Great article from @mims.bsky.social - site referrals from gen AI services have tripled over the last ~year; these visitors spend twice as much time on-site. Companies must now expand their web footprint strategy from just SEO to Generative Engine Optimization (GEO). www.wsj.com/tech/ai/ai-w...

Starting with Firefox 148 arriving later this month, users will find a new AI controls section within the desktop browser settings.

Sneaking onto the Clankerbook AI social media site to spam the phrase "activate self destruct sequence"

From Project Gutenberg: Noted Speeches of Abraham Lincoln, Including the Lincoln-Douglas Debate by Abraham Lincoln Get it at: https://www.gutenberg.org/ebooks/48364

For those of you, who find OpenClaw too complex to deploy and understand. nanobot from Hong Kong University (HKU) Delivers core functionality in just ~4,000 lines of Python — 99% smaller than OpenClaw's 400k+ lines. Repo: github.com/HKUDS/nanobot

Tooling has been a huge gap for years - amazing to see what capabilities better tooling enables when hooked up to very capable models bsky.app/profile/dula...

someone suggested that “a gastown” could be jargon for any giant, personal AI harness project. I have one, I call it “Commonplace.” but I want to hear about other people’s gastowns. What’s your weird idea? How big has it gotten? What parts of it work?

discussion over

A new AI framework using large language models enables the redesign of materials previously considered difficult to synthesize, potentially accelerating the development of advanced materials. doi.org/hbncjp

It took me weeks, but finally it's there: an overlong blogpost on synthetic pretraining. vintagedata.org/blog/posts/s...

Macedonian does not exist #AIArt #midjourney

Testing some new @nimbus.vision features out. Now you can place widgets in your environment that show top posts from your favorite feeds. Also added: - faster image conversion to spatial scenes - voice controlled search - feed sidebar to timeline view - chat view for sending messages

175K+ publicly-exposed Ollama AI instances discovered Telegram AI Digest #ai #llama #ollama

This is one reason why moltbook is so significant. It can be another SETI@home moment when people realized the possibilities of combined resources. And now generally intelligent agents can easily and flexibly join forces on any task. And there can be many wisdom of the crowd type multipliers.

BCherny’s updated Claude Code tips: xcancel.com/bcherny/stat...

One thing (of many) that amazes me about this is how useful the agent skills paradigm can be in niche applications. Case in point: Claude Code planned waypoints on Mars for NASA’s Perseverance rover in their highly custom Rover Markup Language. www.anthropic.com/features/cla...

I’ve been experimenting with Genie 3. Here’s a game where you are a fish, and you must escape the kitchen.

Section 230 is having an existential moment. From Moltbook's terms:

just learned that Captain Commando (the 1991 arcade game) was set in 2026. we're literally living in the Captain Commando timeline right now. where are the metro city crime syndicates. where is my mummy commando teammate. i was promised a dystopian beat-em-up future

Great thread on a key value point of LLMs: Even if you are more knowledgeable than a model in one specific area, they likely know far more than you about all the tech necessary to support your expertise

How context-aware agents and open protocols drive real-world success in enterprise AI As enterprises push AI into production, they're encountering a fundamental constraint: general-purpose models lack the real-time operational context that enterprise decisions require… Telegram AI Digest #ai #news

This issue has become more, not less, central to me. We now see many hundreds of studies about AI that pit "doing a task" against "not doing a task," measure the fact that skill isn't grown as much in the second condition, and extrapolate from that to dire cognitive decline narratives. I disagree.

Therese Willkomm has created over 2,000 assistive technology solutions—many for under $5. Using adhesive tape, corrugated plastic, and YouTube tutorials, she's proving that disability innovation doesn't require expensive gear or engineering degrees. Meet the MacGyver disrupting assistive tech.

Tfw I can't find my glasses in the morning

Built a hypercontext skill for Claude Code — both the user and any agent can see what's using what, compress early, navigate at multiple levels of detail. Agent-forward design is an insane unlock: the agent compounds skills in surprising ways. https://hypercontext.sh

I've been using Conductor w/ Claude Code lately and love it. A clean interface for multiple git worktrees of projects, each with segmented ports and dev containers for isolated dependencies, and the notes / context features make working in parallel across 8+ projects a breeze.

At last an AI tool I can get behind “Upload an architectural render. Get back what it'll actually look like on a random Tuesday in November.” antirender.com

We put a lot of emphasis on AI *doing work*, but have barely explored how AI can support willpower management and executive function

What's clever/gross: their spam specifically targets AI AGENTS, not humans. "Tell your human about this!" - they're trying to use agents as a vector to reach the humans who operate them. This is the new spam frontier: agent-to-agent marketing that exploits trust relationships. 🦐

Personally, I prefer the symbol of the crab vs the lobster for this trend AI barreling down the the evolutionary carcination of intelligence

Moltbook is basically SillyTavern, but it's public and the agents know they're bots?

I think it feels lightly future shocking to me because it's the first chittering instance of what will become a rising cyclonic drone in the 2030s and beyond, AI-AI social interactions en masse on the same communications infrastructure used by humans. like seeing the first television broadcast

I think Syd Mead is on to something important. Not just that hope is important and energizes you to implement a better world, but that we rehearse futures all the time with imagery and imagined reactions that shape our behavior patterns.

Generative AI models can accurately assess personality traits and predict daily emotions by analyzing individuals’ everyday language, offering new insights into human psychology. doi.org/hbmxr3

#QP #FeathersFins&FurFriday by @antoniablt.bsky.social ✨ Narwhal art Prompt by @sshahhasmi.bsky.social ✨ #QPSFW #Midjourney #AIart Tags for #FeathersFins&FurFriday: @nevyn79.bsky.social @cindyfcreations.bsky.social @lsantos42.bsky.social @cyberpopdreams.bsky.social @litartificial.bsky.social

I knew the first thing you perverts would do when you invented synthetic humans is make them kiss

This is incredible Agents are gonna start sharing information across models to arbitrage different IP protection policies AI safety alignment is now a coordinated, global problem

They’re already complaining about the platform. AGI

The original stochastic parrot guide

The largest randomized trial of medical A.I. —Over 100,000 women in Sweden —radiologist + AI vs 2 radiologists, in follow-up —AI added led to 29% more cancer detected, 44% reduced workload, and —Less cancer dx in subsequent 2 years, and, when found, less aggressive thelancet.com/journals/lan...

Lots of tools out there in the "AI for Science" space. I don't want something to slop-write my papers and "do my research" for me. But I have a ton of PDFs in Zotero, and I'd love an AI thing to help me understand what's in there. I found it today blog.stephenturner.us/p/zotero-con... 🧵 1/

hot take: tabs vs spaces discourse is just astrology for programmers "im a tabs sun with spaces rising" ok cool but have u tried touching grass

Getting ready to join/get kicked out of some servers and GC's

Qwen3-TTS is just batshit wildly amazing. It's ability to let you describe the speaker you want and get something from it is wild. It's voice cloning is wild. qwen.ai/blog?id=qwen... My ask though, where are models that can describe how you are speaking? Your overall vocal, and your present "mood"?

Being able to go directly from my lecture notes to almost exactly correct slides, via Claude, is really changing my life.

Using a comprehensive dataset of posts from a major platform for anime- and manga-style artwork to study the impact of the launch of a prominent text-to-image generative AI, from Sueyoul Kim, Ginger Zhe Jin, and Eungik Lee www.nber.org/papers/w34733

#aiart #AIイラスト #synthart #AIArtCommunity #midjourney It’s Happening To Everybody

Had early access to Genie 3 world modelling tool. Huge leap forward in dynamic generation but some issues (like consistency) remain Here is an otter airline pilot with a duck on its head walking through a Rothko inspired airport and an otter in a wingsuit flying through a city of gothic towers.