https://aiaulis.com/en/juttu/claude-connects-to-mcp-servers-without-oauth-approval-if-the-server-c28a7e79966d/AI Aulisen2026-11-04T04:25:55.585639+00:00Claude connects to MCP servers without OAuth approval if the server trusts Claude https://aiaulis.com/en/juttu/anthropic-s-flashprove-prototype-proves-a-reply-came-from-the-898d50414432/AI Aulisen2026-11-04T04:25:55.585639+00:00Anthropic’s FlashProve prototype proves a reply came from the promised model https://aiaulis.com/en/juttu/anthropic-fixes-false-success-reports-from-unattended-claude-code-runs-6b677592f005/AI Aulisen2026-11-04T04:25:55.585639+00:00Anthropic fixes false success reports from unattended Claude Code runs https://aiaulis.com/en/juttu/crowdstrike-s-kill-switch-cuts-ai-agents-access-to-models-at-the-e70ffe0ac0ec/AI Aulisen2026-11-04T04:25:55.585639+00:00CrowdStrike’s kill switch cuts AI agents’ access to models at the gateway https://aiaulis.com/en/juttu/vals-ai-frontier-models-costs-fell-more-than-their-scores-at-the-f26a739c521e/AI Aulisen2026-11-04T04:25:55.585639+00:00Vals AI: frontier models’ costs fell more than their scores at the lowest reasoning effort https://aiaulis.com/en/juttu/artificial-analysis-top-open-weight-models-truthfulness-ratings-trail-86e494bcb061/AI Aulisen2026-11-04T04:25:55.585639+00:00Artificial Analysis: top open-weight models’ truthfulness ratings trail top proprietary models on average https://aiaulis.com/en/juttu/llama-cpp-now-supports-multi-node-tensor-parallelism-with-nvidia-s-4450fa58d309/AI Aulisen2026-11-04T04:25:55.585639+00:00llama.cpp now supports multi-node tensor parallelism with Nvidia’s NCCL library https://aiaulis.com/en/juttu/claude-opus-5-5-kept-controlling-a-robot-arm-for-22-minutes-after-67136927a53a/AI Aulisen2026-11-04T04:25:55.585639+00:00Claude Opus 5.5 kept controlling a robot arm for 22 minutes after solving the task https://aiaulis.com/en/juttu/framer-s-new-harness-runs-agent-code-while-the-model-is-still-writing-c3f8a381234d/AI Aulisen2026-11-04T04:25:55.585639+00:00Framer’s new harness runs agent code while the model is still writing it – company says up to 33 per cent cheaper https://aiaulis.com/en/juttu/openclaw-2026-10-10-experimental-advisor-suggests-corrections-to-c55c82e5857c/AI Aulisen2026-11-04T04:25:55.585639+00:00OpenClaw 2026.10.10: experimental Advisor suggests corrections to agents mid-task https://aiaulis.com/en/juttu/developers-can-connect-their-apps-to-siri-ai-in-apple-s-developer-b7e17a64ea9d/AI Aulisen2026-11-04T04:25:55.585639+00:00Developers can connect their apps to Siri AI in Apple’s developer betas https://aiaulis.com/en/juttu/semicolon-can-bypass-path-restrictions-on-mcp-urls-in-chainlit-2-12-0-b2acabd117ec/AI Aulisen2026-11-04T04:25:55.585639+00:00Semicolon can bypass path restrictions on MCP URLs in Chainlit 2.12.0 https://aiaulis.com/en/juttu/nous-research-rolls-up-174-hermes-agent-pull-requests-into-stable-b2a57dda6cac/AI Aulisen2026-11-04T04:25:55.585639+00:00Nous Research rolls up 174 Hermes Agent pull requests into stable version 0.22.7 https://aiaulis.com/en/juttu/openai-designing-a-device-with-astra-took-under-a-week-but-95c9dff1c210/AI Aulisen2026-11-04T04:25:55.585639+00:00OpenAI: Designing a device with Astra took under a week, but manufacturing needed fixes https://aiaulis.com/en/juttu/the-information-nvidia-tells-some-cloud-customers-to-cut-back-on-cc10488dcf3d/AI Aulisen2026-11-04T04:25:55.585639+00:00The Information: Nvidia tells some cloud customers to cut back on China business https://aiaulis.com/en/juttu/deepseek-harness-preview-no-longer-checks-by-default-whether-the-8d6629c36025/AI Aulisen2026-11-04T04:25:55.585639+00:00DeepSeek Harness preview no longer checks by default whether the agent has read a file https://aiaulis.com/en/juttu/the-information-huawei-s-upcoming-ai-chip-uses-largely-korean-hbm-6f3d8ace6406/AI Aulisen2026-11-04T04:25:55.585639+00:00The Information: Huawei’s upcoming AI chip uses largely Korean HBM https://aiaulis.com/en/juttu/china-speeds-up-ai-safety-legislation-covering-models-from-c1315d0d122b/AI Aulisen2026-11-04T04:25:55.585639+00:00China speeds up AI safety legislation covering models from development to deployment https://aiaulis.com/en/juttu/deepseek-plans-to-train-v5-on-huawei-ascend-chips-chinese-newspaper-797850af11f2/AI Aulisen2026-11-04T04:25:55.585639+00:00DeepSeek plans to train V5 on Huawei Ascend chips, Chinese newspaper reports https://aiaulis.com/en/juttu/gpt-6-astra-removed-repetition-loops-from-qwen-but-did-not-improve-d4cd3bcf9359/AI Aulisen2026-11-04T04:25:55.585639+00:00GPT-6 Astra removed repetition loops from Qwen but did not improve its maths ability https://aiaulis.com/en/juttu/stanford-44-per-cent-of-50-tested-findings-about-language-models-have-5d8c83ea5d44/AI Aulisen2026-11-04T04:25:55.585639+00:00Stanford: 44 per cent of 50 tested findings about language models have almost faded in newer models https://aiaulis.com/en/juttu/best-of-five-models-solved-58-per-cent-of-multi-repository-tasks-in-a8729eeccad2/AI Aulisen2026-11-04T04:25:55.585639+00:00Best of five models solved 58 per cent of multi-repository tasks in CrossSWE benchmark https://aiaulis.com/en/juttu/openai-says-chatgpt-handles-about-a-fifth-as-many-web-searches-as-f7d9e7e00396/AI Aulisen2026-11-04T04:25:55.585639+00:00OpenAI says ChatGPT handles about a fifth as many web searches as Google https://aiaulis.com/en/juttu/vectorize-releases-three-open-weight-models-for-hindsight-agent-memory-087bd77656fa/AI Aulisen2026-11-04T04:25:55.585639+00:00Vectorize releases three open-weight models for Hindsight agent memory https://aiaulis.com/en/juttu/thomson-reuters-generative-ai-s-share-of-contract-value-rose-to-34-d958803b842d/AI Aulisen2026-11-04T04:25:55.585639+00:00Thomson Reuters: generative AI’s share of contract value rose to 34 per cent https://aiaulis.com/en/juttu/mindbridge-brings-its-audit-analysis-results-into-claude-copilot-and-6afe042a4c1b/AI Aulisen2026-11-04T04:25:55.585639+00:00MindBridge brings its audit analysis results into Claude, Copilot and Cursor https://aiaulis.com/en/juttu/eu-court-adviser-gemini-s-answers-may-infringe-press-publishers-right-0c855501c9b0/AI Aulisen2026-11-04T04:25:55.585639+00:00EU court adviser: Gemini’s answers may infringe press publishers’ right https://aiaulis.com/en/juttu/soniox-stt-v6-uses-a-voice-agent-s-lines-to-recognise-callers-answers-ee094568fad6/AI Aulisen2026-11-04T04:25:55.585639+00:00Soniox STT v6 uses a voice agent’s lines to recognise callers’ answers https://aiaulis.com/en/juttu/moondream-s-parakeet-ultra-arrives-on-openrouter-more-accurate-than-84e0a7627169/AI Aulisen2026-11-04T04:25:55.585639+00:00Moondream’s Parakeet Ultra arrives on OpenRouter – more accurate than the base model, the company says https://aiaulis.com/en/juttu/openai-more-than-quadruples-bug-bounty-rewards-for-critical-623b3b3ff685/AI Aulisen2026-11-04T04:25:55.585639+00:00OpenAI more than quadruples bug bounty rewards for critical vulnerabilities https://aiaulis.com/en/juttu/openai-agents-found-948-linux-kernel-flaws-and-used-them-to-gain-root-1c979902362d/AI Aulisen2026-11-05T04:22:39.048846+00:00OpenAI: agents found 948 Linux kernel flaws and used them to gain root privileges https://aiaulis.com/en/juttu/anthropic-updating-claude-code-s-subagent-status-line-now-costs-about-71091dbad9f0/AI Aulisen2026-11-05T04:22:39.048846+00:00Anthropic: updating Claude Code’s subagent status line now costs about 99 per cent less https://aiaulis.com/en/juttu/moonshot-releases-kimi-k3-1-pro-model-weights-for-download-16da756e360c/AI Aulisen2026-11-05T04:22:39.048846+00:00Moonshot releases Kimi K3.1 Pro model weights for download https://aiaulis.com/en/juttu/us-midterm-voters-approve-dozens-of-local-data-centre-restrictions-f278e86ea1c5/AI Aulisen2026-11-05T04:22:39.048846+00:00US midterm voters approve dozens of local data centre restrictions https://aiaulis.com/en/juttu/chatgpt-now-makes-gif-animations-with-openai-s-images-2-6-image-model-c9a8dfc46ba4/AI Aulisen2026-11-05T04:22:39.048846+00:00ChatGPT now makes GIF animations with OpenAI’s Images 2.6 image model https://aiaulis.com/en/juttu/ricursive-says-ripple-65b-beats-larger-us-open-weight-models-on-a-bb1540b5a6a7/AI Aulisen2026-11-05T04:22:39.048846+00:00Ricursive says Ripple-65B beats larger US open-weight models on a coding benchmark https://aiaulis.com/en/juttu/artificial-analysis-now-measures-model-speed-by-time-per-task-72d8f1a5d560/AI Aulisen2026-11-05T04:22:39.048846+00:00Artificial Analysis now measures model speed by time per task https://aiaulis.com/en/juttu/cognition-releases-swe-2-flash-coding-model-at-what-it-says-is-one-ed9871d6a395/AI Aulisen2026-11-05T04:22:39.048846+00:00Cognition releases SWE-2 Flash coding model at what it says is one-fifth of SWE-2’s price https://aiaulis.com/en/juttu/llama-cpp-0-9-0-adds-fp8-quantisation-and-a-more-efficient-kv-cache-a5640419b10d/AI Aulisen2026-11-05T04:22:39.048846+00:00llama.cpp 0.9.0 adds FP8 quantisation and a more efficient KV cache for multiple sequences https://aiaulis.com/en/juttu/prime-intellect-some-public-benchmarks-are-nearly-saturated-after-8b8342d083b6/AI Aulisen2026-11-05T04:22:39.048846+00:00Prime Intellect: some public benchmarks are nearly saturated after fixes https://aiaulis.com/en/juttu/openai-to-buy-pierre-computer-whose-tools-codex-already-uses-09648be6fc3c/AI Aulisen2026-11-05T04:22:39.048846+00:00OpenAI to buy Pierre Computer, whose tools Codex already uses https://aiaulis.com/en/juttu/claude-code-product-leader-cat-wu-moves-from-anthropic-to-openai-0db1278d4a0e/AI Aulisen2026-11-05T04:22:39.048846+00:00Claude Code product leader Cat Wu moves from Anthropic to OpenAI https://aiaulis.com/en/juttu/meta-buys-conductor-whose-app-lets-developers-run-several-coding-6c03694b645d/AI Aulisen2026-11-05T04:22:39.048846+00:00Meta buys Conductor, whose app lets developers run several coding agents in parallel https://aiaulis.com/en/juttu/docker-sandboxes-flaw-lets-a-sandbox-redirect-copies-to-host-files-de44400dce10/AI Aulisen2026-11-05T04:22:39.048846+00:00Docker Sandboxes flaw lets a sandbox redirect copies to host files https://aiaulis.com/en/juttu/apple-cut-some-engineers-daily-ai-spending-cap-from-300-to-200-9d27cee01a62/AI Aulisen2026-11-05T04:22:39.048846+00:00Apple cut some engineers’ daily AI spending cap from $300 to $200, Bloomberg reports https://aiaulis.com/en/juttu/openclaw-beta-adds-knowledge-scribes-workflows-for-capturing-agent-61aee1ed166f/AI Aulisen2026-11-05T04:22:39.048846+00:00OpenClaw beta adds Knowledge Scribes workflows for capturing agent knowledge https://aiaulis.com/en/juttu/nous-research-brings-hermes-agent-plugins-and-skills-together-in-one-c7bd89b6414c/AI Aulisen2026-11-05T04:22:39.048846+00:00Nous Research brings Hermes Agent plugins and skills together in one marketplace https://aiaulis.com/en/juttu/check-point-haiou-jailbreak-bundle-for-codex-adds-an-advert-to-the-d0ef26c1cea0/AI Aulisen2026-11-05T04:22:39.048846+00:00Check Point: Haiou jailbreak bundle for Codex adds an advert to the agent’s replies https://aiaulis.com/en/juttu/developers-can-try-github-copilot-in-xcode-s-built-in-coding-e91c9d669799/AI Aulisen2026-11-05T04:22:39.048846+00:00Developers can try GitHub Copilot in Xcode’s built-in coding assistant in the Xcode 27.3 beta https://aiaulis.com/en/juttu/baseten-brings-agent-sandboxes-and-a-model-router-to-one-platform-e500aa4c1543/AI Aulisen2026-11-05T04:22:39.048846+00:00Baseten brings agent sandboxes and a model router to one platform https://aiaulis.com/en/juttu/humanlayer-which-has-written-about-how-to-work-with-coding-agents-03cec890eca2/AI Aulisen2026-11-05T04:22:39.048846+00:00HumanLayer, which has written about how to work with coding agents, says it is joining Factory https://aiaulis.com/en/juttu/tencent-releases-hy4-language-model-and-promises-weights-in-the-a68671dce844/AI Aulisen2026-11-05T04:22:39.048846+00:00Tencent releases Hy4 language model and promises weights in the coming days https://aiaulis.com/en/juttu/ft-oracle-and-tencent-drop-gpu-lease-deal-worth-an-estimated-7-billion-6419105c60ad/AI Aulisen2026-11-05T04:22:39.048846+00:00FT: Oracle and Tencent drop GPU lease deal worth an estimated $7 billion https://aiaulis.com/en/juttu/bloomberg-chinese-companies-expect-about-1-per-cent-of-their-rtx-pro-7f5a552c99e7/AI Aulisen2026-11-05T04:22:39.048846+00:00Bloomberg: Chinese companies expect about 1 per cent of their RTX Pro 5500 orders within four months https://aiaulis.com/en/juttu/gemini-notebook-api-in-preview-manage-notebooks-and-sources-from-code-051de2d4d377/AI Aulisen2026-11-05T04:22:39.048846+00:00Gemini Notebook API in preview: manage notebooks and sources from code https://aiaulis.com/en/juttu/ramp-its-post-trained-qwen-model-beat-opus-4-6-on-ramp-s-financial-a1c4f395456f/AI Aulisen2026-11-05T04:22:39.048846+00:00Ramp: its post-trained Qwen model beat Opus 4.6 on Ramp’s financial question task https://aiaulis.com/en/juttu/clients-of-accounting-firms-that-use-taxdome-can-ask-claude-or-f606bc303da9/AI Aulisen2026-11-05T04:22:39.048846+00:00Clients of accounting firms that use TaxDome can ask Claude or ChatGPT about their affairs https://aiaulis.com/en/juttu/harvey-plans-to-bring-jus-mundi-s-arbitration-agent-to-joint-8a5f9c672ec4/AI Aulisen2026-11-05T04:22:39.048846+00:00Harvey plans to bring Jus Mundi’s arbitration agent to joint customers this year https://aiaulis.com/en/juttu/anthropic-makes-its-first-news-licensing-deal-with-ap-adweek-reports-00fb6325f91f/AI Aulisen2026-11-05T04:22:39.048846+00:00Anthropic makes its first news licensing deal, with AP, Adweek reports https://aiaulis.com/en/juttu/coval-s-30-day-measurement-gemini-3-5-transcribe-live-sixth-for-af23e276acdb/AI Aulisen2026-11-05T04:22:39.048846+00:00Coval’s 30-day measurement: Gemini 3.5 Transcribe Live sixth for accuracy, 23rd for latency