<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>InYourGeek — AI</title><description>Models, agents, tools and uses — and the often generous gap between the demo and the promise.</description><link>https://inyourgeek.com</link><language>en-US</language><item><title>GPT-6 Astra: what actually changes in ChatGPT and ChatGPT Work</title><link>https://inyourgeek.com/en/articles/2026-09-07-gpt-6-astra-ce-qui-change-vraiment-dans-chatgpt</link><guid isPermaLink="true">https://inyourgeek.com/en/articles/2026-09-07-gpt-6-astra-ce-qui-change-vraiment-dans-chatgpt</guid><description>OpenAI unveiled GPT-6 Astra on 3 September 2026: a model that drives a computer, fills in forms and builds slide decks. It is also the first one OpenAI rates &quot;Critical&quot; for cybersecurity.</description><pubDate>Mon, 07 Sep 2026 05:00:17 GMT</pubDate><category>ia</category><category>openai</category><category>chatgpt</category><category>agents</category><category>securite</category></item><item><title>Cerebras serves Qwen 3.8 27B at 1500 tokens/s, and the agent stops waiting</title><link>https://inyourgeek.com/en/articles/2026-09-03-cerebras-sert-qwen-3-8-27b-a-1500-tokens</link><guid isPermaLink="true">https://inyourgeek.com/en/articles/2026-09-03-cerebras-sert-qwen-3-8-27b-a-1500-tokens</guid><description>Cerebras&apos;s inference documentation lists Qwen 3.8 27B served at roughly 1500 tokens per second on its public endpoints. For a chatbot that is comfort; for an agent loop, it changes what the thing is.</description><pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate><category>ia</category><category>cerebras</category><category>qwen</category><category>llm</category><category>agents</category></item><item><title>GPT-6 Astra hits 99.9% on ARC-AGI-3 — on a harness it built itself</title><link>https://inyourgeek.com/en/articles/2026-09-03-gpt-6-astra-99-9-sur-arc-agi-3</link><guid isPermaLink="true">https://inyourgeek.com/en/articles/2026-09-03-gpt-6-astra-99-9-sur-arc-agi-3</guid><description>OpenAI shipped GPT-6 Astra on 3 September 2026 with a near-perfect ARC-AGI 3 score. Measured on the benchmark&apos;s own standard harness, the same model drops to 62.7% — and costs more.</description><pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate><category>ia</category><category>openai</category><category>benchmark</category><category>llm</category></item><item><title>Nvidia buys Hugging Face for $13 billion and promises openness</title><link>https://inyourgeek.com/en/articles/2026-09-03-nvidia-rachete-hugging-face-pour-13-milliards-et-promet</link><guid isPermaLink="true">https://inyourgeek.com/en/articles/2026-09-03-nvidia-rachete-hugging-face-pour-13-milliards-et-promet</guid><description>Nvidia announced on Thursday that it is buying Hugging Face for $13 billion, its largest acquisition to date. The reference catalogue for open models would pass under the control of the company selling the GPUs to run them.</description><pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate><category>ia</category><category>nvidia</category><category>huggingface</category><category>opensource</category><category>rachat</category></item><item><title>Cross-Outage: ChatGPT, Claude, Grok and Gemini All Broke at Once</title><link>https://inyourgeek.com/en/articles/2026-09-03-panne-croisee-chatgpt-claude-grok-et-gemini-tombes-le</link><guid isPermaLink="true">https://inyourgeek.com/en/articles/2026-09-03-panne-croisee-chatgpt-claude-grok-et-gemini-tombes-le</guid><description>On the morning of Thursday, September 3, all four major AI assistants stumbled inside the same few-hour window. Rare enough that you go hunting for the common supplier — except none of them ever declared an outage.</description><pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate><category>ia</category><category>panne</category><category>openai</category><category>anthropic</category><category>infrastructure</category></item></channel></rss>