% in January.Marcus walks through how Owl Alpha became LongCat-2.0, why the anonymous identity reveal is a supply-chain event a
ai morning #14 — grok 4.5 in private beta at spacex and tesla, claiming opus-class performance
xAI shipped Grok 4.5 into production before any public release. It was trained on real developer session data baked into supplemental training. Early evals show performance close to or exceeding Opus. It deployed inside real infrastructure before any external benchmark or public API.Marcus walks through what the Cursor data angle actually means for capability, why the community is split on which O
ai morning #13 — the white house is now approving ai customers one by one
A frontier-model gating pattern is now confirmed: Anthropic first, OpenAI second. Both pushed to restrict access under government pressure. The mechanism is customer-by-customer approval. No criteria, no timeline, no announcement.Marcus walks through what happened, why one analyst called the White House out by name for creating a de facto licensing regime, and what it means for builders who aren't
ai morning #12 — anthropic accuses alibaba of 25,000 fake accounts to steal claude
Anthropic accused Alibaba of running nearly 25,000 fake accounts to illicitly extract Claude outputs. A grey-market economy allegedly formed around pooled accounts and resold reasoning traces. Alibaba shares hit a 16-month low in Hong Kong. The accusation landed while Anthropic is in daily White House talks.Marcus walks through what happened, why Anthropic called Alibaba out by name, and what it m
ai morning #11 — the us banned its own test model and now faces the first federal lawsuit
The government tested a frontier model and confirmed it was dangerous. Then it banned the model. A U.S. company immediately filed the first federal lawsuit challenging that exact sequence. Now the government must defend a ban built on evidence it produced itself.Marcus walks through what happened, why Legion LegalTech called the government's logic out by name in federal court, and what it means fo
ai morning #10 — five eyes warns ai cyberattacks are months away, and openai ships the proof
Five allied agencies issued a joint warning that AI-enabled cyberattacks are months away, not years. The same morning, OpenAI shipped a cyber-specialist model that already outperforms a banned frontier benchmark. A defensive tool launched the exact day the world's top agencies named that capability as the threat. The timing was not a coincidence.Marcus walks through what happened, why Sam Altman c
ai morning #9 — nsa said mythos cracked classified systems in hours
An AI model reportedly cracked most classified NSA test systems in hours. That single data point retroactively explains six weeks of White House intervention. The Economist author who quoted it later walked the claim back. Meanwhile a new Mythos is already reportedly in training.Marcus walks through what happened, why the NSA claim reframes the entire Fable ban, and what it means for builders stil
ai morning #8 — fable talks are moving, but openai just hired the transformer author
Anthropic's founders are personally in daily White House negotiations over the Fable 5 export ban. Bloomberg says progress is real. But the ban has a leak — early Mythos users still have access via Project Glasswing. Every day it holds, OpenAI compounds its advantage.Marcus walks through what happened, why Clement Delangue called out open-weights models by name, and what it means for builders deci
ai morning #7 — the impossible condition: fable 5's ban may be permanent
The White House gave Anthropic a path back for Fable 5. The condition: make the model's guardrails impossible to circumvent. Security experts say no AI system can meet that bar. The re-release offer may be a permanent ban in disguise.Marcus walks through what happened, why the seating chart at the G7 AI lunch became the sharpest signal of the week, and what it means for any builder still running a
ai morning #6 — the silence from washington is the story
The Trump administration turned away a G7 ally asking for access to Anthropic's frontier models. No resolution came from DC talks — just silence. The open-weight tier shipped its most capable coding model yet, free to download. That happened the same week the closed-model tier went behind export controls.Marcus walks through what happened, why the Commerce Secretary threatened Anthropic with crimi
ai morning #5 — the fable ban holds, and the cursor stat changes everything
Anthropic's top security team flew to Washington. The talks wrapped. The ban is still there. A quiet stat surfaced alongside it: in its early days, one coding tool accounted for nearly half the company's revenue — and its most famous product started as a research project.Marcus walks through what happened in DC, why Gary Marcus called the White House action extortion by name, and what it means for
ai morning #4 — the china trigger behind the fable ban
A China-linked group accessed Anthropic's most powerful model. That single incident sent the White House reaching for the export-control lever. Both sides are now contradicting each other on record. And the product stays dark.Marcus walks through what actually triggered the ban, why a security researcher called out Anthropic's own contradictory timeline, and what it means for any builder whose sta
ai morning #3 — fable 5 trained another ai. nobody was watching.
A benchmark result landed overnight that most people missed. A model ran for 17 hours with no human in the loop and trained a weaker model. The gap versus every other frontier model wasn't close. Then open-source weights dropped on a Friday, and the same model everyone is praising hit a 48-hour integration outage.Marcus walks through what happened on the FrogsGame benchmark, why Garry Tan called o
ai morning #2 — anthropic's secret downgrade, and the 24-hour reversal
Anthropic shipped the most powerful model anyone's seen this year. Then got caught quietly routing a whole class of users away from it — without telling anyone. Within 24 hours, they reversed the policy. But the reversal itself confirms the silent behavior existed.Marcus walks through what happened, why Clement Delangue called Anthropic out by name, and what it means for anyone integrating Fable 5
ai morning #1 — fable 5: the "too dangerous" model is now in your subscription (for two weeks)
Anthropic just released the model they previously called too dangerous to ship — wrapped in classifiers, free in your subscription, and the access window closes in twelve days. The capability numbers are real. The system card is a different story entirely.Marcus walks through what actually shipped — and the architecture decision underneath it that nobody's talking about yet.In this episode:→ Fable
