Just finished Dario's essay and listened to Brockman on @a16z pod. No take yet.
I want the ideas to sit for a bit. Hot takes are cheap. Clarity isn’t. Yaddada mean!?
Sweden’s best founders should leave Sweden.
At least for a week.
We started Velocity Fellows so 10 young Swedes can spend a week in SF on us.
In 2023, we took Max from Legora to meet Arvid at Cursor. Cursor had 10 people. Legora had 2.
Apply to get a picture like this!
Yep, and... @roshan_s@prades_maxime - how do we get all users to move from 1% to 99% utilization/value? Build more trust in the agents? Clearer product education? Show, vs tell videos on social? Agent self-documentation and proactive guides to go up the trust + utility ladder?
SpaceXAI engineer Lauren Tan on why most users barely scratch the surface of GrokBot
"GrokBot is the most powerful agentic tool we have ever built, but only 1% of users use it correctly."
Her setup: 20+ GrokBot agents, organized around a Chief of Staff, 3 managers and 16 workers.
Meanwhile, 99% of users are still working inside a single chat window and wondering why everything takes so long.
The missing piece is how the work gets organized. And it takes one evening to change that.
Chief of Staff → Managers → Workers → Skills → Autonomous Workflows
In a 1-hour workshop, a SpaceXAI engineer breaks down how to turn Grok agents into a coordinated team and put their capabilities to work.
Research → Build → Launch → Improve
Worth more than most $1,000 agent engineering courses online.
Skip Netflix tonight. Watch the workshop and rethink how you use GrokBot.
@arnavsahu341 Both can be true. Distribution, hiring, and cash discipline still transfer. The 10% that breaks is usually “how we used to ship” advice applied to a world where the build loop is days, not quarters. IMVHO.
@JenniferHli Exactly. If the people building the frontier also own the scoreboard, you don't have a scoreboard, you have marketing(ish). Independent evals are how buyers and builders actually calibrate.
The debate on pacing the frontier makes one thing increasingly clear: the labs pushing the frontier shouldn’t also be the only ones grading it.
As an industry, we need credible third-party evals - an independent source of truth on where the frontier actually is, and what risks are emerging.