The law of laziness is that it will propagate to whatever leverage/abstraction you're working at. For example, I know the right way to prompt to get what i want. I'm feeling too lazy to put the work in 😭

Manifold is building open-source foundations for agents that handle long-horizon, autonomous tasks across both GUIs and APIs - the kind of complex workflows humans do daily. Everything we develop will be open science. Join us in creating AI that can truly navigate our digital

Building AI reasoning models with extremely long context lengths - think days, weeks, even years of context - is the next big challenge in AI. that's why i'm extremely excited about the latest work from Ao Qu Ao Qu, incoming PhD student in our group, on MEM1: RL for Memory

LLMs are getting so damn good at hallucinating

We (E2B) raised a $21M Series A. And we are hiring the best team to help us grow even more and power next generation of AI agents with our sandboxes. ✴️Check out e2b.dev/careers Very proud of team ❤️

IN NEWS: Julius AI raises a $10M round. We spoke with Founder & CEO, Rahul Sonwalkar (rahul) about data accessibility within a company. "Companies have gold mines worth of data and they aren't able to get the insights they need because every time somebody on the team

Context is king.

Frontiers talks bring experts from the frontier of fields beyond just AI to discuss key problems, challenges, and opportunities. Really excited to have our first talk on the topic/area of ZK proofs from the legend Prover Killer

There are new scaling laws to be discovered folks - keep pushing!

Missed Community Research Call #5? The recording is available now! We unveiled new insights into LLM calibration, benchmarked the next generation of multimodal action models, plus modular space system assembly and crowded orbit navigation. Check out the recording below!

1/ we're extending the application deadline for @Conviction Embed, our grant program for AI-native startups to 8/14 it's the best time in history to start a company

Another of our ongoing projects is VLM Calibration, which aims to answer the question... How well can vision–language models really judge the correctness of their own answers? Here's what we've found so far

This Saturday at 10 AM PDT is our next Frontiers Talk: Breaking ZK Provers to Build a Stronger Ethereum with Prover Killer, founder of Prooflab. More detail below!

One of our ongoing programs is Aethermorph: modular swarm robotics for in-space assembly, repair, and reconfiguration. We’re developing methods for building modular structures and new methods for safe navigation in crowded orbits. Here’s what we’ve built so far.

after Fetchr, I spent some time soul searching and reflecting during that time Julie Zhuo helped me better understand myself and my relationship with conviction as Dalton Caldwell said during my YC batch, “conviction is the only KPI that matters”

We are recruiting an OS Operations Fellow to help architect the systems behind open, high-impact research. You will cultivate Manifold’s community, design impactful events, and communicate cutting-edge research to the world. Apply below👇

People often ask how did the Unitree robots get so good all of a sudden. It wasn't all of a sudden, and it's because they ship their hardware and open source their SDKs. Arguably these robots are nearly useless out of the box, but you have full dev control of them. Because of

1-3 will all exist

Another legendary run incoming

Think about technology the way you think about your diet. There’s junk food. Empty calories. Things that fill you but don’t really nourish. You consume, but you don’t grow. There’s food you cook. It takes time, effort, presence. You slice, stir, wait. Because you invested