fpl9000.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

Pro-tip: Press Ctrl-Y in Gemini CLI to toggle YOLO mode. :)

submitted 1 day ago • 0 comments

Today we’re releasing a prototype of Genesys, an autonomous multi-agent LLM discovery system that aims to discover new types of language model architectures. We found Genesys can discover novel architectures competitive with the industry-standard transformer. 🧵

submitted 2 days ago • 1 comment

Just using Gemini CLI to modify its own source code.

submitted 2 days ago • 0 comments

My notes on Gemini CLI, including poking around in their system prompt which I've extracted into a more readable rendered Gist simonwillison.net/2025/Jun/25/...

submitted 4 days ago • 5 comments

Claude just got hands-on. 🛠️✨ Now you can build, host, and share AI-powered artifacts—with zero backend hassle. Jump into the new artifacts space to remix examples, organize your projects, and embed Claude’s smarts directly into your creations.

submitted 4 days ago • 2 comments

Mindscape 319 | Bryan Van Norden @bryanvannorden.bsky.social on Philosophy From the Rest of the World. There are many interesting footnotes, not all of them to Plato! www.preposterousuniverse.com/podcast/2025...

submitted 7 days ago • 2 comments

Fin Moorhouse interviews @tobyord.bsky.social about the future of AI and the risks it involves. Clear and illuminating. pnc.st/s/forecast/5...

submitted 13 days ago • 0 comments

I'm using Claude Sonnet 4 and coding agent Cline (a VSCode extension) to create a graphical hex dump utility in Go. I'm not a GUI developer, so to fix GUI bugs, I just upload a screenshot showing the issue and ask Claude to fix it.

submitted 10 days ago • 0 comments

No signs of an end to rapid gains in AI ability at ever-decreasing costs, yet I did my best to update my chart to take into account the price drop in o3 & new models released by Google GPT-4 was released 2.25 ago, so its worth noting the trend when considering the future of AI capabilities & cost

submitted 12 days ago • 3 comments

Generative AI News https://nicolehennig.com/gen-ai-news/ (my free monthly newsletter) #AI #Substack #newsletters

submitted 229 days ago • 0 comments

OpenHands (formerly OpenDevin) released a coding CLI. " - Has top accuracy (similar to Claude Code) - Is completely open source, MIT licensed - Is model agnostic, use an API or bring your own - Is simple to install and run `pip install openhands-ai` and `openhands` (no Docker!) "

submitted 12 days ago • 3 comments

Stephen Welch (of Welch Labs) has published a great video explaining backpropagation, the mechanism used to train neural networks. There's some calculus, but the graphics make it easy to follow. www.youtube.com/watch?v=VkHf...

submitted 12 days ago • 0 comments

My notes on Anthropic's substantial essay about how they built their multi-agent research system, which has finally talked me around to taking multi-agent LLM prompt engineering seriously simonwillison.net/2025/Jun/14/...

submitted 15 days ago • 4 comments

By surveying workers and AI experts, this paper gets at a key issue: there is both overlap and large mismatches between what workers want AI to do & what AI is likely to do. AI is going to change work. It is critical that we take an active role in shaping how it plays out. arxiv.org/pdf/2506.06576

submitted 16 days ago • 2 comments

I really don't know if I should waste any time pointing out what a load of crap this article and the Apple paper it references are. Has someone already provided a good debunk I could just link to instead?

submitted 18 days ago • 4 comments

"[...] the average [ChatGPT] query uses about 0.34 watt-hours, about what an oven would use in a little over one second, or a high-efficiency lightbulb would use in a couple of minutes. It also uses about 0.000085 gallons of water" - Sam Altman in blog.samaltman.com/the-gentle-s...

submitted 19 days ago • 15 comments

Finally, the perfect blog post. The only thing wrong with it is that I didn't write it.

submitted 27 days ago • 0 comments

Mindscape Ask Me Anything | June 2025. We talk about the anthropic principle, AI and copyright, the Heisenberg uncertainty principle, the current disaster in science funding, and more. #MindscapePodcast www.preposterousuniverse.com/podcast/2025...

submitted 28 days ago • 6 comments

I recently participated in the High-Level Expert Panel on AGI, convened by the Council of Presidents of the UN General Assembly, which has just released its final report on the Governance of the Transition to AGI. Report below ⬇️ uncpga.world/agi-uncpga-r...

submitted 29 days ago • 0 comments

Some notes on the new DeepSeek-R1-0528 - a completely different model from the R1 they released in January, despite having a very similar name Terrible LLM naming has managed to infect the Chinese AI labs too simonwillison.net/2025/May/31/...

submitted 29 days ago • 2 comments

My local Wendy's drive-thru has an AI taking orders. It works surprisingly well. I asked it a series of detailed questions about the 6 kinds of Biggie Bag, and it answered every one with the info I needed. I would never have tried that if it was mumble-mouthed teenager behind the speaker.

submitted 35 days ago • 1 comment

Mindscape 316 | Niayesh Afshordi @nafshordi.bsky.social and Phil Halper on The Big Bang and Before. #MindscapePodcast www.preposterousuniverse.com/podcast/2025...

submitted 35 days ago • 3 comments

I started reading the 120 page Claude 4 System Card and found myself unable to put it down. If you're in the mood for some excellent science fiction (or you miss Person of Interest on CBS) I cannot recommend it enough. I made detailed notes on my highlights here: simonwillison.net/2025/May/25/...

submitted 36 days ago • 7 comments

submitted 38 days ago • 0 comments

Google’s AI video tools are booming—meet Sparkify, their latest experiment. Just drop in a question or idea, and Sparkify uses Gemini + Veo AI to instantly generate a 2-minute animated explainer—perfect for education.

submitted 39 days ago • 3 comments

We're ready for Microsoft Recall and the automatic screenshots it takes of everything on your desktop. Signal Desktop on Windows now includes support for a new "Screen security" feature designed to block screenshots of your Signal chats. signal.org/blog/signal-...

submitted 39 days ago • 15 comments

The article fails to name the coding agent from Anthropic. Is this the intelligent backend of Claude Code coupled tightly to GitHub APIs? Is this drawing from the same pool of compute that serves Claude customers? www.reuters.com/business/mic...

submitted 40 days ago • 0 comments

The NotebookLM mobile app is rolling out! This initial version has an MVP feature set with more to come! Listen to Audio Overviews offline: Download them for offline playback anytime, anywhere. Also, play them in the background while using your device, online or offline.🧵1/4

submitted 41 days ago • 1 comment

Techies propose the Agent Name Service: It's like DNS but for AI agents

submitted 41 days ago • 1 comment

Mindscape 315 | Branden Fitelson on the Logic and Use of Probability. Finally we decide whether all ravens are black. #MindscapePodcast www.preposterousuniverse.com/podcast/2025...

submitted 42 days ago • 6 comments

Research finds similarity between the way large language models (LLMs) like ChatGPT process information and how the brains of people with Wernicke’s aphasia function. In both cases, fluent but often incoherent output is produced, suggesting rigid internal processing patterns that can distort meaning

submitted 43 days ago • 0 comments

I wish these skeptical AI articles (this is from the NYTimes) would actually grapple with the growing body of research that AI can do original research & perform key unstructured tasks across the spectrum of high-end white collar employment. AI criticism is important, but it should be clear-eyed.

submitted 44 days ago • 6 comments

Great conversation with Boris and Cat, the lead dev and PM for Claude Code, on the Latent Space podcast. They said ~80% of Claude Code's source was written by Claude. 😲

submitted 51 days ago • 1 comment

Nineteen years after his death, the universe is still struggling to catch up with the vast creative force that was Stanisław Lem. Get a great deal on the ebook edition of his classic sci-fi tale, "His Master's Voice," over at BookBub: www.bookbub.com/books/his-ma... #scifi

submitted 52 days ago • 1 comment