Model Follower.

Up to date

Rebuilt
2026-10-26 07:36 UTC
Release radar
116 on watch
Hub snapshot
2026-10-26
Re-checked
daily · every number sourced and dated

Links / Past days

Saturday, 3 October 2026

The AI links the most newsletters and link sources had shared by the end of the day (UTC), each source counted once.

As of 2026-10-03 23:59 UTC Status sealed Rebuilt from issues on file

Reconstructed after the day from the issues on file (dates as published). An issue counts on the date it carries, even if it was collected later.

That day 36-hour window

  1. Shared by 1

    Google Cloud launched a similar feature in July, called Spend Caps, which lets you "set a monthly financial cap on…

    Reported bySimon Willison

    cloud.google.com First shared 2026-10-03
  2. Shared by 1

    I just sent the September edition of my sponsors-only monthly newsletter.

    Reported bySimon Willison

    github.com First shared 2026-10-03
  3. Shared by 1

    If you are a sponsor (or start a sponsorship now) you can access it here.

    “I just sent the September edition of my sponsors-only monthly newsletter. If you are a sponsor (or start a sponsorship now) you can access it here.” via Simon Willison

    Reported bySimon Willison

    github.com First shared 2026-10-03
  4. Shared by 1

    Here's a copy of the August newsletter as a preview of what you'll get.

    Reported bySimon Willison

    github.com First shared 2026-10-03
  5. Shared by 4

    DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

    “DeepSeek's V4.1-Flash paper gets long-context KV cache down to 890 bytes per token at 1M context, a quarter of the prior generation's footprint (arXiv)” via AI Breakfast

    Reported byAI Breakfast, Turing Post, The Batch, Last Week in AI

    AboutDeepSeek-V4.1-Flash

    How 1 source put it
    arxiv.org First shared 2026-09-19
  6. Shared by 1

    OpenAI safety leader quits, warning AI company's culture is 'broken'

    Discussed onHacker News 268 pts · 3 comments

    theguardian.com First shared 2026-10-03
  7. Shared by 2

    I made my iPhone a second GPU for my 24 GB MacBook: Qwen 3.8 27B prefills 29–44% faster & my holds part of the CTX window.

    “(Activity: 1192): OP built backburner, a llama.cpp fork / distributed inference setup that offloads part of Qwen 3.8 27B IQ4_XS from a 24 GB M4 Pro MacBook to an iPhone 17 Pro Max over 10 Gb/s USB-C: the Mac runs layers 1–40, streams…” via Latent Space / AINews

    Reported byLatent Space / AINews

    Discussed onr/LocalLLaMA 2,293 upvotes · 328 comments

    reddit.com Reddit thread First shared 2026-10-02
  8. Shared by 1

    The work by Valve's Timur Kristóf on improving old AMD GPUs on Linux

    Discussed onHacker News 499 pts · 109 comments

    phoronix.com First shared 2026-10-03
  9. Shared by 2

    Qwen3.8-27B-Humanlike-Chat 2.0: texts like a human, now with tool calls and better instruction following

    “(Activity: 805): LessThanThreeAI released Qwen3.8-27B-Humanlike-Chat 2.0, a merged LoRA over huihui-ai’s abliterated Qwen3.8-27B, available as GGUF/BF16/LoRA on Hugging Face with a demo Space. v2 replaces plain SFT with on-policy distillation: the…” via Latent Space / AINews

    Reported byLatent Space / AINews

    Discussed onr/LocalLLaMA 940 upvotes · 191 comments

    AboutQwen3.8-27B

    reddit.com Reddit thread First shared 2026-10-02
  10. Shared by 1

    We checked 12 subreddits, 544 Twitters and no further Discords.

    “AI News for 9/22/2026-9/23/2026. We checked 12 subreddits, 544 Twitters and no further Discords. AINews’ website lets you search all past issues. As a reminder, AINews is now a section of Latent Space. You can opt in/out of email frequencies!” via Latent Space / AINews

    Reported byLatent Space / AINews

    x.com First shared 2026-09-24
  11. Shared by 2

    New in llama.cpp: Decision Models

    “(Activity: 574): The post announces Decision Models support in llama.cpp: local “Jev/Jeff-like” models intended to act more like controllers/classifiers—selecting among actions, continuations, or behavioral choices—rather than purely free-form generators. No benchmark numbers or low-level implementation details were…” via Latent Space / AINews

    Reported byLatent Space / AINews

    Discussed onr/LocalLLaMA 501 upvotes · 122 comments

    reddit.com Reddit thread First shared 2026-10-02
  12. Shared by 1

    promo code LWAI takes an additional 15% off any pass.

    “— promo code LWAI takes an additional 15% off any pass.” via Last Week in AI

    Reported byLast Week in AI

    odsc.ai First shared 2026-09-17
  13. Shared by 1

    Get started at langfuse.com

    “; generous free tier, no credit card required.” via Last Week in AI

    Reported byLast Week in AI

    langfuse.com First shared 2026-09-17
  14. Shared by 1

    LeCun has "zero concerns" about AI wiping out humanity, recent "rogue" incidents

    Discussed onHacker News 423 pts · 846 comments

    fortune.com First shared 2026-10-03
  15. Shared by 2

    Read the full article here

    “The Economist address the effective altruism (EA) belief system, which it calls “this century’s biggest idea”:” via Exponential View

    Reported byForward Future, Exponential View

    How 1 source put it
    economist.com First shared 2026-10-02
  16. Shared by 1

    The company says drones will mainly take over mid-distance orders (3 to 5 miles) that normally slow down Dashers, while…

    “…ification, clearing it to run its own drone delivery fleet, DoorDash Air, on top of its existing marketplace. The company says drones will mainly take over mid-distance orders (3 to 5 miles) that normally slow down Dashers, while human drivers keep handling shorter, trickier deliveries. DoorDash says drones averaged 25-minute deliveries in 2025. You can see…” via Superhuman AI

    Reported bySuperhuman AI

    about.doordash.com First shared 2026-10-03
  17. Shared by 4

    Getting the most out of Opus 5.5 in Claude and Claude Code

    “📝 Opus Playbook: Anthropic published a blog on getting the most out of Opus 5.5, and it racked up 5K bookmarks within 2 hours. It covers prompting the new model, steering long tasks, checking your results, and more.” via Superhuman AI

    Reported bySuperhuman AI, The Neuron, Don't Worry About the Vase

    Discussed onHacker News 249 pts · 161 comments

    How 1 source put it
    claude.dev First shared 2026-09-23
  18. Shared by 1

    Agents don't need memory, they need documentation

    Discussed onHacker News 383 pts · 309 comments

    liao.gg First shared 2026-10-03
  19. Shared by 2

    Pi now supports MCP

    “, and makes a case for why after saying it never would.” via Ben's Bites

    Reported byBen's Bites, Latent Space / AINews

    How 1 source put it
    earendil.com First shared 2026-10-01
  20. Shared by 1

    (Activity: 1192): OP built backburner, a llama.cpp fork / distributed inference setup that offloads part of Qwen 3.8…

    “… window. (Activity: 1192): OP built backburner, a llama.cpp fork / distributed inference setup that offloads part of Qwen 3.8 27B IQ4_XS from a 24 GB M4 Pro MacBook to an iPhone 17 Pro Max over 10 Gb/s USB-C: the Mac runs layers 1–40, streams activations, and the phone runs layers 41–64 using Metal 4 tensor ops. Reported end-to-end prefill gains vs Mac-only…” via Latent Space / AINews

    Reported byLatent Space / AINews

    github.com First shared 2026-10-03

The week to that day two or more sources

  1. Shared by 9

    Gemini 4 Argon: our next era of frontier intelligence

    Google announces Gemini 4 Argon, a frontier model launching at $2 per million input tokens and $10 per million output tokens. Details internal wins like a 40% quantum optimization beat, 1M token output limits, and a 91.7% LVBench score. cited

    Reported byLatent Space / AINews, The Rundown AI, Superhuman AI, TLDR AI, Ben's Bites, Don't Worry About the Vase, The Neuron, AI Breakfast

    Discussed onHacker News 1,702 pts · 1,190 comments

    Where it was shared
    blog.google announcement · 6 min read First shared 2026-09-30
  2. Shared by 11

    More on the model here

    “: An upgrade for coding, computer use, and professional work. OpenAI claims near-Astra intelligence at one-fifth of Astra’s standard input and output token prices. Available through the API and listed paid ChatGPT plans.” via Turing Post

    Reported bySimon Willison, Turing Post, Latent Space / AINews, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Data Points, Ben's Bites, Don't Worry About the Vase

    Discussed onHacker News 1,067 pts · 955 comments

    How 1 source put it
    openai.com First shared 2026-09-29
  3. Shared by 10

    Dots. OpenAI has given us… dots

    “OpenAI’s DevDay gives us even more reasons to question why would we need a laptop. Dots, its new personal agents, have their own cloud computers, browsers, and connections to your applications. OpenAI says they can continue working on projects and carry context across ChatGPT, Slack, and Teams. And soon texts, as well.” via Turing Post

    Reported byTuring Post, Latent Space / AINews, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Data Points, Ben's Bites, Don't Worry About the Vase

    Discussed onHacker News 769 pts · 647 comments

    How 1 source put it
    openai.com First shared 2026-09-29
  4. Shared by 9

    GLM-5.3 and the spread of advanced cyber capabilities

    Anthropic reports attackers bypass GLM-5.3 safeguards 64% to 100% of the time using simple techniques. Includes ExploitBench results showing 50 of 410 end-to-end exploits and details the $4,400 cost to abliterate the model. cited

    Reported bySimon Willison, Latent Space / AINews, Last Week in AI, TLDR AI, AI Breakfast, The Neuron, Don't Worry About the Vase, The Batch

    Discussed onHacker News 256 pts · 241 comments

    AboutGLM-5.3-Flash, GLM-5.3, GLM-5

    Where it was shared
    anthropic.com analysis · 8 min read First shared 2026-09-29
  5. Shared by 9

    Introducing Claude Sonnet 5.5

    Anthropic launches Claude Sonnet 5.5, a 30% faster model that costs up to 30% less per task than its predecessor. Includes benchmark tables showing 70.6% on Terminal-Bench 4.0 and pricing details for cache reads at $0.20 per million tokens. cited

    Reported bySimon Willison, Latent Space / AINews, The Rundown AI, The Neuron, TLDR AI, Ben's Bites, Turing Post, AI Breakfast, Don't Worry About the Vase

    Where it was shared
    anthropic.com announcement · 8 min read First shared 2026-09-28
  6. Shared by 5

    The Future Is for Everyone: Muse for Small Business

    Meta expands its personal AI agent with new skills and connectors to help small businesses run operations. Includes connectors for Instagram, Facebook, and Meta ads, plus integrations with Canva, Shopify, Stripe, and QuickBooks. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, Ben's Bites

    Where it was shared
    about.fb.com announcement · 2 min read First shared 2026-09-30
  7. Shared by 6

    OpenAI puts a 24/7 agent inside ChatGPT

    “: Luna answers developer-defined questions with finite answer choices for classification, routing, and action selection. Limited preview, with broader release planned shortly.” via Turing Post

    Reported byTuring Post, The Rundown AI, The Neuron, Superhuman AI, AI Breakfast, Ben's Bites

    How 1 source put it
    openai.com First shared 2026-09-29
  8. Shared by 3

    Tavus’ Griffin passes for human on video calls

    “It’s called Griffin. And instead of the usual AI stack where speech, an LLM, voice, and avatar animation take turns, Griffin watches your video and listens while it’s talking.” via The Neuron

    Reported byThe Rundown AI, The Neuron, Superhuman AI

    How 1 source put it
    tavus.io First shared 2026-10-02
  9. Shared by 3

    The Dot and the Swarm

    “Instead, it feels much more tool-dependent. You need it to connect to tools where your life happens to feel the magic of Dots. Your calendar and email are the start. But think Slack, Notion, Jira, your messaging apps and more. Your Dot should take it from there. I think this post from Ethan explains this shift better.” via Ben's Bites

    Reported byBen's Bites, The Neuron, TLDR AI

    How 1 source put it
    • Ben's Bites2026-10-01issue ↗ Instead, it feels much more tool-dependent. You need it to connect to tools where your life happens to feel the magic of Dots. Your calendar and email are the start. But think Slack, Notion, Jira, your messaging apps and more. Your Dot should take it from there. I think this post from Ethan explains this shift better.
    • The Neuron2026-10-02issue ↗
    • TLDR AI2026-10-02issue ↗
    oneusefulthing.org First shared 2026-10-01
  10. Shared by 3

    Our first streaming transcription model debuts at no. 1 on Artificial Analysis

    Microsoft launches MAI-Transcribe-2-Streaming, ranking no. 1 on Artificial Analysis for final and partial transcript accuracy. Includes first hypotheses in just over 100ms and an introductory price of $0.54 per hour of audio through the end of the year. cited

    Reported byThe Rundown AI, TLDR AI, AI Breakfast

    Where it was shared
    microsoft.ai announcement · 3 min read First shared 2026-10-02