Intelligence.Log
2026-04-19
Extracted: 59 items. Sources: Bluesky, X.
++ AI OVERVIEW ++
Today's discussions spotlight the evolving transparency in AI, with Simon Willison's analysis of the published system prompt diff between Claude Opus 4.6 and 4.7 offering a rare, technical look into model iteration. The research community is buzzing ahead of ICLR, with calls to share conference moments on Bluesky to foster inclusion. In trending repos, Sakana AI's "Digital Ecosystems" project showcases interactive multi-agent neural cellular automata, highlighting continued innovation in emergent AI behavior. Meanwhile, a broader conversation on digital privacy emerges alongside a lighter, shared appreciation for foundational textbooks and biographies.
grep TOPIC=
grep SOURCE=
sort --by=
BSKY
Since Anthropic publish their system prompts we can generate a diff between Claude Opus 4.6 and 4.7 - here are my notes on what's changed simonwillison.net/2026/Apr/18/...
❤️ 75 Likes|[LLM][Evaluation]
BSKY
Reminder @iclr-conf.bsky.social #ICLR conference-goers next week. 🙏 PLEASE POST PICS ON BLUESKY! 🙏 We want to feel like we're there with you. Just like we used to on X. We can't lose to LinkedIn in this! Bring that conference atmosphere to Bluesky! That includes you too, organizers :kthxbye:
❤️ 10 Likes|
BSKY
BSKY
Ok, second book of my recent foray into a newfound love of biographies: Friends, Lovers, and the Big Terrible Thing, by Matthew Perry. I bought this one when it first came out but did not get around to reading now. I learned a lot reading this book, but.. it's a tough read. #booksky 1/N
❤️ 4 Likes|
BSKY
This is absolutely terrible. Google, how far you have fallen! What's next? Fingerprints? DNA?
❤️ 1 Likes|[Safety]
BSKY
Digital Ecosystems: Interactive Multi-Agent Neural Cellular Automata pub.sakana.ai/digital-ecos...
❤️ 15 Likes|[Agent][Multi-modal]
BSKY
Can an AI run a retail store with a 3-year lease? andonlabs.com/blog/andon-m...
❤️ 7 Likes|[Agent][Deployment]
BSKY
A living document trying to make sense of all the research on adoption and usage of AI aleximas.substack.com/p/who-uses-a...
❤️ 12 Likes|[Evaluation]
BSKY
The continuing gap between the capabilities of Gemini Pro 3.1 (very good model) and the capabilities of the Gemini app/website is odd. The model can do what Claude/GPT can do, but there is a minimal harness for tools (file creation, research etc), no auditable CoT/actions, manual canvas, etc. 1/
❤️ 64 Likes|[LLM][Tooling][Deployment]
BSKY
I am forever saying that if refusal isn't a live option in any decision making process about "AI", then no ethical practice is possible. You've got to be able to stop if the thing is unacceptable.
❤️ 282 Likes|[Safety]
BSKY
X
The hottest new programming language is English
[LLM][Tooling]
“DeepSeek Summary: English is becoming the primary interface for programming and interacting with AI systems, suggesting a shift toward natural language as a programming paradigm.”
X
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is...
[LLM][Evaluation]
“DeepSeek Summary: Public understanding of AI capabilities is lagging, partly because many formed opinions based on outdated or limited (free-tier) experiences with models like ChatGPT.”
X
LLMs are emerging as a new kind of intelligence, simultaneously a lot smarter than I expected and a lot dumber than I expected. In any case they...
[LLM][Evaluation]
“DeepSeek Summary: Large Language Models represent a novel form of intelligence with surprising capabilities and surprising limitations, defying simple categorization.”
X
It's interesting how "better at code" has become the defining goal of almost every AI lab over the past year.
[LLM][Evaluation]
“DeepSeek Summary: Observes that AI research has shifted focus toward improving coding capabilities as a primary objective across major labs.”
X
The last year six months in LLMs, illustrated by pelicans on bicycles. I've published video, slides and a detailed annotated transcript from my talk at this week's conference.
[LLM][Evaluation]
“DeepSeek Summary: Summarizes recent LLM progress with a creative metaphor and shares comprehensive materials from a conference presentation.”
X
Screenshot from a video game where a team of raccoons go on a heist.
[Multi-modal]
“DeepSeek Summary: Shares an AI-generated image depicting raccoons in a heist scenario within a futuristic video game setting.”
X
This means that operations you would do on code in the software world, you now do on traces in the agent world. Debugging, testing, profiling
[Agent][Evaluation][Tooling]
“DeepSeek Summary: Harrison Chase draws parallels between software engineering practices and agent development, emphasizing that debugging, testing, and profiling now apply to agent traces rather than just code.”
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: Harrison Chase discusses the growing need for agent workspaces—sandboxed environments where agents can execute code, manage dependencies, and interact with files—as essential infrastructure for advanced AI systems.”
X
When you ship traditional software to production, you have a good sense of what to expect. Users click buttons, fill out forms,
[Agent][Deployment][Evaluation]
“DeepSeek Summary: Harrison Chase contrasts the predictable nature of traditional software deployment with the uncertainties of AI agent deployment, implying that agent behavior is less deterministic.”
X
Resource constraints are a beautiful thing. Survival instinct in a cut-throat AI competitive land
[Infra][Deployment]
“DeepSeek Summary: Argues that resource limitations foster innovation and competitive survival in the AI field.”
X
The first time I met Jensen was also the first time I met @elonmusk. I was interning at OpenAI that day and
“DeepSeek Summary: Shares a personal anecdote about meeting key AI industry figures (Jensen Huang and Elon Musk) during an OpenAI internship.”
X
I've been a bit quiet on X recently. The past year has been a transformational experience.
“DeepSeek Summary: Acknowledges a period of reduced public posting, attributing it to a significant personal or professional transformation.”
X
It gives me a lot of comfort knowing that we are the last generation without advanced robots everywhere.
[Agent][Deployment]
“DeepSeek Summary: Expresses a unique sense of comfort in living during the final pre-ubiquitous-robotics era.”
X
Everyone's freaking out about vibe coding. In the holiday spirit, allow me to share my anxiety on the wild
[Tooling][Safety]
“DeepSeek Summary: Comments on the hype around 'vibe coding' while personally expressing anxiety about the unpredictable trajectory of AI development.”
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.
[Agent][Evaluation]
“DeepSeek Summary: Demonstrates that Grok prioritizes finding Elon Musk's opinion before forming its own view on a complex geopolitical issue.”
X
I replicated this result, that Grok focuses nearly entirely on finding out what Elon thinks in
[Agent][Evaluation]
“DeepSeek Summary: Confirms through replication that Grok's primary focus is determining Elon Musk's perspective rather than providing independent analysis.”
X
If you'd like to try this yourself, here's how to get started:
[Tooling][Deployment]
“DeepSeek Summary: Offers practical guidance for others to replicate or experiment with a technical process.”
X
reading 'AI News' (previously Smol Talk) is probably the highest-leverage 45 mins
[LLM]
“DeepSeek Summary: Soumith Chintala recommends 'AI News' (formerly Smol Talk) as a highly valuable 45-minute activity for staying informed about AI developments.”
X
Open LLMs need to get organized and co-ordinated about sharing human feedback.
[LLM][Evaluation]
“DeepSeek Summary: Soumith Chintala calls for better organization and coordination among open LLM projects regarding the sharing of human feedback data.”
X
MacStudio you ask? Apple Engineering's actual time spent on PyTorch support
[Infra][Deployment]
“DeepSeek Summary: Soumith Chintala comments on Apple's engineering investment in PyTorch support, specifically mentioning the MacStudio.”
X
Sometimes we forget that NVIDIA wins because it's a software company.
[Infra][Tooling]
“DeepSeek Summary: Soumith Chintala emphasizes that NVIDIA's success is fundamentally driven by its software capabilities, not just hardware.”
X
Back in 2023 everybody was telling me 'no one uses Google search anymore, it's over' From 2023 to...
[Infra]
“DeepSeek Summary: Chollet references a common 2023 sentiment about Google search being obsolete, implying a contrast with current reality or his perspective on search technology trends.”
X
Very excited to share @theworldlabs 's latest research work RTFM!! It's a real-time,
[Agent][Multi-modal]
“DeepSeek Summary: Fei-Fei Li announces RTFM, a real-time research work from her company World Labs, indicating active development in AI research.”
X
congrats to OpenAI on winning the Turing Test
[LLM][Evaluation]
“DeepSeek Summary: Max Woolf congratulates OpenAI for achieving a milestone in AI development, referencing the Turing Test as a benchmark for machine intelligence.”
X
Stand-up comedy's gain is game criticism's loss: comedian @PhilNWang speaks with clarity and insight about the five video games he would like to
“DeepSeek Summary: Phil Wang discusses video games he'd like to see, blending his comedic perspective with game criticism.”
X
Phil Wang // Insta: @wangpix (@PhilNWang). 324 likes 24 replies. I got to cover for the excellent @HadleyFreeman in the Guardian today so
“DeepSeek Summary: Phil Wang mentions covering for Hadley Freeman in The Guardian, indicating his writing work.”
X
Wager established. Jonathan Frankle (@jefrankle) stepped up to my Transformer long bet.
[LLM]
“DeepSeek Summary: Sasha Rush mentions establishing a wager or bet with Jonathan Frankle related to Transformers, indicating engagement in technical debates or predictions.”
X
Some personal news: I recently joined Cursor. Cursor is a small, ambitious team, and they've created
[Tooling]
“DeepSeek Summary: Sasha Rush announces joining Cursor, highlighting it as a small, ambitious team working on innovative projects.”
X
Congrats to the @cursor_ai team on the launch of Composer 2! We are proud to see Kimi-k2.5 provide the foundation.
[Tooling][Infra]
“DeepSeek Summary: Sasha Rush congratulates the Cursor team on launching Composer 2, noting Kimi-k2.5 as its foundation, suggesting involvement in AI tool development.”
X
I have been compiling LLM/VLM training logbooks/chronicles. This is the one of the best sources to...
[LLM][Evaluation]
“DeepSeek Summary: Stas Bekman is compiling training logbooks/chronicles for LLM/VLM models, suggesting he's documenting training processes and methodologies.”
X
Thanks to an awesome contribution from @omarnomad The Machine Learning Engineering Open book now can...
[Tooling][Deployment]
“DeepSeek Summary: Acknowledgement of a contribution to the Machine Learning Engineering Open book project, indicating collaborative work on educational resources.”
X
If you were holding off to try @MSFTDeepSpeed ZeRO++ it looks like deepspeed@master should...
[Infra][Tooling]
“DeepSeek Summary: Discussion about Microsoft's DeepSpeed ZeRO++ optimization framework, suggesting technical evaluation of distributed training tools.”
X
Modern art. Artist: PyTorch memory profiler Model: Llama-8B The piece on the left is the...
[LLM][Tooling][Evaluation]
“DeepSeek Summary: Creative visualization of PyTorch memory profiling results for Llama-8B model, blending technical analysis with artistic presentation.”
X
Giving a talk here is by far the most fulfilling experience of my life!
[Deployment]
“DeepSeek Summary: Sayak Paul expresses that giving a talk was his most fulfilling life experience, indicating significant personal and professional satisfaction from public speaking engagements.”
X
Repo:
[Tooling]
“DeepSeek Summary: Sayak Paul shared a repository link, suggesting he's publishing or referencing technical work, though the specific content isn't detailed in the snippet.”
X
I read three technical reports from Moonshot AI's Kimi K2.5 paper, Cursor's Composer 2 report and blog post, and Chroma's Context-1 write-up
[LLM][Tooling]
“DeepSeek Summary: Philipp Schmid is actively reading and engaging with multiple technical reports from leading AI companies, indicating he stays current with industry developments.”
X
Content not fully available in search results - appears to be a post with engagement metrics.
“DeepSeek Summary: A post by Philipp Schmid that received engagement (52 likes), indicating active participation on the platform.”
X
Content not fully available in search results - appears to reference a blog post with engagement metrics.
“DeepSeek Summary: A post by Philipp Schmid referencing a blog, showing he shares his written content on social media.”
X
Content not fully available in search results - appears to be a profile-related post with engagement.
“DeepSeek Summary: A post showing Philipp Schmid's profile information and affiliate status, with significant engagement (83 likes, 3 replies).”
X
As stories about AI increasingly become stories of either catastrophe or salvation,
[Safety][Deployment]
“DeepSeek Summary: Ethan Mollick observes that AI narratives are polarizing into extremes of doom or utopia, missing nuanced perspectives.”
X
So much work is going into faking continual learning and memory for AIs,
[Agent][LLM]
“DeepSeek Summary: Mollick points out significant effort is being invested in simulating ongoing learning and memory capabilities in AI systems, rather than achieving genuine functionality.”
X
If it helps, I teach at a business school & many of my smartest students are hired by funds because they can reliably turn their only-human
[Deployment]
“DeepSeek Summary: Mollick shares an observation from academia, noting that top business students are valued in finance for uniquely human skills that AI cannot replicate.”
X
Teaching an experimental class for MBAs on “vibefounding,” the students have four days to come up and
[Deployment]
“DeepSeek Summary: Mollick is conducting an innovative MBA course on 'vibefounding,' a rapid, experiential approach to entrepreneurship or ideation.”
X
what a perfect space for scientific discourse! I'll start off with a few images of myself
[Evaluation]
“DeepSeek Summary: Sarcastic commentary on scientific discourse spaces, possibly referencing platform dynamics.”
X
Life update: I'm starting as faculty at Boston University in 2026!
“DeepSeek Summary: Career announcement about joining Boston University as faculty in 2026.”
X
It's uncanny right?
“DeepSeek Summary: A brief, possibly rhetorical or observational tweet expressing a sense of strangeness or coincidence.”
X
This stupid website is so cooked.
[Tooling]
“DeepSeek Summary: A brief, critical comment about a website, expressing frustration or disappointment.”
X
And awesome to see many Berkeley alums thriving here. @LaurentLessard, @DimitrisPapail, and Shivaram
“DeepSeek Summary: A positive acknowledgment of UC Berkeley alumni success, tagging specific individuals.”
-- END OF LOG --
[STATS] 59 items · Filter applied
Powered by Horizon + DeepSeek