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

Tuesday, 20 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 2026-10-20 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 4

    of Homeland Security formally banned DeepSeek from civilian agency systems, giving departments 60 days to request…

    DHS excludes DeepSeek products from federal civilian procurements and removes them from agency IT systems. The order also mandates removing third-party products and services developed using DeepSeek implementations from civilian agency information technology. cited

    Reported byLatent Space / AINews, The Rundown AI, The Neuron, Interconnects

    Where it was shared
    public-inspection.federalregister.gov news report · 1 min read First shared 2026-10-20
  2. Shared by 4

    Diffusion at Scale: Introducing Mercury 3

    Inception launches Mercury 3, a diffusion LLM that runs 4–8x faster at 1,000+ tokens per second than GPT-6 Luna and Claude Haiku 5.5. Details pricing at $0.20 per million input and $0.70 per million output, plus a 50% launch discount and 1M token context window. cited

    Reported byThe Rundown AI, The Neuron, TLDR AI

    Discussed onHacker News 257 pts · 92 comments

    Where it was shared
    inceptionlabs.ai announcement · 4 min read First shared 2026-10-19
  3. Shared by 3

    MiniMax M3.1 Flash: Redefining Agent Performance and Cost

    MiniMax launches M3.1 Flash, an agent model that reduces usage costs by 70% while scoring 76.2 on BankerToolBench. Details flat pricing of $0.15 input and $0.60 output per 1M tokens alongside 200+ TPS output speeds. cited

    Reported byThe Rundown AI, The Neuron, TLDR AI

    Where it was shared
    minimax.io announcement · 2 min read First shared 2026-10-20
  4. Shared by 7

    Investing in the public benefit

    Anthropic commits 5% of shares to a new Public Benefit Fund, adding 2% for every 50% valuation increase. Details how the LLC avoids lobbying, funds labor impact studies, and combines with founder pledges to reach 20% total public good equity. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Latent Space / AINews, Forward Future

    Where it was shared
    anthropic.com announcement · 2 min read First shared 2026-10-19
  5. Shared by 2

    (tweet)Join the Meta Global AI Developer Hackathon, Oct 26–Nov 4.

    “For some background, from the Muse Spark Open model card on Hugging Face:” via Interconnects

    Reported byInterconnects, Ben's Bites

    How 1 source put it
    huggingface.co First shared 2026-10-17
  6. Shared by 5

    Claude reveals ECDSA curve origins

    Claude identified NIST ECDSA curve seeds as outputs from a C library's lrand48 generator seeded with Unix timestamps. Includes the specific generator seeds for base points, the recovery algorithm using LLL basis reduction, and code to reproduce the findings. cited

    Reported byTLDR AI, AI Breakfast, Latent Space / AINews, Forward Future

    Discussed onHacker News 1,065 pts · 391 comments

    Where it was shared
    anthropic.com paper · 8 min read First shared 2026-10-16
  7. Shared by 2

    Steam Fans Are Begging for a No-AI Filter as Steam Next Fest Is Filled With Slop

    “This morning (via Hacker News) I found Steam Fans Are Begging for a No-AI Filter as Steam Next Fest Is Filled With Slop. I wouldn't personally dream of submitting Battle Drones to an event like this - it's still way too low effort to warrant consideration as a legitimate indie gaming project.” via Simon Willison

    Reported bySimon Willison

    Discussed onHacker News 183 pts · 231 comments

    ign.com First shared 2026-10-20
  8. Shared by 2

    deepseek-ai/DeepInfer: DeepInfer: High-performance, large-scale inference framework for DeepSeek models

    DeepSeek releases DeepInfer, an open-source inference framework supporting DeepSeek-V4.1-Flash on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs. Includes configuration for DSpark verification, SWA bounded replay, and batch-invariant inference, plus build instructions requiring Python 3.12 and CUDA 13. cited

    Reported byInterconnects

    Discussed onHacker News 2,340 pts · 891 comments

    github.com repo · 2 min read First shared 2026-10-20
  9. Shared by 5

    DeepSeek-V4.1-Flash-1017 (Hugging Face Repo)

    “The brain first. This is a minor update to DeepSeek's cheaper "Flash" tier, and the weights are free to download under an MIT license, so anyone can run it or build a paid service on it. On DeepSWE v1.1, a coding-agent test, DeepSeek reports 77.0, versus 74.0 for Claude Opus 5 and 73.0 for GPT-5.6 Sol.” via The Neuron

    Reported byThe Neuron, TLDR AI, AI Breakfast, Interconnects

    Discussed onHacker News 959 pts · 724 comments

    AboutDeepSeek-V4.1-Flash-1017, DeepSeek-V4.1-Flash

    How 1 source put it
    • The Neuron2026-10-19issue ↗ The brain first. This is a minor update to DeepSeek's cheaper "Flash" tier, and the weights are free to download under an MIT license, so anyone can run it or build a paid service on it. On DeepSWE v1.1, a coding-agent test, DeepSeek reports 77.0, versus 74.0 for Claude Opus 5 and 73.0 for GPT-5.6 Sol.
    • TLDR AI2026-10-19issue ↗
    • AI Breakfast2026-10-19issue ↗
    • Interconnects2026-10-20issue ↗
    huggingface.co First shared 2026-10-17
  10. Shared by 2

    Nscale and Anthropic are reportedly in talks for yet another $45B compute deal

    Bloomberg reports Anthropic and Nscale are in talks for a second $45 billion compute deal. Includes details that both companies are currently pursuing IPOs while investors worry about spending too much money. cited

    Reported byThe Neuron, TLDR AI

    Where it was shared
    techcrunch.com news report · 1 min read First shared 2026-10-20
  11. Shared by 2

    One hacker’s AI agents break into 100+ companies

    NorForge found over 20 exposed Cairn panels used by Chinese and English-speaking actors for cyberattacks. Includes details on a Chinese cluster running 1100 projects and an English cluster targeting financial theft. cited

    Reported byThe Rundown AI, The Neuron

    Where it was shared
    norforge.com news report First shared 2026-10-20
  12. Shared by 2

    Black Forest Labs released FLUX 3 Max, a sped-up FLUX 3 built with fal that it says matches Seedance 2.5 on quality at…

    Reported byThe Rundown AI, The Neuron

    How 1 source put it
    • The Rundown AI2026-10-20issue ↗ Black Forest Labs released FLUX 3 Max, a sped-up FLUX 3 built with fal that it says matches Seedance 2.5 on quality at 22x the speed, rendering 10s clips in 12 seconds.
    • The Neuron2026-10-20issue ↗
    fal.ai First shared 2026-10-20
  13. Shared by 1

    This site is tracking software updates via this GitHub repository - Linux distros need to update their ca-certificates…

    “Any software that bundles a list of trusted certificates needs to be updated. This site is tracking software updates via this GitHub repository - Linux distros need to update their ca-certificates packages.” via Simon Willison

    Reported bySimon Willison

    github.com First shared 2026-10-20
  14. Shared by 1

    In Python world uv 0.13.4 and pip 26.2.2 and certifi 2026.10.19 have been released with upgraded certificate lists.

    Reported bySimon Willison

    github.com First shared 2026-10-20
  15. Shared by 1

    In Python world uv 0.13.4 and pip 26.2.2 and certifi 2026.10.19 have been released with upgraded certificate lists.

    Reported bySimon Willison

    github.com First shared 2026-10-20
  16. Shared by 1

    In Python world uv 0.13.4 and pip 26.2.2 and certifi 2026.10.19 have been released with upgraded certificate lists.

    Reported bySimon Willison

    github.com First shared 2026-10-20
  17. Shared by 1

    There's a very informative thread about this on the official Python forum.

    Reported bySimon Willison

    discuss.python.org First shared 2026-10-20
  18. Shared by 3

    Disney agrees to hire the majority of Character AI’s workforce

    Disney agreed to hire at least 60 Character.AI employees for an estimated $60 million in annual compensation. Details how the vast majority of offers were accepted despite high costs and potential litigation-related liabilities. cited

    Reported byThe Neuron, Superhuman AI, Forward Future

    Where it was shared
    techcrunch.com news report · 1 min read First shared 2026-10-19
  19. Shared by 2

    Claiming this is pleasantly easy: navigate to Settings -> Billing and select the API organization that should benefit…

    “Open Claude and upload the job description you want to build the interview around.” via Superhuman AI

    Reported bySuperhuman AI, Simon Willison

    How 1 source put it
    claude.ai First shared 2026-09-28
  20. Shared by 1

    Orchestrate subagents at scale with dynamic workflows

    “This research task is really an experiment in running Claude Code for web in "Ultracode" mode, as described in Orchestrate subagents at scale with dynamic workflows in the Claude Code documentation.” via Simon Willison

    Reported bySimon Willison

    code.claude.com First shared 2026-10-16

The week to that day two or more sources

  1. Shared by 7

    Investing in the public benefit

    Anthropic commits 5% of shares to a new Public Benefit Fund, adding 2% for every 50% valuation increase. Details how the LLC avoids lobbying, funds labor impact studies, and combines with founder pledges to reach 20% total public good equity. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Latent Space / AINews, Forward Future

    Where it was shared
    anthropic.com announcement · 2 min read First shared 2026-10-19
  2. Shared by 4

    of Homeland Security formally banned DeepSeek from civilian agency systems, giving departments 60 days to request…

    DHS excludes DeepSeek products from federal civilian procurements and removes them from agency IT systems. The order also mandates removing third-party products and services developed using DeepSeek implementations from civilian agency information technology. cited

    Reported byLatent Space / AINews, The Rundown AI, The Neuron, Interconnects

    Where it was shared
    public-inspection.federalregister.gov news report · 1 min read First shared 2026-10-20
  3. Shared by 4

    Diffusion at Scale: Introducing Mercury 3

    Inception launches Mercury 3, a diffusion LLM that runs 4–8x faster at 1,000+ tokens per second than GPT-6 Luna and Claude Haiku 5.5. Details pricing at $0.20 per million input and $0.70 per million output, plus a 50% launch discount and 1M token context window. cited

    Reported byThe Rundown AI, The Neuron, TLDR AI

    Discussed onHacker News 257 pts · 92 comments

    Where it was shared
    inceptionlabs.ai announcement · 4 min read First shared 2026-10-19
  4. Shared by 5

    Claude reveals ECDSA curve origins

    Claude identified NIST ECDSA curve seeds as outputs from a C library's lrand48 generator seeded with Unix timestamps. Includes the specific generator seeds for base points, the recovery algorithm using LLL basis reduction, and code to reproduce the findings. cited

    Reported byTLDR AI, AI Breakfast, Latent Space / AINews, Forward Future

    Discussed onHacker News 1,065 pts · 391 comments

    Where it was shared
    anthropic.com paper · 8 min read First shared 2026-10-16
  5. Shared by 5

    DeepSeek-V4.1-Flash-1017 (Hugging Face Repo)

    “The brain first. This is a minor update to DeepSeek's cheaper "Flash" tier, and the weights are free to download under an MIT license, so anyone can run it or build a paid service on it. On DeepSWE v1.1, a coding-agent test, DeepSeek reports 77.0, versus 74.0 for Claude Opus 5 and 73.0 for GPT-5.6 Sol.” via The Neuron

    Reported byThe Neuron, TLDR AI, AI Breakfast, Interconnects

    Discussed onHacker News 959 pts · 724 comments

    AboutDeepSeek-V4.1-Flash-1017, DeepSeek-V4.1-Flash

    How 1 source put it
    • The Neuron2026-10-19issue ↗ The brain first. This is a minor update to DeepSeek's cheaper "Flash" tier, and the weights are free to download under an MIT license, so anyone can run it or build a paid service on it. On DeepSWE v1.1, a coding-agent test, DeepSeek reports 77.0, versus 74.0 for Claude Opus 5 and 73.0 for GPT-5.6 Sol.
    • TLDR AI2026-10-19issue ↗
    • AI Breakfast2026-10-19issue ↗
    • Interconnects2026-10-20issue ↗
    huggingface.co First shared 2026-10-17
  6. Shared by 3

    MiniMax M3.1 Flash: Redefining Agent Performance and Cost

    MiniMax launches M3.1 Flash, an agent model that reduces usage costs by 70% while scoring 76.2 on BankerToolBench. Details flat pricing of $0.15 input and $0.60 output per 1M tokens alongside 200+ TPS output speeds. cited

    Reported byThe Rundown AI, The Neuron, TLDR AI

    Where it was shared
    minimax.io announcement · 2 min read First shared 2026-10-20
  7. Shared by 7

    Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date

    Meta releases Muse Spark Open, an 816B-parameter multimodal model designed for coding and agentic work. Details how two Apple M5 Ultra Mac Studios run the model locally via MLX and lists hosted inference partners. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Interconnects, TheSequence

    Where it was shared
    research.meta.ai announcement · 3 min read First shared 2026-10-16
  8. Shared by 3

    Disney agrees to hire the majority of Character AI’s workforce

    Disney agreed to hire at least 60 Character.AI employees for an estimated $60 million in annual compensation. Details how the vast majority of offers were accepted despite high costs and potential litigation-related liabilities. cited

    Reported byThe Neuron, Superhuman AI, Forward Future

    Where it was shared
    techcrunch.com news report · 1 min read First shared 2026-10-19
  9. Shared by 5

    Anthropic update on independent evaluators

    Anthropic embedded six METR employees with hardware and access matching its internal risk teams. Details how Anthropic can require redactions for confidential IP while METR publishes findings without approval. cited

    Reported byThe Rundown AI, The Neuron, TLDR AI, Forward Future, Turing Post

    Where it was shared
    anthropic.com announcement · 4 min read First shared 2026-10-16
  10. Shared by 4

    How the Claude Code team does code review / claude.dev Blog

    Claude Code merged PRs doubled from June to September while reverts fell over 40 percent. Includes a risk tier table mapping low, medium, and high changes to zero, one, or two human approvals. cited

    Reported bySimon Willison, TLDR AI, Turing Post

    Discussed onHacker News 487 pts · 480 comments

    Where it was shared
    claude.dev guide · 13 min read First shared 2026-10-16