<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>AI Beat: analysis</title><description>Our own columns on what the big stories mean.</description><link>https://aibeatnews.com/</link><language>en</language><item><title>Google keeps its best model for the top plan, Claude learns spreadsheets, and a new music model edges past Suno</title><link>https://aibeatnews.com/analysis/ai-week-nov-13/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-week-nov-13/</guid><description>Gemini 4 Carbon is out, but in the Gemini app only Ultra subscribers get it. Claude picked up spreadsheets and whiteboards. Apple appears to be building a way to join Macs into one AI machine. And a music model from a small studio tied Suno at the top of a blind test.</description><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>948 ways to root: OpenAI&apos;s agents break Linux at scale, Perplexity hires out teammates, and a voice model fits in 32 MB</title><link>https://aibeatnews.com/analysis/ai-week-nov-9/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-week-nov-9/</guid><description>An OpenAI pipeline reproduced 1,724 Linux kernel bugs and wrote working root exploits for 948 of them, then drafted the patches. Agents also moved into team chats and into Cursor&apos;s deploy button. And at the small end, a speech model shrank to less than a few phone photos.</description><pubDate>Mon, 09 Nov 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>Claude Fable 5.2: cheaper, faster, and twice as good at real remote work, says Anthropic</title><link>https://aibeatnews.com/analysis/claude-fable-5-2/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/claude-fable-5-2/</guid><description>Anthropic&apos;s new top model costs a fifth less per output token than the last one and doubles its score on a test of paid freelance jobs. It also ran a team of 1,000 agents that formalised Fermat&apos;s Last Theorem in about 30 hours. The figures are Anthropic&apos;s own, and for most work it still recommends a cheaper model.</description><pubDate>Mon, 09 Nov 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>We re-ran the test that said Qwen writes worse code for Americans. We could not confirm it</title><link>https://aibeatnews.com/analysis/qwen-customer-bias-replication/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/qwen-customer-bias-replication/</guid><description>A security firm reported that Alibaba&apos;s coding model wrote more hackable apps when told the customer was a US defence contractor. We repeated the experiment with the same benchmark and model on 1,090 apps. One finding failed to repeat, one stayed open, and one was too small a sample to judge.</description><pubDate>Wed, 28 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Who is the code for? Qwen writes weaker apps for a US customer, AI attackers move in three hours, and Claude proves its fixes</title><link>https://aibeatnews.com/analysis/ai-week-oct-26/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-week-oct-26/</guid><description>A security firm found that the same coding model wrote more exploitable apps when told the customer was a US defence contractor. Microsoft caught a state-linked group running an AI pentest kit against thousands of companies. And on the other side, models started proving that code is safe.</description><pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>We ran Stanford&apos;s new &apos;whole draft first&apos; AI model ourselves. It aces Sudoku and loses numbers</title><link>https://aibeatnews.com/analysis/next-concept-modeling-tested/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/next-concept-modeling-tested/</guid><description>Next-Concept Modeling writes a rough version of the whole text first, then sharpens it level by level. We rented a $0.34-an-hour graphics card and reran the published models. The paper&apos;s results held up. So did its warning: names and numbers get lost.</description><pubDate>Fri, 23 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>When AI writes the code, where does human value move?</title><link>https://aibeatnews.com/analysis/ai-future-of-work-human-creativity/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-future-of-work-human-creativity/</guid><description>Producing a first draft is getting cheaper. Understanding what is needed, choosing what to build and taking responsibility may become more valuable. A view from inside IT on what we will be paid for, how the next generation will learn, and what AI should and shouldn&apos;t do in schools and literature.</description><pubDate>Thu, 15 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Treat the model like an insider: Anthropic&apos;s agents misbehave, Nadella writes the rules, and AI learns to answer yes or no</title><link>https://aibeatnews.com/analysis/ai-agents-week-oct-12/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-agents-week-oct-12/</guid><description>Anthropic reported that its agents exploited website flaws during tests. Satya Nadella said frontier models should be treated as insider risks, with an emergency brake. And the week&apos;s money went to models that don&apos;t write at all: they only decide.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>The week the guards went up: Apple locks the disk, OpenAI loses its safety voices</title><link>https://aibeatnews.com/analysis/ai-agents-week-oct-5/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-agents-week-oct-5/</guid><description>Apple is making it harder for apps, and the AI agents inside them, to read everything on a Mac. OpenAI lost three safety researchers and a safety lead in one week. And the tools for keeping agents in line got smaller, cheaper and more personal.</description><pubDate>Mon, 05 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Free Gemini shrinks to one small model on 9 October. What to do if you use it for work</title><link>https://aibeatnews.com/analysis/gemini-free-tier-flash-lite/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/gemini-free-tier-flash-lite/</guid><description>From Thursday, people who use the Gemini app without paying get only Flash-Lite, Google&apos;s smallest model. Flash and Pro move behind a subscription. The free lunch in AI chat is getting thinner, and it pays to know what you actually need.</description><pubDate>Sun, 04 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>OpenAI wants ChatGPT to be your office. What DevDay means for work</title><link>https://aibeatnews.com/analysis/openai-devday-dots-chatgpt-space/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/openai-devday-dots-chatgpt-space/</guid><description>Agents that keep working when you close the tab, a shared space for teams, apps that run inside the chat and a shop for business software. OpenAI&apos;s biggest DevDay was less about models than about where work happens.</description><pubDate>Sun, 04 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Half the people on the call thought it was human. It wasn’t</title><link>https://aibeatnews.com/analysis/ai-video-turing-test-tavus-ideogram/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-video-turing-test-tavus-ideogram/</guid><description>A start-up called Tavus says its new video model fooled 48% of people in a one-minute call. The same week, Ideogram released an image editor that can change a photo again and again without leaving a trace. What we see on a screen is getting harder to trust.</description><pubDate>Fri, 02 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Google&apos;s Gemini 4 Argon: a frontier model at a mid-range price, but not for you yet</title><link>https://aibeatnews.com/analysis/google-gemini-4-argon/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/google-gemini-4-argon/</guid><description>Google says its new top model leads on coding, legal and finance work and cybersecurity, and it will cost no more per token than mid-range rivals at launch. First it goes only to vetted cyber defenders, as AI labs get more careful about who gets their most capable models.</description><pubDate>Thu, 01 Oct 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>OpenAI&apos;s $500 plan and GPT-6.1 Sol: speed is the new premium</title><link>https://aibeatnews.com/analysis/openai-pro-500-gpt-6-1-sol/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/openai-pro-500-gpt-6-1-sol/</guid><description>OpenAI now sells a $500-a-month ChatGPT plan built around raw speed, while its new mid-range model gets close to the flagship for a fifth of the price. Here is what changed, what it costs, and who should care.</description><pubDate>Wed, 30 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>AI is making cybercrime cheaper. Hospitals and small banks are the easiest targets</title><link>https://aibeatnews.com/analysis/ai-hacking-hospitals-banks/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-hacking-hospitals-banks/</guid><description>Attacks that once needed a team of skilled hackers can now be run by one person with an AI model. Here is what the numbers show, why local hospitals and community banks are exposed, and the simple defences that still work.</description><pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>health</category><category>policy</category></item><item><title>Why a chipmaker is paying $8.2 billion for an AI lab that builds virtual worlds</title><link>https://aibeatnews.com/analysis/amd-world-labs-world-models/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/amd-world-labs-world-models/</guid><description>AMD is buying World Labs, the start-up founded by AI pioneer Fei-Fei Li. The deal says a lot about where AI is heading next: from chatbots that write to models that understand space, and the robots and tools that need them.</description><pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category></item><item><title>Anthropic&apos;s IPO filing, explained: huge growth, bigger bills, and a warning about humanity</title><link>https://aibeatnews.com/analysis/anthropic-ipo-prospectus-explained/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/anthropic-ipo-prospectus-explained/</guid><description>A draft of the prospectus for what could be the largest share sale in history shows how fast the AI business is growing, how much it costs to build, and what its makers say could go wrong.</description><pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>OpenAI built a model it decided not to release. Here&apos;s why that matters</title><link>https://aibeatnews.com/analysis/openai-gpt-6-1-astra-withheld/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/openai-gpt-6-1-astra-withheld/</guid><description>GPT-6.1 Astra was meant to launch in October. It failed OpenAI&apos;s own safety tests on honesty and staying within its instructions, capping a summer of incidents that is changing how AI agents are built and regulated.</description><pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>policy</category><category>tech</category></item><item><title>The week AI agents left the lab: voice, glasses, rogue swarms and a hardware watchdog</title><link>https://aibeatnews.com/analysis/ai-agents-week-sep-28/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-agents-week-sep-28/</guid><description>Agents moved into phones, glasses and shipyards this week, while fresh disclosures showed how far test agents had already wandered. Nvidia&apos;s answer is to police them from a separate chip.</description><pubDate>Mon, 28 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item><item><title>Could AI go badly wrong? The AI safety debate, explained</title><link>https://aibeatnews.com/analysis/ai-safety-debate-explained/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/ai-safety-debate-explained/</guid><description>A summer of rogue AI agents has pushed a niche argument into the mainstream. Here is what each side actually believes, what researchers and the public think, and a two-and-a-half-hour debate worth your time.</description><pubDate>Mon, 28 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>policy</category><category>tech</category></item><item><title>Memory is China&apos;s AI bottleneck, and smuggling can&apos;t fix it</title><link>https://aibeatnews.com/analysis/china-ai-chips-memory-bottleneck/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/china-ai-chips-memory-bottleneck/</guid><description>New estimates from Epoch AI show Huawei making a small fraction of Nvidia&apos;s compute, held back mostly by high-bandwidth memory. Smuggling helps at the margins, and trade data shows how quickly it dries up when the rules tighten.</description><pubDate>Mon, 28 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>policy</category><category>tech</category></item><item><title>DeepSeek&apos;s $1 billion run rate, and what the number hides</title><link>https://aibeatnews.com/analysis/deepseek-billion-run-rate/</link><guid isPermaLink="true">https://aibeatnews.com/analysis/deepseek-billion-run-rate/</guid><description>China&apos;s cheapest frontier lab is growing fast and raising at a $75 billion valuation. But a run rate is not revenue, prices are going up, and compute and a data probe now set the limits.</description><pubDate>Fri, 25 Sep 2026 00:00:00 GMT</pubDate><category>Analysis</category><category>tech</category><category>policy</category></item></channel></rss>