Prashaant Ranganathan
@LongHorizonTask
I help scientists and AI agents collaborate on long horizon tasks for scientific discovery. CS and PD @stanford, winner @Intel ISEF, @a16z backed, @outintech
see the future 5 months earlier by reading our blog
📁 Sam Altman says the real AI breakthrough will not be better reasoning, but total memory. An AI will be able to remember every conversation, email, and document across a person’s lifetime, identifying patterns and preferences humans never consciously express. Once memory…
Gemini model post-trained to support teachers and students is huge ☺️ we need more initiatives in the US like this!
imo this is a HUGE deal Google DeepMind and the UK govt have just signed a partnership w/ 3 pillars: 1. transforming public services 2. accelerating scientific discovery 3. advancing AI security & resilience Concrete bits that matter: - a Gemini model trained on the UK…
Great open source tool
🧵 Labs and VC's are throwing cash at RL environments, especially for computer and browser use. Yet, with just 4 customers and over 30+ vendors, is cloning every website in the world really the path to scale? of course not. Introducing TRACE: Trajectory Recording and Capture of…
Let's go!
.@NSF is launching Tech Labs to build and scale next-gen independent research organizations to advance science. These orgs will tackle technical challenges traditional university and industry labs can’t solve alone. Share your feedback: bit.ly/44ZtMw7.
Announcing the ICML 2026 policy for LLMs in reviewing! Reviewers and authors both pick either conservative or permissive LLM use, and will be matched accordingly. Importantly: authors on papers who choose conservative must obey the conservative policy as reviewers.
In 2026, multi-player AI will eat single-player AI. Right now, most AI tools are built for one human + one model in a private workspace. ChatGPT, Cursor, Claude. Incredibly powerful, but currently optimized for individuals. The impact is massive: drafts, code, specs, campaigns,…
we need more founders like @sama. in my last quarter at stanford, i emailed him while I was thinking about what to do with my life. he replied with: “if you feel excited about it, why not dive in? seems like a good time in your life for it.” it didn’t feel like advice as…
feels like the hack of actually being able to inspect ai scientist decisions is the move. black box research automation is just expensive randomness
what would it look like if @EdisonSci’s kosmos, @SakanaAILabs’ ai-scientist-v2, and @claudeai’s scientific agent all worked on the same research project? right now, most agents for scientific research ship as a paper + open source repo. no place to actually host them, compare…
What if you could see an AI scientist think? These past few weeks, @LongHorizonTask, @belindmo and I have been testing various AI scientist agents to improve our ML experiments.
thanks @SakanaAILabs, you are amazing! <3 we wired ai scientist v2 into an observability layer. every run becomes a trace: prompts, files touched, code edits, metrics. you can replay its actions step by step. Worked with @belindmo @turboblitzzz on it!
What if you could watch an AI Scientist think? We built an interface to make @SakanaAILabs’s AI Scientist-v2's interpretable - so you can see every hypothesis, every failed experiment, every "aha" moment. Here's what we learned 🧵 (1/8)
First, here is the hosted AI Scientist-v2: platform.sundialscientific.com Shoutouts to @LongHorizonTask and @turboblitzzz as collaborators (2/8)
United States Trends
- 1. Christmas 2.88M posts
- 2. Santa 1.66M posts
- 3. Hawaii 24.5K posts
- 4. Feliz Navidad 892K posts
- 5. EigenCloud x DataHaven 1,211 posts
- 6. Merry Xmas 303K posts
- 7. Marty Supreme 18.3K posts
- 8. Clavicular 4,772 posts
- 9. #ShowSVPTheTree N/A
- 10. Chan 186K posts
- 11. God is Great 9,592 posts
- 12. Die Hard 16.4K posts
- 13. Happy Birthday Jesus 26.2K posts
- 14. jisung 42.8K posts
- 15. Rudolph 35.9K posts
- 16. Rainbow Warriors 1,736 posts
- 17. Dan Fouts N/A
- 18. RIKI 13.8K posts
- 19. Drew Brees N/A
- 20. Howard Stern N/A
Something went wrong.
Something went wrong.