Archive
Latest — page 6
Older stories, newest first.
Weak June jobs report puts AI automation in the frame — but not on trial
A 57,000 payrolls print is a real signal about the labour market. Whether it is a signal about AI is a separate, and much harder, question.
Anthropic is in talks with Samsung to build a custom AI chip
The reported Samsung talks would give Anthropic a fourth silicon supplier — and another lever on the per-token economics now defining the frontier.
UN and ITU launch the AI for Good Global Commission with frontier CEOs
The UN's first frontier-AI governance body seats the CEOs building the technology — a calculated bet that inclusion shapes norms faster than regulation imposed from the outside.
Claude Fable 5 and Mythos 5 are back — but the export-control precedent survives
The controls are gone, but the power that imposed them was never tested in court — and that is the part that lasts.
What Claude Sonnet 5 does to the economics of running agents
Anthropic has put near-Opus quality at a Sonnet price, and the real story is what that does to the cost of running agents at scale.
Meituan's LongCat-2.0: the first trillion-parameter model built without Nvidia
A 1.6-trillion-parameter open model is the smaller story. The compute stack underneath it — 50,000 domestic chips, no Nvidia — is the one that changes the map.
xAI's Grok 4.5 enters private beta at SpaceX and Tesla
xAI has put its 1.5-trillion-parameter Grok 4.5 in front of SpaceX and Tesla engineers first. With no benchmarks and no system card, Musk's 'beats Opus' claim is untestable.
The Fable 5 jailbreak wasn't unique — every frontier model shared it
The export order that pulled Claude Fable 5 traced to one cyber jailbreak. Anthropic's later investigation found the same bypass worked across the frontier.
OpenAI gates GPT-5.6 behind a federal access wall — and objects
The real story of GPT-5.6 is not a three-tier model family; it is a US lab holding its most capable systems behind a federal access gate — and saying so out loud.
The end of tokenmaxxing: why AI buyers are routing to cheaper open-weight models
The assumption that intelligence could be metered by the token, and that usage would only rise, has quietly become negotiable.
Why DeepMind researchers are leaving Google for AI rivals
A single week of departures — a Transformer co-author and a Nobel laureate among them — shows how pre-IPO equity and compute politics are redrawing the frontier-lab map.
Anthropic accuses Alibaba of the largest known distillation attack on Claude
The accusation is unusually specific, the venue was a Senate committee rather than a press release, and Alibaba denies all of it. The mechanism — not the numbers — is what makes it matter.
Gemini 3.5 Pro's 'next month' has arrived — and Pro hasn't
Google promised the Pro model 'next month' at I/O. The month is up; the model isn't out.
AI as scientific instrument: what OpenAI's June wave actually demonstrates
Four results in four days — rare-disease diagnosis, a near-autonomous chemist, a sobering benchmark. Tracing how each was validated is the only read that matters.
The 2026 voice and TTS landscape, mapped
AI voice became indistinguishable from human this year. The real story is the structural split it's created — and which camp wins which kind of work.
The 2026 open-weight surge, explained
A year of downloadable models that narrowed the gap with the closed frontier — mostly from China.
Qwen and the rise of open-source AI from China
One of the most-downloaded model families anywhere — and the strategy behind it.
Three days after launch, Washington pulls Fable 5 — and the precedent is the story
An export-control directive reaching every foreign national forced Anthropic to switch its best model off for everyone. The mechanism matters more than the model.
Kimi K2.7-Code and the metronome of Chinese open weights
Another month, another capable open model from a Chinese lab. The cadence is the story.
Mistral's €20bn raise and Europe's sovereign-AI bet
Nearly doubling its valuation in months — the thesis is as much geopolitical as financial.
Humanoid robots in 2026: from pilot to platform
Past the demo reels, the real story is deployment economics — and the AI making it possible.
Anthropic wants the world to be able to hit pause
A serious argument about recursive self-improvement — landing in an awkward commercial moment.
Decart's Oasis 3 and the wager on 'world models'
A photorealistic driving simulator you can prompt in plain language — and an attempt to do for world models what OpenAI did for language.
Fable 5 and the cost of the frontier: what Anthropic's launch actually signals
The launch is less about the benchmark scores than what they imply about price, safety and the next eighteen months.