LLMs are not the only transformer-based models that you can explore in repos like HuggingFace. #ai #mlmodels #huggingface #modelecosystem
07.03.2026 08:12
👍 1
🔁 0
💬 0
📌 0
LLMs are not the only transformer-based models that you can explore in repos like HuggingFace. #ai #mlmodels #huggingface #modelecosystem
Published some notes on the situation at Qwen - they released the Qwen 3.5 family (an outstanding family of open weight models) but now their lead researcher and several others all appear to have resigned within the past 24 hours simonwillison.net/2026/Mar/4/q...
When D3 was released in February 2011, it transformed how we build data visualizations for the web and ushered in a new age of unbridled creativity in information design, data journalism, and beyond. Today, it remains the backbone of modern, interactive data viz.
Happy 15th birthday, D3 🎂
d3js.org
New chapter of my Agentic Engineering Patterns guide. This one is about having coding agents build custom interactive and animated explanations to help fight back against cognitive debt simonwillison.net/guides/agent...
Mind-blowing, quite cool but also slightly unsettling: "46,000-Year-Old Worms Wake Up" www.seti.org/news/life-af...
This article is a must-read if you run an engineering team - it certainly made me aware of things we need to do to manage the transition better into AI-assisted development: www.ivanturkovic.com/2026/02/25/a... #engineering #ai #softwaredev #developers
What Claude Code Actually Chooses:
In this study, they pointed Claude Code at real repos 2,430 times and watched what it chose. No tool names in any prompt. Open-ended questions only. Insightful read: amplifying.ai/research/cla... #claude #ai #codegen
Shared memory is really working in making hardware for AI more affordable for developers - check out this article on the DGX Spark by @sebastianraschka.com sebastianraschka.com/blog/2025/dg... #llms #dgxspark #machinelearning
Insightful post by @simonwillison.net that aligns with my own experience guiding my engineering team over the last year of AI code-gen adoption - "Good code still has a cost." #ai #codegeneration #agenticai simonwillison.net/guides/agent...
Enjoying a quiet week in Nice, France and sharing a couple of visual moments...
Interesting...I might consider using it for some of my desktop-AI app experiments: github.com/alibaba/zvec
In honor of International Day of Women and Girls in Science, we present our list of 11 Women to Watch in Science. Prepare to be inspired! #womeninscience
CL-bench, a benchmark that tests whether LMs can learn new knowledge from context: 500 complex contexts, 1,899 tasks, and 31,607 verification rubrics. The test: solving each task requires learning something from context that is not already in the model's pre-training. www.clbench.com #llms #ai
TL;DR Highlights
* LLMs were effective for boilerplate, parsing logic, and refactoring, but not as autonomous system designers
* Human engineers still had to own architecture, correctness, and debugging
* Progress came from tight human-in-the-loop iteration, not “prompt and pray”
Anthropic published a fascinating engineering write-up on building a C compiler with LLM assistance — a great, grounded look at where AI actually helps in complex systems work (and where it doesn’t).
👉 www.anthropic.com/engineering/...
#AI #SoftwareEngineering #LLM #Compilers #HumanInTheLoop
Useful tips for token counting: www.dsdev.in/counting-tok...
We just posted the replay of our December Headwraps & Hacks Hacker Hour event 🎥
Blessing and I share 4 real builds that support our businesses and clients every day. Watch to learn more about
@fibery.com, #Asana, @n8n.io, @notion.com and custom GPTs.
For those who have a paid ChatGPT account, note that the promotional period for the Codex app involves double your standard quota - might be worth it if you have a side-project shelved that you could test this out on: openai.com/index/introd... #codex #developers #startups #chatgpt
❤️🩹🙏🏾❤️🩹
Sounds like a plan to make Mars become an actual requirement by making earth unlivable. Insanity.
As someone that has an almost identical experience starting with copy-pasting code from ChatGPT to Cursor and now to agents like Claude Code, I found this article very enlightening from @mariozechner.at mariozechner.at/posts/2025-1... #coding #aicodegen #agenticai #claudecode
This is pretty neat if you're into designing your own games (something I've always wanted to do but have found platform support a bit time-consuming): www.adventuregamestudio.co.uk
NetBird aligns with a modern networking sweet spot: WireGuard + zero-trust + identity-based access, without legacy VPN pain. From their site: Clean, fast, and built for distributed teams. netbird.io
#ZeroTrust #WireGuard #Networking #DevOps
After years of using portable drives to keep files backed-up (notably photos) - I finally went down the NAS road and it's working well but there are a few lessons on HDD sector health and such. This article confirmed a suspicion I had: www.xda-developers.com/why-not-to-s...
This is weird but it also sucks you in a little and you can't help peeping at the posts lol www.moltbook.com #socials #ai #agents
Might be fun to do on a spare Android device: www.xda-developers.com/turned-phone...
As someone that actually remembers the original Clippy fiasco, this is quite amusing… #clippylives #aichatbots #aiassistants #producthunt
This is helpful to deterministically fan-out step execution in parallel within a workflow.
With over 2,500 total device recalls occurring annually, focusing on the initial design phase is the most critical step for patient safety. This is especially interesting given the rise of AI usage in development but design still proves to be a category for human in the loop! #SoftwareEngineering
