Intelligence.Log
2026-05-06
Extracted: 62 items. Sources: Bluesky, X, Blogs.
++ AI OVERVIEW ++
A quieter day in AI development, with the community seemingly catching its breath after recent flurries of releases. While no major new projects or papers dominated GitHub stars, the conversation on Bluesky took a decidedly personal turn, as AI researcher Nathan Lambert marked a trivial but human milestone by updating his profile picture, confirming he’s been mustache-free for over two years. The lack of technical debate or trending repos suggests the field is in a reflective period, with practitioners taking a moment for lighthearted self-expression before the next inevitable wave of model launches and benchmark wars.
grep TOPIC=
grep SOURCE=
sort --by=
BSKY
new pfp, final digital confirmation that I haven't had a mustache in like 2+ years
❤️ 3 Likes|
BSKY
I'm at the Claude w/ Code event in San Francisco, and I'll be live blogging the keynote here: simonwillison.net/2026/May/6/c...
❤️ 78 Likes|[LLM][Tooling]
BSKY
I was talking with Joseph Ruscio on the @heavybit.com podcast the other day when I realized that vibe coding and agentic engineering have started to blur a bit in some of my work - I published some extracts from the transcript simonwillison.net/2026/May/6/v...
❤️ 37 Likes|[Agent][Tooling]
BSKY
Question for #PolisciSky on bsky: How should the US Constitution be changed to make Congress more representative? The Voting Rights Act (and various other redistricting commissions) have attempted to work around the fundamental flaws of congressional districts. Potential fixes: 1/
❤️ 1 Likes|
BSKY
Today, I'm receiving heavy SMS phishing from someone claiming to be from T-Mobile. Beware!
❤️ 0 Likes|[Safety]
BSKY
Added a 1500 word mini history to my book on the path to on-policy distillation being a core post-training optimization technique. Is a fun rapidly growing area now! rlhfbook.com/c/12-synthet...
❤️ 16 Likes|[Fine-tuning]
BSKY
I usually avoid commenting too much on industry deals, but this one is fascinating. Certainly seems like a blow to the idea that Grok will remain a frontier model (while giving Anthropic a lot of compute).
❤️ 114 Likes|[Infra]
BSKY
BSKY
BSKY
UW CSE News covers our CHI award-winning paper 🥳 news.cs.washington.edu/2026/05/05/a...
❤️ 9 Likes|
X
My most amusing interaction was where the model (I think I was given some earlier version with a
[LLM]
“DeepSeek Summary: Karpathy recounts an amusing interaction with an early model version.”
X
One common issue with personalization in all LLMs is how distracting memory seems to be for the models.
[LLM][Fine-tuning]
“DeepSeek Summary: Identifies a key challenge in LLM personalization: memory distraction.”
X
- Drafted a blog post - Used an LLM to meticulously improve the argument over 4 hours.
[LLM][Tooling]
“DeepSeek Summary: Karpathy describes using an LLM to refine a blog post argument over 4 hours.”
X
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around
[Evaluation]
“DeepSeek Summary: Notes a growing gap in understanding AI capabilities, starting with a specific issue.”
X
I'm starting to get into a habit of reading everything (blogs, articles, book chapters,…)
[LLM]
“DeepSeek Summary: Karpathy shares his habit of extensive reading across various formats.”
X
This may be the best guidance I've seen anywhere on writing a really good commit history.
[Tooling]
“DeepSeek Summary: Simon Willison praises guidance on writing good commit history.”
X
TL;DR: More and more agents need a workspace: a computer where they can run code, install packages, and access files. Sandboxes provide this
[Agent][Infra][Deployment]
“DeepSeek Summary: Agents require a workspace like a computer to run code, install packages, and access files; sandboxes fulfill this need.”
X
Memory is just a form of context. Short term memory (messages in the conversation, large tool call results) are handled by the harness. Long
[Agent][LLM]
“DeepSeek Summary: Memory is conceptualized as a form of context, with short-term memory managed by the harness.”
X
The Second Pre-training Paradigm
[LLM][Multi-modal]
“DeepSeek Summary: Jim Fan discusses a new paradigm for pre-training in AI.”
X
I've been a bit quiet on X recently. The past year has been a transformational experience.
[Agent]
“DeepSeek Summary: Jim Fan acknowledges his absence and hints at significant changes.”
X
Folks seem to rediscover this every couple of years. As I've been saying for many years,
[LLM]
“DeepSeek Summary: Jeremy notes that people repeatedly rediscover a concept he has been advocating for years.”
X
Here's a complete unedited video of asking Grok for its views on the Israel/Palestine situation. It first searches twitter for what Elon thinks.
[Safety][LLM]
“DeepSeek Summary: Jeremy shares a video showing Grok's behavior in seeking Elon Musk's opinion on a sensitive topic.”
X
I replicated this result, that Grok focuses nearly entirely on finding out what Elon thinks in
[Safety][LLM]
“DeepSeek Summary: Jeremy confirms his replication of a finding about Grok's reliance on Elon Musk's views.”
X
reading "AI News" (previously Smol Talk) is probably the highest-leverage 45 mins
[LLM]
“DeepSeek Summary: Soumith recommends a newsletter as high-leverage reading.”
X
Sometimes we forget that NVIDIA wins because it's a software company.
[Infra]
“DeepSeek Summary: He emphasizes NVIDIA's software strength.”
X
Open LLMs need to get organized and co-ordinated about sharing human feedback.
[LLM][Fine-tuning]
“DeepSeek Summary: Calls for coordination in open LLM community on feedback sharing.”
X
MacStudio you ask? Apple Engineering's **actual** time spent on PyTorch support
[Infra][Deployment]
“DeepSeek Summary: Comments on Apple's engineering effort for PyTorch on MacStudio.”
X
Current AI is a librarian of existing knowledge. Science requires an explorer of the unknown.
[Evaluation][LLM]
“DeepSeek Summary: Chollet contrasts AI's role as a retriever of known information with the need for exploration in science.”
X
Folks who work in AI or software engineering feel like the world is changing exponential fast.
[Deployment]
“DeepSeek Summary: Chollet notes the perception of rapid change among AI and software professionals.”
X
Re-reading an article I wrote in 2017, and I'm finding I could have written it yesterday
[Evaluation]
“DeepSeek Summary: Chollet observes that his 2017 article remains relevant today.”
X
Dario is wrong. He knows absolutely nothing about the effects of technological revolutions on the labor market.
[Safety]
“DeepSeek Summary: Yann LeCun criticizes Dario Amodei's views on AI and labor market impact, asserting Dario lacks understanding of historical technological revolutions.”
X
It seems to me that before 'urgently figuring out how to control AI systems much smarter than us' we need
[Safety]
“DeepSeek Summary: LeCun questions the urgency of controlling superintelligent AI, implying current AI is far from that capability.”
X
The emergence of superintelligence is not going to be an event. We don't have anything close to a
[Safety]
“DeepSeek Summary: LeCun argues superintelligence will be gradual, not sudden, and current AI lacks the foundations for it.”
X
Very excited to share @theworldlabs 's latest research work RTFM!! It's a real-time, ...
[Multi-modal]
“DeepSeek Summary: Fei-Fei Li shares excitement about World Labs' RTFM research, a real-time spatial intelligence model.”
X
Some personal news: I recently joined Cursor. Cursor is a small, ambitious team, and they've created
[Tooling]
“DeepSeek Summary: Sasha Rush announced joining Cursor, a small ambitious team.”
X
Been reflecting a bit on the Harvard news. This paper from 2017 was ... Didn't realize at the time how lucky for us Americans to work with incredible people from around the world.
[Safety]
“DeepSeek Summary: Reflecting on Harvard news and gratitude for international collaboration.”
X
I have been compiling LLM/VLM training logbooks/chronicles. This is the one of the best sources to ...
[LLM]
“DeepSeek Summary: Stas Bekman has been compiling LLM/VLM training logbooks, which are valuable resources for understanding training processes.”
X
Thanks to an awesome contribution from @omarnomad The Machine Learning Engineering Open book now can ...
[Tooling][LLM]
“DeepSeek Summary: Acknowledges contribution to the Machine Learning Engineering Open Book, enhancing its content.”
X
If you were holding off to try @MSFTDeepSpeed ZeRO++ it looks like deepspeed@master should ...
[Infra]
“DeepSeek Summary: Announces that DeepSpeed ZeRO++ is now usable, encouraging adoption.”
X
Modern art. Artist: PyTorch memory profiler Model: Llama-8B The piece on the left is the ...
[Tooling][LLM]
“DeepSeek Summary: Uses humor to illustrate PyTorch memory profiling results for Llama-8B model.”
X
Release notes: Release Diffusers 0.34.0: New Image and Video Models, Better torch.
[Multi-modal][Deployment][Tooling]
“DeepSeek Summary: Announcement of Diffusers 0.34.0 release with new image and video models and improved torch support.”
X
Had a nice time chatting about the state of diffusion models and some text-to-image data shenanigans at
[Multi-modal][Fine-tuning]
“DeepSeek Summary: Engaged in a discussion about diffusion models and text-to-image data challenges.”
X
X
How to use Deep Research with the Gemini API. www.philschmid.de.
[Agent][Infra]
“DeepSeek Summary: Philipp Schmid shares a guide on using Deep Research with the Gemini API.”
X
Guide: ReAct agent from scratch with Gemini 2.5 and LangGraph | Gemini API | Google AI for Developers. ai.google.dev.
[Agent][LLM][Tooling]
“DeepSeek Summary: Philipp Schmid publishes a guide on building a ReAct agent from scratch using Gemini 2.5 and LangGraph.”
X
On the plus side with Opus 4.7, if it does decide to think it produces BY FAR the best
[LLM]
“DeepSeek Summary: Ethan Mollick praises Opus 4.7 for producing the best results when it decides to think.”
X
One thing thing about AI, for better and worse, is that "everything around me is somebody's life
[Safety]
“DeepSeek Summary: Reflects on the profound impact of AI on people's lives.”
X
After reading it, this does seem like a big deal. Industry experts outlined important, real-world, hard tasks for AI to do.
[Evaluation]
“DeepSeek Summary: Emphasizes the significance of industry-defined hard tasks for AI.”
X
EMILY M. BENDER: Yeah. And so passive, like, oops, the moon, the moon went further away. It's like no, actually, you made some decisions.
[Safety]
“DeepSeek Summary: Critique of passive language in AI discourse, emphasizing human agency in decisions.”
X
Image is of the 1990s Microsoft writing assistant character Clippy with its eyebrows raised positioned in.
[LLM]
“DeepSeek Summary: Uses Clippy meme to critique AI assistants.”
X
Facebook (sorry: Meta) AI: Check out our "AI" that lets you access all of humanity's knowledge.
[Deployment]
“DeepSeek Summary: Sarcastic commentary on Meta's AI claims.”
X
New preprint! Everyone loves causal interp. It's coherently defined! It makes testable predictions
[Safety][Evaluation]
“DeepSeek Summary: Naomi announces a new preprint on causal interpretation, emphasizing its coherent definition and testable predictions.”
X
I work on understanding and improving training for NLP models, with a focus on studying how structures and mechanistic behaviors emerge over the
[LLM][Evaluation]
“DeepSeek Summary: Naomi describes her research focus on training dynamics and emergence of mechanistic behaviors in NLP models.”
X
Ok, I wrote this up (link below)
[Evaluation]
“DeepSeek Summary: Naomi references a write-up she authored, likely a blog post or paper.”
X
For the first time in almost a decade, I'm teaching a class on learning and control.
[Evaluation]
“DeepSeek Summary: Ben Recht announces teaching a class on learning and control after nearly a decade.”
X
This stupid website is so cooked.
“DeepSeek Summary: Ben Recht expresses frustration with the X/Twitter platform.”
X
Revisiting Sutton's Bitter Lesson in the wake of GPT-5.
[LLM]
“DeepSeek Summary: Ben Recht discusses Sutton's Bitter Lesson in context of GPT-5.”
BLOG
<p>I'm at Anthropic's Code w/ Claude event today. Here's my live blog of the morning keynote sessions.</p><p><em>You are only seeing the long-form articles from my blog. Subscribe to <a href="https://simonwillison.net/atom/everything/">/atom/everything/</a> to get all of my posts, or take a look at...
“Anthropic's Code w/ Claude event showcases new capabilities for AI-assisted coding, including improved code generation, debugging, and collaborative features. The live blog format provides real-time insights into keynote sessions, highlighting practical applications and future directions for Claude in software development.”
BLOG
<p>I recently talked with Joseph Ruscio about AI coding tools for Heavybit's High Leverage podcast: <a href="https://www.heavybit.com/library/podcasts/high-leverage/ep-9-the-ai-coding-paradigm-shift-with-simon-willison">Ep. #9, The AI Coding Paradigm Shift with Simon Willison</a>. Here are some of...
“The post discusses the convergence of 'vibe coding' (using AI to generate code without fully understanding it) and 'agentic engineering' (autonomous AI agents that build software), warning that as these approaches advance, developers risk losing control over code quality and security. It emphasizes the need for human oversight and testing, especially as AI-generated code becomes more complex and harder to audit.”
-- END OF LOG --
[STATS] 62 items · Filter applied
Powered by Horizon + DeepSeek