@coinbase laid off 14%. @blocks laid off 40% from 10,000 to just under 6,000.
Both got there through layoffs. @jack got there through a manifesto - "From Hierarchy to Intelligence". And that a great read. To an extent that @ycombinator has RFS that's exactly his manifesto as a product for other companies.
His article names three roles inside a company:
→ ICs build the work
→ DRIs own the metric
→ Player-coaches build capacity in the people around them by doing the work alongside them
Being AI-native at a mature company means flattening hierarchy to get to this. Being AI-native at an early-stage company is different. Every employee, along with the founder, is already a mixture of IC and DRI, with the added responsibility of being a player-coach to the founder.
Somewhere along my career, I've played all three and somehow across startups and a mature company.
At @gohighlevel, the metric that I started working on last summer was trial-to-paid conversion. After enough customer calls I found a motivation no team was targeting - a marketer who didn't want his job made easier. Automating the boring tasks still didn't give him his evenings back. His kids were in college and he was spending the evenings on ad campaigns instead of his wife.
So, I proposed an end-to-end email marketing agent with dollar value coming in as the metric to optimize. Not a workflow automator. An optimizer.
DRI and IC, same person. And in this case, AI.
That's what I built Pepper for. Pepper is everything I left mid-way there - but for any optimizable metric across product like latency
Two jobs in one platform:
1. Figure out what to build, then build it — give Pepper a high-level goal and it spins up an agent swarm to get there.
2. Once you have a business going, point Pepper at a metric and it runs the optimization loop: analysis → hypothesis → experiment design → execution → result. On loop. No human in the middle.
This week I used Pepper (powered by pi coding harness) to ship Claude Agents SDK support inside Pepper itself, so users can bring their own Claude tokens (After the recent news, I know but nevertheless it's still there). Shipped in a day. Only human effort: UAT.
Pepper building Pepper and eventually Pepper running Pepper!
Video below shows Pepper running the optimization loop on an ads → landing → activation funnel. Workspace creation up 10%.
Honest status: all four layers in infancy. Shipping daily. Looking for feedback, and for more work, from anyone running team-of-one who wants the intelligence layer behind them and don't want to build it by themselves.
#buildinpublic #soloprenuers
You've been told AI agents will run companies and single person unicorns will be possible repeatedly Sam Altman and others.
The problem isn't the models but the tooling!
Then you ship one feature with them and realize: you're still the planner, the memory, the QA, re-pasting context into 7 amnesiac parallel sessions to ship one thing quickly.
You are the context bus.
We built @buildwithpepper , the OS for AI native companies to take that job away from you. 🌶️
Startups don’t fail because people can’t build.
They fail because they pick the wrong idea… and can’t get distribution.
I’m building something to fix both.
#buildinpublic#indiehackers#entrepreneur
Startups don’t fail because people can’t build.
They fail because they pick the wrong idea… and can’t get distribution.
I’m building something to fix both.
#buildinpublic#indiehackers#entrepreneur
@CuriousGuy2020@waitbutwhy You'd also find names such as Guru Nanak (just below Babur) and Mahatma Gandhi.
The list may seem biased but it has attempted to acknowledge a few prominent names from the Eastern world.