", "via": "inbox", "person": "x-openainewsroom", "meaning": "OpenAI says it found a hidden recording device in one of its offices, used its own models to work out what it was and where it came from, and referred the matter to the FBI.", "concepts": [], "story": "2026-10-14-openai-engineer-arrested-for-allegedly-bugging-executive-con", "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" }, { "id": "bsky-emollick.bsky.social-3mxrncnjjjk2q", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxrncnjjjk2q", "posted_at": "2026-10-13T18:16:45Z", "text": "Using Gemini 4 makes you realize how many other models must be distilling Claude/GPT (or using a similar training pipeline). It feels quite different than other models.", "links": [], "context": null, "likes": 142, "reposts": 4, "via": "feed", "meaning": "Mollick observes that using Gemini 4 highlights how distinct its behavior is, suggesting many other models rely on distillation or similar training pipelines mimicking leading systems like Claude and GPT. Builders might care when evaluating model diversity and performance.", "concepts": [ "distillation", "foundation-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-13" }, { "id": "x-2110038898009407584", "platform": "x", "url": "https://x.com/sundarpichai/status/2110038898009407584", "posted_at": "2026-10-13T16:04:10Z", "date_only": false, "text": "Gemini 4 Argon is here.\n\nIt’s available globally today for developers and enterprises in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search.\n\nHope you all enjoy it!", "context": null, "embed_html": "We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re…
— OpenAI Newsroom (@OpenAINewsroom) October 13, 2026
", "via": "inbox", "person": "sundar-pichai", "meaning": "Google announces the release of Gemini 4 Argon, available globally for developers, enterprises, and subscribers across its platforms.", "concepts": [ "foundation-model", "multimodal-model" ], "story": "2026-10-13-google-releases-gemini-4-argon-frontier-model", "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-14" }, { "id": "bsky-emollick.bsky.social-3mxr77gpvik2q", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxr77gpvik2q", "posted_at": "2026-10-13T14:04:25Z", "text": "A worthwhile (and unnerving) reminder from Anthropic that we should be paying attention to AI ability thresholds, not just steady climbs\n\nIn this case, the threshold is the capability needed for a self-replicating AI worm on IoT devices like off-brand TVs www.anthropic.com/research/sel...", "links": [ { "text": "www.anthropic.com/research/sel...", "url": "https://www.anthropic.com/research/self-replicating-ai" } ], "context": null, "likes": 75, "reposts": 6, "via": "feed", "meaning": "Mollick highlights an Anthropic warning to focus on sudden AI ability thresholds rather than steady capability climbs, citing a threshold involving self-replicating AI worms on IoT devices. Developers and safety researchers care about identifying such critical capability jumps.", "concepts": [ "ai-safety", "alignment" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-13" }, { "id": "bsky-marcorus.bsky.social-3mxqgz6l6ho2h", "person": "marco-russo", "platform": "bluesky", "url": "https://bsky.app/profile/marcorus.bsky.social/post/3mxqgz6l6ho2h", "posted_at": "2026-10-13T06:51:25Z", "text": "Do early adopters win?\n\nFabric comes: Power BI is over.\nConversational BI: Dashboards and reports are dead.\nClaude dashboards: Was conversational BI a thing?\n\nCalm down.\n\nAs a professional, you want to stay up to date.\nAs a company, FOMO bleeds money.\n\nBe ready; don't rush.", "links": [], "context": null, "likes": 8, "reposts": 0, "via": "feed", "meaning": "Russo advises professionals and companies to stay updated on tools like Microsoft Fabric and conversational BI without panicking over rapid technology cycles or bleeding money from FOMO. Data professionals can benefit from balancing readiness with measured adoption.", "concepts": [ "microsoft-fabric", "copilot-in-power-bi" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-13" }, { "id": "bsky-emollick.bsky.social-3mxq4bhezas2q", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxq4bhezas2q", "posted_at": "2026-10-13T03:39:12Z", "text": "GPT-4 was the undisputed best model in the world for a full year and we all debated whether anyone could catch up or if OpenAI had some secret formula that nobody else knew about.\n\nNow there are are 19 AI models that can run on a modern smartphone that are better than GPT-4!", "links": [], "context": null, "likes": 202, "reposts": 30, "via": "feed", "meaning": "Mollick notes that while GPT-4 was the top model for a year, there are now 19 smartphone-capable AI models that outperform it. This illustrates rapid progress in edge and small language models that builders can leverage.", "concepts": [ "small-language-model", "foundation-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-13" }, { "id": "bsky-natolambert.bsky.social-3mxp3cejcy62p", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mxp3cejcy62p", "posted_at": "2026-10-12T17:49:09Z", "text": "If you announce a model's benchmarks without releasing it (weights, product, api, etc \"later\"), feels like almost a sure thing that someone releases a better model in your category before you get it out.\n\n... Gemini Argon, Reflection Beam, who's next on this list?", "links": [], "context": null, "likes": 23, "reposts": 0, "via": "feed", "meaning": "Lambert argues that announcing model benchmarks without immediate availability means competitors will likely release a better model first. Developers should watch launch strategies closely.", "concepts": [ "benchmark", "open-weight-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-12" }, { "id": "bsky-emollick.bsky.social-3mxoxbpm5nk2z", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxoxbpm5nk2z", "posted_at": "2026-10-12T16:37:12Z", "text": "If you haven’t made sure to turn on two-factor authentication on all accounts, have passwords with family in case you get voice calls asking for money, etc, now would be a good time.\n\nAI gains mean the costs of mass targeting individuals with very sophisticated attacks are dropping.", "links": [], "context": null, "likes": 123, "reposts": 27, "via": "feed", "meaning": "Mollick warns that AI tools make sophisticated targeted attacks cheaper, advising users to enable two-factor authentication and establish family verification passwords.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-12" }, { "id": "bsky-emollick.bsky.social-3mxnnvoyesk2a", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxnnvoyesk2a", "posted_at": "2026-10-12T04:16:45Z", "text": "GPT-6 Pro finds supposed errors in academic papers all the time, just as part of other work. Confirming the errors aren't nitpicks & learning their importance would still take a lot of time today, but it will get faster. We don't have the right mechanisms in place to deal a flood of corrections.", "links": [], "context": null, "likes": 88, "reposts": 6, "via": "feed", "meaning": "Mollick notes that AI models frequently spot errors in academic papers, but researchers lack efficient mechanisms to handle the resulting flood of corrections.", "concepts": [ "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-12" }, { "id": "x-2109361546505966046", "platform": "x", "url": "https://x.com/karpathy/status/2109361546505966046", "posted_at": "2026-10-11T19:12:37Z", "date_only": false, "text": "Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less \"two groups of people speaking past each other\" and more a sharp funnel.\n\n- Napkin math somewhere around 6B people (~75% of the population) have barely come… https://t.co/79xGhQdf1n", "context": null, "embed_html": "Gemini 4 Argon is here.
— Sundar Pichai (@sundarpichai) October 13, 2026
It’s available globally today for developers and enterprises in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search.
Hope you all enjoy it! pic.twitter.com/2jyAgqoylU
", "via": "inbox", "person": "andrej-karpathy", "meaning": "Karpathy says the gap in shared understanding of what LLMs can do is widening, from two groups talking past each other into a sharp funnel. The full post on X has his napkin math.", "concepts": [ "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" }, { "id": "bsky-marcorus.bsky.social-3mxlikvqqpn2w", "person": "marco-russo", "platform": "bluesky", "url": "https://bsky.app/profile/marcorus.bsky.social/post/3mxlikvqqpn2w", "posted_at": "2026-10-11T07:35:55Z", "text": "𝗖𝗼𝗱𝗲 𝗮𝘀 𝗮 𝗵𝗼𝗯𝗯𝘆\n\nThere are two types of developers, hyper-logical and hyper-creative, who feel very differently about AI writing their code.\nIt seems to me that how you feel about AI writing code depends on why you write code.\n\nRead my thoughts:", "links": [], "context": { "kind": "link", "url": "https://marcorusso.com/writing/code-as-a-hobby/", "title": "Code as a hobby — marcorusso.com", "text": "There are two types of developers, hyper-logical and hyper-creative, who feel very differently about AI writing their code." }, "likes": 3, "reposts": 0, "via": "feed", "meaning": "Russo observes that developer attitudes toward AI coding assistants depend on whether they view coding as a hyper-logical or hyper-creative hobby.", "concepts": [ "coding-assistant" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-11" }, { "id": "bsky-natolambert.bsky.social-3mxjv24xzzo2z", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mxjv24xzzo2z", "posted_at": "2026-10-10T16:13:51Z", "text": "We've been monitoring a decline in HuggingFace downloads across the top open LLMs we track for interconnects. It looks like there was a slight change in counting of downloads, as we saw a ~30% download rate drop for many top models, but this is continuing over time. \ndashboard.interconnects.ai", "context": null, "likes": 47, "reposts": 4, "via": "feed", "meaning": "Lambert notes a sustained decline in HuggingFace downloads across top open LLMs, initially linked to counting changes but continuing over time.", "concepts": [ "open-weight-model", "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "dashboard.interconnects.ai", "url": "https://dashboard.interconnects.ai/" } ] }, { "id": "x-2108931348857827686", "platform": "x", "url": "https://x.com/satyanadella/status/2108931348857827686", "posted_at": "2026-10-10T14:43:10Z", "date_only": false, "text": "https://t.co/MjwFy73LOF", "context": null, "embed_html": "Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
— Andrej Karpathy (@karpathy) October 11, 2026
- Napkin math somewhere around 6B people (~75% of the population) have barely come… https://t.co/79xGhQdf1n
— Satya Nadella (@satyanadella) October 10, 2026", "via": "inbox", "person": "satya-nadella", "note": "Links to his essay Models as Insider Risks in the Super Intelligence Era", "meaning": "Nadella shares his essay arguing that companies should treat AI models as possible insider risks. Our story on what the essay says is linked below.", "concepts": [ "ai-safety" ], "story": "2026-10-11-microsoft-ceo-satya-nadella-says-ai-models-are-compromised", "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" }, { "id": "bsky-natolambert.bsky.social-3mxhwx5ydch2t", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mxhwx5ydch2t", "posted_at": "2026-10-09T21:42:40Z", "text": "I’m more bought into AI becoming far more efficient, which is a massive lever on diffusion, than I am on AI models making massive breakthroughs that result in eliminating all the known weaknesses of LLMs.\nwww.interconnects.ai/p/i-expect-r...", "context": null, "likes": 46, "reposts": 2, "via": "feed", "meaning": "Lambert argues that AI becoming more efficient is a much larger lever than models making massive breakthroughs that eliminate all known weaknesses of LLMs.", "concepts": [ "llm", "scaling-laws" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "www.interconnects.ai/p/i-expect-r...", "url": "https://www.interconnects.ai/p/i-expect-rapid-progress-but-not-towards" } ] }, { "id": "bsky-emollick.bsky.social-3mxhfe5axps2o", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxhfe5axps2o", "posted_at": "2026-10-09T16:27:48Z", "text": "Randomized trials with GPT-4o: \"AI access raises test scores, and a smaller gain persists a week later. Gains remain for students who use AI as a tutor (“augmentation”) and fade for students who have AI write for them (“automation”)\"\n\nAcross all experiments, positive effects arxiv.org/abs/2607.08849", "context": null, "likes": 68, "reposts": 12, "via": "feed", "meaning": "Mollick shares that randomized trials show AI access raises test scores, with gains persisting when used as a tutor but fading when the AI writes for the student.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "arxiv.org/abs/2607.08849", "url": "https://arxiv.org/abs/2607.08849" } ] }, { "id": "bsky-simonwillison.net-3mxh2nlvc7s26", "person": "simon-willison", "platform": "bluesky", "url": "https://bsky.app/profile/simonwillison.net/post/3mxh2nlvc7s26", "posted_at": "2026-10-09T13:16:14Z", "text": "I built a new feature for my blog entirely by voice with Codex Desktop, while I was cooking dinner simonwillison.net/2026/Oct/9/b...", "context": { "kind": "link", "url": "https://simonwillison.net/2026/Oct/9/built-using-my-voice/", "title": "A new feature for my blog, built using my voice", "text": "I shipped a new feature for my blog today: the Newsletters page, which offers an index of all of the newsletters I’ve sent out, both my free weekly Substack and …" }, "likes": 63, "reposts": 3, "via": "feed", "meaning": "Willison built a new feature for his blog entirely by voice using Codex Desktop while cooking dinner.", "concepts": [ "coding-assistant" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "simonwillison.net/2026/Oct/9/b...", "url": "https://simonwillison.net/2026/Oct/9/built-using-my-voice/" } ] }, { "id": "bsky-emollick.bsky.social-3mxg56pvsvc2c", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxg56pvsvc2c", "posted_at": "2026-10-09T04:28:56Z", "text": "From The Lancet: in an urgent care setting, the advice of the obsolete Gemini 2.5 Pro & Gemini 2.5 Flash (without access to patient medical records) were rated of similar quality to doctors by other physicians. There was no safety issues spotted.\n\nModels have gotten significantly better since.", "context": null, "likes": 107, "reposts": 17, "via": "feed", "meaning": "Mollick highlights a Lancet study where urgent care advice from older Gemini models without patient records was rated of similar quality to doctors by other physicians.", "concepts": [ "multimodal-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "x-2108404561413349695", "platform": "x", "url": "https://x.com/claudeai/status/2108404561413349695", "posted_at": "2026-10-09T03:49:54Z", "date_only": false, "text": "We’ve loved seeing so many founders join the Claude Startups program. Unfortunately, we underestimated demand for the program, and we need to pause the Claude Team and $1,000 API credit offers.\n\nWith hundreds of thousands of applicants, we're re-reviewing applications so the…", "context": null, "embed_html": "
", "via": "inbox", "person": "x-claudeai", "meaning": "Anthropic says demand for its Claude Startups program was far higher than it expected, so it is pausing the Claude Team and $1,000 API credit offers while it re-reviews applications.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" }, { "id": "bsky-natolambert.bsky.social-3mxerqsbr5s2t", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mxerqsbr5s2t", "posted_at": "2026-10-08T15:31:38Z", "text": "The fact that all the AI news is so bizarre, kids vibe hacking with claude code, governments vibe governing, all massively increases the uncertainty of what's coming with AI. The meme-est AI timeline is also one of the scariest.", "context": null, "likes": 37, "reposts": 2, "via": "feed", "meaning": "Lambert comments that bizarre AI news and informal usage increase uncertainty and make the current timeline feel scary.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "x-2108216753604248000", "platform": "x", "url": "https://x.com/simonw/status/2108216753604248000", "posted_at": "2026-10-08T15:23:37Z", "date_only": false, "text": "New open source cross-platform (Windows, macOS, Linux) sandboxing library from Microsoft - looks very promising, uses processcontainer/bubblewrap/seatbelt under the hood https://t.co/WMW0AXeBoG https://t.co/1xXIohaD5g", "context": null, "embed_html": "We’ve loved seeing so many founders join the Claude Startups program. Unfortunately, we underestimated demand for the program, and we need to pause the Claude Team and $1,000 API credit offers.
— Claude (@claudeai) October 9, 2026
With hundreds of thousands of applicants, we're re-reviewing applications so the…
", "via": "inbox", "person": "simon-willison", "meaning": "Willison highlights a new cross-platform sandboxing library from Microsoft that utilizes processcontainer, bubblewrap, and seatbelt under the hood.", "concepts": [ "agent-sandbox" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-14" }, { "id": "bsky-emollick.bsky.social-3mxeh745nhk2x", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxeh745nhk2x", "posted_at": "2026-10-08T12:22:47Z", "text": "As a researcher who did early some work on the productivity impacts of AI chatbots using RCTs, I’d note a lack of similar studies since the dawn of true agents last fall\n\nPartially that is newness & partially research design challenges, but I suspect we are missing some large & important effects", "context": null, "likes": 47, "reposts": 1, "via": "feed", "meaning": "Mollick notes a lack of recent randomized controlled trials studying the productivity impacts of true AI agents compared to early chatbot studies.", "concepts": [ "agent" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-emollick.bsky.social-3mxd2yu3fnc2h", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxd2yu3fnc2h", "posted_at": "2026-10-07T23:11:53Z", "text": "Anyhow, I think we are now done with the last round of skeptical responses to AI after Navier-Stokes. Yes, AI is now capable of producing novel proofs to hard problems that have eluded us\n\nThe next line will be that this is only constrained to math because of RL training. I suspect not, we will see.", "context": null, "likes": 102, "reposts": 7, "via": "feed", "meaning": "Mollick claims AI is now capable of producing novel proofs to hard problems, suggesting skepticism regarding math constraints is ending.", "concepts": [ "reinforcement-learning", "reasoning-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-emollick.bsky.social-3mxczf26uy22h", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxczf26uy22h", "posted_at": "2026-10-07T22:42:54Z", "text": "Some early first-hand accounts of the experience of encountering a narrow superhuman intelligence as mathematicians grapple with the hundreds of big AI proofs released by OpenAI.\n\nProblems solved in inhuman ways that make us wonder what it means to actually know things... scottaaronson.blog?p=10169", "context": null, "likes": 101, "reposts": 15, "via": "feed", "meaning": "Mollick shares early accounts of mathematicians grappling with hundreds of major AI-generated proofs solved in inhuman ways.", "concepts": [ "reasoning-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "scottaaronson.blog?p=10169", "url": "https://scottaaronson.blog/?p=10169" } ] }, { "id": "bsky-simonwillison.net-3mxctkjruk22c", "person": "simon-willison", "platform": "bluesky", "url": "https://bsky.app/profile/simonwillison.net/post/3mxctkjruk22c", "posted_at": "2026-10-07T20:58:36Z", "text": "Pelicans and a bunch of notes on the pricing of the new Claude Haiku 5.5 simonwillison.net/2026/Oct/7/c...", "context": { "kind": "link", "url": "https://simonwillison.net/2026/Oct/7/claude-haiku-5-5/", "title": "Introducing Claude Haiku 5.5", "text": "As previously promised, here’s Anthropic’s new fast, low cost model: Introducing Claude Haiku 5.5. The previous Haiku, 4.5, was very much showing its age. It came out almost a year …" }, "likes": 80, "reposts": 4, "via": "feed", "meaning": "Willison shares notes and pricing details on the newly released Claude Haiku 5.5 model from Anthropic, highlighting its low cost and speed compared to its predecessor.", "concepts": [ "llm", "foundation-model" ], "story": "2026-10-08-anthropic-releases-claude-haiku-5-5-and-aws-integration", "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "simonwillison.net/2026/Oct/7/c...", "url": "https://simonwillison.net/2026/Oct/7/claude-haiku-5-5/" } ] }, { "id": "x-2107894039626277339", "platform": "x", "url": "https://x.com/claudeai/status/2107894039626277339", "posted_at": "2026-10-07T18:01:16Z", "date_only": false, "text": "Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model we’ve ever released.\n\nOn average, it costs around 75% less to run than Claude Haiku 4.5.", "context": null, "embed_html": "New open source cross-platform (Windows, macOS, Linux) sandboxing library from Microsoft - looks very promising, uses processcontainer/bubblewrap/seatbelt under the hood https://t.co/WMW0AXeBoG https://t.co/1xXIohaD5g
— Simon Willison (@simonw) October 8, 2026
", "via": "inbox", "person": "x-claudeai", "meaning": "Anthropic introduces Claude Haiku 5.5, promoting it as their cheapest, fastest, and most capable small model while costing about 75 percent less to run than the previous version.", "concepts": [ "small-language-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-14" }, { "id": "bsky-emollick.bsky.social-3mxbcxh52gs2c", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxbcxh52gs2c", "posted_at": "2026-10-07T06:28:56Z", "text": "When people talk about AI easily replacing workers, I think about this ethnography of copier repair technicians in the 1990s.\n\nReading it, you see how work is complicated and improvised by workers and informal and badly documented.\n\n(But eventually technology did largely eliminate this job)", "context": null, "likes": 68, "reposts": 4, "via": "feed", "meaning": "Mollick compares AI workplace replacement fears to historical studies of copier repair technicians, noting that while work is often improvised and poorly documented, technology eventually eliminates those jobs.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-emollick.bsky.social-3mxasq2guls2c", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mxasq2guls2c", "posted_at": "2026-10-07T01:38:28Z", "text": "After a brief but institution-eroding slop science era, it increasingly looks like we are going to have 2 revolutions from AI:\n1) Everything ever published will be re-read and re-judged in ways that human scientists never anticipated\n2) Novel discoveries will start to come faster than we can absorb", "context": null, "likes": 163, "reposts": 11, "via": "feed", "meaning": "Mollick argues that AI will drive two major revolutions: enabling the re-evaluation of all published science in unforeseen ways, and accelerating novel discoveries past our capacity to absorb them.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-simonwillison.net-3mxap2pofos2t", "person": "simon-willison", "platform": "bluesky", "url": "https://bsky.app/profile/simonwillison.net/post/3mxap2pofos2t", "posted_at": "2026-10-07T00:32:51Z", "text": "I released an LLM plugin for sending prompts to OpenAI's new Jev-clone decision model that's based on GPT-6 Luna simonwillison.net/2026/Oct/6/l...", "context": { "kind": "link", "url": "https://simonwillison.net/2026/Oct/6/llm-openai-decisions/", "title": "Release: llm-openai-decisions 0.1a0", "text": "LLM plugin for the OpenAI Decisions API" }, "likes": 39, "reposts": 3, "via": "feed", "meaning": "Willison released an LLM plugin to send prompts to OpenAI's new decision model based on GPT-6 Luna.", "concepts": [ "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "simonwillison.net/2026/Oct/6/l...", "url": "https://simonwillison.net/2026/Oct/6/llm-openai-decisions/" } ] }, { "id": "bsky-emollick.bsky.social-3mx7q4lupt22f", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mx7q4lupt22f", "posted_at": "2026-10-06T15:19:08Z", "text": "I think most people who have been closely watching AI for the past few years are likely surprised by how relatively rare terrible AI incidents have been, given a billion users overall and hundreds of millions of corporate users. Doesn’t mean that will continue, but still its surprising.", "context": null, "likes": 179, "reposts": 11, "via": "feed", "meaning": "Mollick expresses surprise that relatively few terrible AI incidents have occurred given the hundreds of millions of corporate and consumer users.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-natolambert.bsky.social-3mx7n6zwwus2z", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mx7n6zwwus2z", "posted_at": "2026-10-06T14:26:48Z", "text": "Too many people are analyzing the cyber risks of open models in a narrow lens, assuming China doesn't care about safety, and fear mongering about private information on open-weight cyber attacks. The public information we have paints a very different picture.\nwww.interconnects.ai/p/the-cyber-...", "context": { "kind": "link", "url": "https://www.interconnects.ai/p/the-cyber-risk-discourse-is-broken", "title": "The Cyber Risk Discourse is Broken", "text": "Open-weights, ideology, and acknowledging trade-offs." }, "likes": 23, "reposts": 5, "via": "feed", "meaning": "Lambert argues that current public discussions around the cyber risks of open-weight models rely on narrow assumptions and fear-mongering, and points to his article for a broader analysis.", "concepts": [ "open-weight-model", "ai-safety" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "www.interconnects.ai/p/the-cyber-...", "url": "https://www.interconnects.ai/p/the-cyber-risk-discourse-is-broken" } ] }, { "id": "bsky-marcorus.bsky.social-3mx76lefq5u2l", "person": "marco-russo", "platform": "bluesky", "url": "https://bsky.app/profile/marcorus.bsky.social/post/3mx76lefq5u2l", "posted_at": "2026-10-06T10:05:16Z", "text": "Defining the right contract is a key activity for a DAX developer using AI.\nIn this article and video, we introduce the concept of a contract through simple examples.\n\nhttps://www.sqlbi.com/articles/understanding-dax-contracts-when-using-ai/\n\n#dax #ai", "links": [ { "text": "https://www.sqlbi.com/articles/understanding-dax-contracts-when-using-ai/", "url": "https://www.sqlbi.com/articles/understanding-dax-contracts-when-using-ai/" } ], "context": null, "likes": 1, "reposts": 0, "via": "feed", "meaning": "Russo explains that defining the right contract is a key activity when using AI for DAX development, sharing examples in a new article and video.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b", "added": "2026-10-11" }, { "id": "bsky-emollick.bsky.social-3mx63cx3bjs2o", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mx63cx3bjs2o", "posted_at": "2026-10-05T23:34:13Z", "text": "Big swing happening from this summer's trend to have AI running on virtual machines on your own computer (Claude and ChatGPT apps) to moving them back into you having a dedicated VM the cloud (Claude Cowork just announced they were doing this, also Grok Bot, Dot, and others). Cloud is mostly better.", "context": null, "likes": 65, "reposts": 4, "via": "feed", "meaning": "Mollick highlights a trend of AI applications shifting from running locally on virtual machines back to dedicated cloud virtual machines.", "concepts": [], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-natolambert.bsky.social-3mx5sodby3b26", "person": "nathan-lambert", "platform": "bluesky", "url": "https://bsky.app/profile/natolambert.bsky.social/post/3mx5sodby3b26", "posted_at": "2026-10-05T20:59:31Z", "text": "Reflection joins the list of Nvidia and Thinking Machines who have tried to release their strongest models and come up behind Chinese counterparts. There's a lot of ways people will overthink this, but the clearest takeaway should be that the Chinese are very very good at building LLMs.", "context": null, "likes": 83, "reposts": 8, "via": "feed", "meaning": "Lambert notes that several labs have released strong models that fall behind Chinese counterparts, demonstrating that Chinese developers are exceptionally skilled at building LLMs.", "concepts": [ "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-emollick.bsky.social-3mx5qe4df2k2c", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mx5qe4df2k2c", "posted_at": "2026-10-05T20:18:01Z", "text": "AI policy right now, especially from the Labs, makes me think of this poem.\n\nIf you believe superintelligence is near, you don't need to make hard decisions about how to build AI to make the world better. Just wait for ASI & it will decide for us, for better or worse. But if that doesn't happen...", "context": null, "likes": 61, "reposts": 5, "via": "feed", "meaning": "Mollick criticizes current AI lab policies, suggesting that belief in near-term superintelligence lets labs avoid making hard present-day decisions about building better AI.", "concepts": [ "agi" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [] }, { "id": "bsky-emollick.bsky.social-3mx2prvytik2c", "person": "ethan-mollick", "platform": "bluesky", "url": "https://bsky.app/profile/emollick.bsky.social/post/3mx2prvytik2c", "posted_at": "2026-10-04T15:29:51Z", "text": "\"create the ultimate brutalist city builder. you should have lots of controls over buildings in ways that are delightful and intuitive to people...\"\n\nHere is Fable 5.1, a little over a year since GPT-5 below & a single shot (compared with many rounds with GPT-5.6)\n\nQuite good: brut-city.netlify.app", "context": { "kind": "quote", "url": "", "title": "Quoted post by @emollick.bsky.social", "text": "When GPT-5 came out, I created a procedural brutalist city builder as a demo (you can see it at the video's start)\n\nI used GPT-5.6 Sol in Codex to do the same thing, touching no code.\n\nLess than a year...\n\nPlay with it (its fun, if you like architecture): monument-brutalist-city-builder.netlify.app" }, "likes": 64, "reposts": 7, "via": "feed", "meaning": "Mollick demonstrates how a coding assistant built a brutalist city builder app in a single shot using newer models compared to older versions, showcasing rapid yearly progress.", "concepts": [ "coding-assistant" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:qwen/qwen3.8-27b", "added": "2026-10-11", "links": [ { "text": "brut-city.netlify.app", "url": "https://brut-city.netlify.app/" } ] }, { "id": "x-1886192184808149383", "platform": "x", "url": "https://x.com/karpathy/status/1886192184808149383", "posted_at": "2025-02-02T23:17:15Z", "date_only": false, "text": "There's a new kind of coding I call \"vibe coding\", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper…", "context": null, "embed_html": "Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model we’ve ever released.
— Claude (@claudeai) October 7, 2026
On average, it costs around 75% less to run than Claude Haiku 4.5. pic.twitter.com/jm06cZJfkV
", "via": "inbox", "person": "andrej-karpathy", "note": "classic: coined the term vibe coding", "meaning": "Karpathy coins the term vibe coding: you describe what you want, accept what the AI writes, and stop looking at the code itself. He says it works because LLMs in tools like Cursor have become that good.", "concepts": [ "coding-assistant", "llm", "vibe-coding" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "groq:openai/gpt-oss-120b (note edited by the editor)", "added": "2026-10-14" }, { "id": "x-1790075827666796666", "platform": "x", "url": "https://x.com/sama/status/1790075827666796666", "posted_at": "2024-05-13T17:45:09Z", "date_only": false, "text": "her", "context": null, "embed_html": "There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper…
— Andrej Karpathy (@karpathy) February 2, 2025
", "via": "inbox", "person": "sam-altman", "note": "classic: posted on the day of the GPT-4o launch and widely read as a nod to the film Her", "meaning": "The whole post is one word. It went up on the day OpenAI launched GPT-4o, whose voice mode many people compared to the AI in the film Her.", "concepts": [ "multimodal-model" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" }, { "id": "x-1617979122625712128", "platform": "x", "url": "https://x.com/karpathy/status/1617979122625712128", "posted_at": "2023-01-24T20:14:18Z", "date_only": false, "text": "The hottest new programming language is English", "context": null, "embed_html": "her
— Sam Altman (@sama) May 13, 2024
", "via": "inbox", "person": "andrej-karpathy", "note": "classic: an early line that summed up prompting as programming", "meaning": "Karpathy's point: with large language models you can program a computer by writing plain English, so writing prompts becomes a kind of programming.", "concepts": [ "prompt-engineering", "llm" ], "story": null, "model": "gemini:gemini-3.5-flash-lite", "checked": "editor", "added": "2026-10-14" } ] }The hottest new programming language is English
— Andrej Karpathy (@karpathy) January 24, 2023