# 2026 W04 - 01-25 - Agents and long-running tasks

First time here ? Check the introduction [here]({% post_url hn_digest/2026-01-01-Init %})


# What you Missed this Week 

Not that much this week.

AI adoption also reached researchers, where some references were hallucinated.


# Articles

## Model Training 

- 636 pt  [[Link](<https://qwen.ai/blog?id=qwen3tts-0115>) / [HN](<https://news.ycombinator.com/item?id=46719229>)] Qwen3-TTS family is now open sourced: Voice design, clone, and generation
- 497 pt  [[Link](<https://huggingface.co/sweepai/sweep-next-edit-1.5B>) / [HN](<https://news.ycombinator.com/item?id=46713106>)] Show HN: Sweep, Open-weights 1.5B model for next-edit autocomplete 
- 319 pt  [[Link](<https://huggingface.co/zai-org/GLM-4.7-Flash>) / [HN](<https://news.ycombinator.com/item?id=46679872>)] GLM-4.7-Flash 
- 84 pt  [[Link](<https://huggingface.co/collections/Linum-AI/linum-v2-2b-text-to-video>) / [HN](<https://news.ycombinator.com/item?id=46721488>)] Show HN: Text-to-video model from scratch (2 brothers, 2 years, 2B params) 

## Coding Assistant 

- 587 pt  [[Link](<https://hugodaniel.com/posts/claude-code-banned-me/>) / [HN](<https://news.ycombinator.com/item?id=46723384>)] I was banned from Claude for scaffolding a Claude.md file? 
- 581 pt  [[Link](<https://www.media.mit.edu/publications/your-brain-on-chatgpt/>) / [HN](<https://news.ycombinator.com/item?id=46712678>)] Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant 
- 541 pt  [[Link](<https://www.anthropic.com/news/claude-new-constitution>) / [HN](<https://news.ycombinator.com/item?id=46707572>)] Claude's new constitution 
- 512 pt  [[Link](<https://github.com/anthropics/original_performance_takehome>) / [HN](<https://news.ycombinator.com/item?id=46700594>)] Anthropic's original take home assignment open sourced 
- 180 pt  [[Link](<https://blog.emilburzo.com/2026/01/running-claude-code-dangerously-safely/>) / [HN](<https://news.ycombinator.com/item?id=46690907>)] Running Claude Code dangerously (safely) 
- 168 pt  [[Link](<https://github.com/mastra-ai/mastra>) / [HN](<https://news.ycombinator.com/item?id=46693959>)] Show HN: Mastra 1.0, open-source JavaScript agent framework from the Gatsby devs 
- 142 pt  [[Link](<https://simonwillison.net/2026/Jan/19/scaling-long-running-autonomous-coding/>) / [HN](<https://news.ycombinator.com/item?id=46686418>)] Scaling long-running autonomous coding 
- 138 pt  [[Link](<https://rijnard.com/blog/the-code-only-agent>) / [HN](<https://news.ycombinator.com/item?id=46674416>)] The Code-Only Agent 
- 122 pt  [[Link](<https://blog.silennai.com/claude-code>) / [HN](<https://news.ycombinator.com/item?id=46676554>)] I was a top 0.01% Cursor user, then switched to Claude Code 2.0 

## Personal Blogs

- 276 pt  [[Link](<https://calquio.com/finance/compound-interest>) / [HN](<https://news.ycombinator.com/item?id=46673809>)] Show HN: I quit coding years ago. AI brought me back 
- 217 pt  [[Link](<https://cyberlaw.stanford.edu/publications/how-ai-destroys-institutions/>) / [HN](<https://news.ycombinator.com/item?id=46705606>)] How AI Destroys Institutions 
- 195 pt  [[Link](<https://sean.heelan.io/2026/01/18/on-the-coming-industrialisation-of-exploit-generation-with-llms/>) / [HN](<https://news.ycombinator.com/item?id=46676081>)] The coming industrialisation of exploit generation with LLMs 
- 117 pt  [[Link](<https://philipotoole.com/why-talking-to-llms-has-improved-my-thinking/>) / [HN](<https://news.ycombinator.com/item?id=46728197>)] Talking to LLMs has improved my thinking 
- 91 pt  [[Link](<https://kconner.com/2024/08/02/ai-is-a-horse.html>) / [HN](<https://news.ycombinator.com/item?id=46686402>)] AI Is a Horse (2024) 
- 80 pt  [[Link](<https://www.404media.co/comic-con-bans-ai-art-after-artist-pushback/>) / [HN](<https://news.ycombinator.com/item?id=46705952>)] Comic-Con Bans AI Art After Artist Pushback 

## Politics

- 285 pt  [[Link](<https://www.valueaddedresource.net/ebay-bans-ai-agents-updates-arbitration-user-agreement-feb-2026/>) / [HN](<https://news.ycombinator.com/item?id=46711574>)] eBay explicitly bans AI "buy for me" agents in user agreement update 
- 250 pt  [[Link](<https://dbushell.com/2026/01/22/proton-spam/>) / [HN](<https://news.ycombinator.com/item?id=46729368>)] Proton Spam and the AI Consent Problem 
- 204 pt  [[Link](<https://github.com/ghostty-org/ghostty/blob/main/AI_POLICY.md>) / [HN](<https://news.ycombinator.com/item?id=46730504>)] Ghostty's AI Policy 

## Unclassified

- 869 pt  [[Link](<https://gptzero.me/news/neurips/>) / [HN](<https://news.ycombinator.com/item?id=46720395>)] GPTZero finds 100 new hallucinations in NeurIPS 2025 accepted papers 
- 194 pt  [[Link](<https://ploum.net/2026-01-19-exam-with-chatbots.html>) / [HN](<https://news.ycombinator.com/item?id=46688954>)] Giving university exams in the age of chatbots 
- 188 pt  [[Link](<https://en.wikipedia.org/wiki/Wikipedia:WikiProject_AI_Cleanup>) / [HN](<https://news.ycombinator.com/item?id=46677106>)] Wikipedia: WikiProject AI Cleanup 
- 177 pt  [[Link](<https://openai.com/index/scaling-postgresql/>) / [HN](<https://news.ycombinator.com/item?id=46725300>)] Scaling PostgreSQL to power 800M ChatGPT users 
- 174 pt  [[Link](<https://github.com/jordanhubbard/nanolang>) / [HN](<https://news.ycombinator.com/item?id=46684958>)] Nanolang: A tiny experimental language designed to be targeted by coding LLMs 
- 171 pt  [[Link](<https://so-long-sucker.vercel.app/>) / [HN](<https://news.ycombinator.com/item?id=46698370>)] Which AI Lies Best? A game theory classic designed by John Nash 
- 158 pt  [[Link](<https://www.nibzard.com/agentic-handbook>) / [HN](<https://news.ycombinator.com/item?id=46701969>)] The Agentic AI Handbook: Production-Ready Patterns 
- 101 pt  [[Link](<https://www.anthropic.com/research/assistant-axis>) / [HN](<https://news.ycombinator.com/item?id=46684708>)] The assistant axis: situating and stabilizing the character of LLMs 
- 88 pt  [[Link](<https://www.theregister.com/2026/01/19/copper_chief_cops_it_after/>) / [HN](<https://news.ycombinator.com/item?id=46679657>)] West Midlands police chief quits over AI hallucination 
- 84 pt  [[Link](<https://teendifferent.substack.com/p/apply_chat_template-is-the-safety>) / [HN](<https://news.ycombinator.com/item?id=46675271>)] Bypassing Gemma and Qwen safety with raw strings 
- 65 pt  [[Link](<https://github.com/browseros-ai/BrowserOS>) / [HN](<https://news.ycombinator.com/item?id=46721474>)] Show HN: BrowserOS – "Claude Cowork" in the browser 
- 54 pt  [[Link](<https://walters.app/blog/composing-apis-clis>) / [HN](<https://news.ycombinator.com/item?id=46722074>)] Composing APIs and CLIs in the LLM era