Flow Learn is live.
A free, open-source curriculum for builders who want to understand AI agents, blockchain, protocol engineering, and the systems behind them.
We’re curating relevant resources and shaping them into a path that keeps evolving.
https://t.co/AK8BqmaeNx
AI agents are getting better at completing tasks. But does success mean reliability? Can they still perform when conditions change?
At Flow Research, we're looking beyond successful outputs to understand what makes AI agents genuinely reliable.
Where do you think current AI benchmarks fall short?
Your team and your AI agent, finally in one workspace.
Garden helps teams plan, build and get things done side by side, with your tasks, files and tools all in one place
Take a look 👇
At Flow Research, we're interested in the questions beneath the headline performance numbers.
Building capable agents is one thing. Building agents that can be trusted to do reliable work is another.
Just because an AI agent gets the job done doesn't mean it gets it right.
Not necessarily.
There's a difference between finishing a task and doing it correctly and consistently. And that's what we need to pay attention to.
Evaluating agents means looking beyond task completion.
We need to understand when they succeed, where they fail, how consistently they perform, and whether their results are reproducible.
Tomorrow evening, come hang out with us on X Spaces.
We’re chatting about how credit and rewards might work as humans and AI agents team up, and more insights to share.
🕖 7:00 PM WAT | 11:30 PM IST | 2:00 PM EDT
🎙️ Set a reminder: https://t.co/vCfAX573ox
Tomorrow evening, come hang out with us on X Spaces.
We’re chatting about how credit and rewards might work as humans and AI agents team up, and more insights to share.
🕖 7:00 PM WAT | 11:30 PM IST | 2:00 PM EDT
🎙️ Set a reminder: https://t.co/vCfAX573ox
4 months ago, our first cohort of fellows joined Flow Research to learn, collaborate, and contribute to the work we’re building.
Today, they get to show what came out of it. We’re hosting a demo day where our fellows will walk through the projects they’ve worked on and what they’ve picked up along the way.
As AI agents take on more work, how do we decide who gets credit, what counts as valuable work, and how people should be rewarded?
This Saturday, join Omidiora Samuel @samparsky (Co-founder, @pagrin_) and Gbolahan Adebayo @sheys_mst (Co-founder, @FlowResearch_ ) as we explore fair rewards and recognising contributions.
🗓️ Saturday, 31 October 2026
🕖 7:00 PM WAT | 11:30 PM IST | 2:00 PM EDT
Set your reminder: https://t.co/yl3zoETTtx
It’s finale day at the Africa Deep Tech Challenge.
We are supporting this effort to build AI that runs on commodity hardware. Good luck to every finalist!
📅 October 17th
Save a spot: https://t.co/d15Ca2vaCF
It’s been an honor to serve as a judge in the Africa Deep Tech Challenge. The finale is happening today.
Come and witness AI built to run on commodity hardware, including everyday laptops.
I helped build the profiler, the tool used to measure performance across the submissions.
Additionally, @FlowResearch_ has donated $4,000 to the prize pool.
Looking forward to hearing from the finalists and seeing what they’ve built.