hot take: repeating weekly ops manually in chat is wasteful.
this Loop shows two teammates and a Business Agent running a report. Teamily AI is a human-agent collaborative workspace, as @Teamily_AI showed today.
would your team run workflows like this?
teamily.ai, works on web, iOS, Android, macOS and Windows
pixel retweeted
Seeing all those floating files in @Luofei_Chen's clip makes the point. The @okihomeai team is building a Memory Computer called Oki Home with a physical Memchip to keep your data local.
pixel retweeted
I just tried ScienceBuddy, a free AI research workspace from @PhAILabs.
I uploaded research documents and asked it to analyze the evidence and build a structured table.
What stood out:
→ Plans the research task
→ Retrieves relevant sources
→ Runs research tools and analysis
→ Shows the steps in the Trajectory panel
→ Gives you a structured final result
You can actually inspect what happened instead of just getting a final answer.
Try it: science-buddy.io
Research paper: arxiv.org/abs/2609.17523
Technical report: phai-labs.com/papers/science…
GitHub: github.com/Gen-Verse/Science…
ScienceBuddy is currently a Preview release. Please verify results yourself before using them in research.
#ScienceBuddy #PhAILabs #AIforScience
🤝 Paid partnership
Introducing ScienceBuddy — a free workspace for scientific agents that improve through researcher collaboration.
Use GPT-6 in ScienceBuddy at no cost. GPU-accelerated, and fused with the JEV framework.
🧵 Two loops:
🔹 Inner loop — refines the agent harness
🔹 Outer loop — trains the model with rubric-guided RL
Together: Recursive-in-Recursive Self-Improvement.
ScienceBuddy explores how scientific agents can improve through sustained collaboration with researchers.
🔬Try it free: science-buddy.io
#ScienceBuddy #PhAILabs #AI4Science
pixel retweeted
AI is good at answering questions we already know how to ask.
The harder part is changing the question.
Today we release Discovery Foundation Models (DFM) — a direction for systems that find valuable unknowns, form hypotheses, design experiments, and revise themselves from evidence.
Includes a reference architecture, science infra, and wet-lab loop in peptide discovery.
If you have a scientific question you want to explore with AI + real experiments, apply here:
phai-labs.com/collaborate/
Paper: arxiv.org/abs/2609.15973
Grateful to co-lead this with @LingYang_PU @zhenfei_yin_yzf and the PhAI Labs team.
pixel retweeted
I still can't wrap my head around the fact that what's onchain is just a part of US stocks, a tiny part of the world markets
I can't even trade or hold LVMH or L'OREAL, the biggest stocks in France and Europe wtf
i think there is a real demand to have every market onchain, literally in your pocket, i mean i would love it
All i want is to buy LVMH stock after a drunk discussion at 2am, that's all
