- 3The Great Copper Crunch: Why the Next Bull Run Is Just Getting Started (prinsights.substack.com)
- 1P/ – A simple and clear programming language
- 3Ask HN: How much Adderall do you use?
ask - 2Understanding an AI agent's memory through Harry Potter (squidler.io)
- 5My home internet died for half a day. So did every agent I had (bostrat.ai)
- 2Fedora CoreOS to Enable Systemd-Oomd and ZRAM Swap by Default (phoronix.com)
- 2Ask HN: When should you open-source your solution and why?
ask - 3TDD inside the agent loop – theater or actual value? (martinfowler.com)
- 3Show HN: HyperSAE – Hyperbolic Sparse Autoencoders for LLM Interpretability (github.com)
- 2LTX 2.5 a stronger foundation for what's being built (ltx.io)
- 2Ask HN: How do you keep 54 LLM workflows on the right models?
ask - 5OpenAI's head of ethics leaves startup less than one year after joining (ft.com)
- 1A Preliminary Study on Simultaneous Coscheduling for Discrete GPU vs. Fused GPU (arxiv.org)
- 1Show HN: SmokeOps – describe a flow, get checks you rerun without AI (github.com)
- 2Imprint – Fine-tune MoE LLMs bigger than your RAM (github.com)
- 1Sam's News – Free AI News Aggregator (samnews.news)
- 95How we used to get jobs: A newspaper classifieds story (ironicsans.ghost.io)
- 3Nvidia building 1T-parameter Nemotron 4 to rival open AI models (reuters.com)
- 2Cynthion: Open-source USB test instrument (greatscottgadgets.com)
- 5Elon Musk ran a decoy super PAC to compare Trump with RGB (google.com)
- 1The worst part of using multiple LLM providers is not the API differences (github.com)
- 2A bug becoming a feature (1972 edition) (unsung.aresluna.org)
- 7Ethical Cold Outreach (blog.val.town)
- 2Report from the Security Mines (blog.val.town)
- 3Docfooding: Eating Our Own Documentation (blog.val.town)
- 1Hack's Law (en.wikipedia.org)
- 1Would Appreciate Any Feedback (deskfellow.app)
- 14ChatGPT Desktop App for Linux (twitter.com)
- 4Why Are Rivers So Mathematical? (quantamagazine.org)
- 2(2,1)-C1P is NP-complete (zenodo.org)