Latest

The US calls the UK's plan to rank "trustworthy" news censorship

The Guardrails

The US calls the UK's plan to rank "trustworthy" news censorship

Washington has filed a formal objection to Britain's plan to make platforms push approved news to the top of your feed, OpenAI's own chart says the longer an agent works the more it needs you, and a Chinese city has decided inference bills are a talent problem. The Trump administration has told the UK that forcing platforms to promote "trustworthy" news providers would be censorship — and that it would treat it as such. The US Embassy in London filed a response to DCMS's Green Paper "Watch this

Quick Hits — September 7, 2026

Quick Hits

Quick Hits — September 7, 2026

The midday round: a memory crunch starts setting phone prices, China's model labs move onto Tmall's shelves, and Cisco's numbers say a clean single-turn safety score tells you almost nothing. Cisco tested 15 frontier models against attackers who get to come back for a second turn — and none of them held. Its researchers split harmful requests across a conversation rather than firing them off in one prompt, and multi-turn attack success rates landed between 7.89% and 88.30%, against 2.19% to 64.

Who's liable when a $3.2B AI data center catches fire?

The Guardrails

Who's liable when a $3.2B AI data center catches fire?

Two stories this afternoon about the gap between what AI companies promise and what anyone can actually hold them to — one measured in megawatts and dry hydrants, one in a president's AI-generated feed. A fire at a half-built AI data center in western New York exposed how thin the accountability really is. In early June, crews responding to a blaze in an unfinished building at the Lake Mariner campus in Somerset found no working alarm, no suppression system, and three dead hydrants — and the ch

How to — spot prompt injection in a product you use

The Guardrails

How to — spot prompt injection in a product you use

By the end of this you'll be able to answer one question about any AI product you rely on: if a stranger hid a sentence inside something your assistant reads tomorrow, what is the worst thing that could happen, and would you notice? Six moves, no security background required. Prompt injection is not a hack of the software — it's text. Somebody writes instructions into a page, a document, an email or a file name, your assistant reads that text while helping you, and follows it. The reason this i

Tencent's EVIE turns a scanned page into something you can actually search

The Stack

Tencent's EVIE turns a scanned page into something you can actually search

Two releases land in the same place today: Tencent putting open-weight visual document retrieval models on Hugging Face, and vLLM publishing what happens when you run speculative decoding on AMD silicon instead of Nvidia's. One is about what AI can read; the other is about how cheaply you can run it. Tencent has released EVIE-8B and EVIE-4.5B, open-weight models built to retrieve pages of documents by what they look like — and the smaller one nearly matches the big one. The pair scored 66.75 an

Open Source Radar — September 7: agents leave the chat window

The Stack

Open Source Radar — September 7: agents leave the chat window

Today's board splits cleanly between things that give your agent a body — a desk, a voice, a cluster — and things that hand it a wallet. Five fresh picks, none of them repeats from this week's earlier radars. AIPOCH Open Science (TypeScript, 4,037 stars) — A local-first research workbench where you describe a research goal in plain language and the agents read your files, search the web, run Python and R, query scientific data sources, and hand back reports, tables, and figures with a trail you

LG makes you liable for anyone its TV's AI mic records

The Stack

LG makes you liable for anyone its TV's AI mic records

Two hardware stories and one biology model this hour: a TV that listens when it looks off, a phone chip that runs a 30B model locally, and a four-month-old startup that says it beat AlphaFold 3 where it hurts. LG's smart TVs were caught capturing voice audio while the screen was off — and its terms put the legal risk on you. Hardware teardown channel Gamers Nexus published a 135-minute investigation that Notebookcheck summarized: during bench tests the TV recorded clean microphone audio while t

Deep Dive — Robots are learning what their actions do, not what to do

The Frontier

Deep Dive — Robots are learning what their actions do, not what to do

Between May and August, at least ten Chinese embodied-AI companies shipped a world model. That is a land rush in four months, and it arrived just as the World Robot Conference in Beijing turned into a showcase for the same idea. When a field this young converges this fast, the useful question is not who is winning — it is whether everyone is building the same thing. They are not, and the split tells you more about robotics in 2026 than any leaderboard does. The clearest statement of the split c

Mostik's bridge puts a 753B model's thinking inside a 4B phone model

The Frontier

Mostik's bridge puts a 753B model's thinking inside a 4B phone model

Two stories from the same frontier: a startup arguing that the next capability jump comes from wiring models together rather than scaling them up, and a Chinese robotics lab arguing that a world model is only useful if it knows what the robot's own actions did. Mostik, a four-month-old startup whose chief scientist is 2010 Fields Medalist Stanislav Smirnov, says it has built a translation layer that lets a 753-billion-parameter GLM-5.2 pass its internal state straight into a 4-billion-parameter

DeepSeek plans a 160,000-chip Huawei cluster in Inner Mongolia

The Arena

DeepSeek plans a 160,000-chip Huawei cluster in Inner Mongolia

Monday's stories are about what happens after the models ship: a Chinese lab committing to domestic silicon at a scale nobody has attempted, a record graduating class walking into a hiring market AI is quietly reshaping, and a retailer giving away its AI stack to move more hardware. DeepSeek plans to install at least 160,000 Huawei Ascend 950DT accelerators at a data center it is building in Inner Mongolia, according to people familiar with the matter — a deployment that would rank among the la

China's office agents: 20% of users burn 87% of the compute

The Everyday

China's office agents: 20% of users burn 87% of the compute

A million-user desktop agent just opened its logs, and the numbers say agent adoption is nothing like the chatbot curve. Also: OpenAI takes its newsroom AI program to Ukraine. China's first office-agent behavior report is out, and usage is extremely top-heavy. Published in Beijing on September 7, the report draws on real user data from LobsterAI, an open-source desktop office agent from a major Chinese tech company that passed one million users in August. The headline finding: the top 20% of us

AI 101 — What is a transformer?

The Frontier

AI 101 — What is a transformer?

A transformer is the neural network design almost every modern AI system is built on — an architecture that lets a model look at every word in its input at once and decide which other words matter for understanding it. That mechanism is called attention, and it is the reason AI went from clunky to conversational. The name comes from a 2017 paper by eight researchers at Google, titled "Attention Is All You Need." Before it, language models read text one word at a time, in order, like someone run

See all