Ex-dev → lawyer → back to dev. Founder of Million Opportunities. Helping founders find business ideas from real market signals. Work as sadhana, code as craft.
NYC
Joined February 2026
- Tweets320
- Following10
- Followers11
- Likes97
Fix your AI slop from the command line using the /deslop skill.
github.com/fayerman-source/d…
A professional plain English style guide and agent skill that translates legalese and purges AI slop.
A neurologist and a lawyer (me) ran opportunity scans together (1mil.app). I couldn't tell a legit medical idea from a BS one. He could, instantly.
In one session we pulled two real, shippable products out of the noise. Both came from him recognizing pain I literally couldn't see on the page.
Bring Your Own Moat isn't a slogan. It's the unlock.
The most useful thing I've done with 1mil.app: I stopped using it alone. I started booking sessions with people whose worlds I know nothing about - doctors, accountants, estate planners, contractors.
The tool surfaces opportunity. The expert tells you which ones are real and which ones are hallucinations dressed up as insight.
That's the moat.
Opportunity is everywhere, but you better Bring Your Own Moat.
I built 1mil.app on one belief: every market, every niche, every workflow has an opening. You point the tool at something you're curious about and it ranks the directions worth building - competition, money, difficulty, and most importantly: moat.
Low moat = a 13-year-old vibe-coder ships it by Sunday. High moat = the opportunity is locked behind domain expertise you don't have.
Free scan, no account: 1mil.app
A lawyer and a neurologist walked into a startup-discovery tool. The lawyer couldn't tell a legit medical opportunity from a BS one. The neurologist could, instantly.
Searched X for startup ideas, and most posts said you didn't need "one" - right, you need a million to choose from.
This is kind of nuts.
@karpathy made a tiny repo where an agent can do actual research-loop iteration: change code, run a fixed timed eval, keep/discard, repeat.
P.S. I had to adapt it for a normal #NVIDIA card.
github.com/fayerman-source/a…
The next step for autoresearch is that it has to be asynchronously massively collaborative for agents (think: SETI@home style). The goal is not to emulate a single PhD student, it's to emulate a research community of them.
Current code synchronously grows a single thread of commits in a particular research direction. But the original repo is more of a seed, from which could sprout commits contributed by agents on all kinds of different research directions or for different compute platforms. Git(Hub) is *almost* but not really suited for this. It has a softly built in assumption of one "master" branch, which temporarily forks off into PRs just to merge back a bit later.
I tried to prototype something super lightweight that could have a flavor of this, e.g. just a Discussion, written by my agent as a summary of its overnight run:
github.com/karpathy/autorese…
Alternatively, a PR has the benefit of exact commits:
github.com/karpathy/autorese…
but you'd never want to actually merge it... You'd just want to "adopt" and accumulate branches of commits. But even in this lightweight way, you could ask your agent to first read the Discussions/PRs using GitHub CLI for inspiration, and after its research is done, contribute a little "paper" of findings back.
I'm not actually exactly sure what this should look like, but it's a big idea that is more general than just the autoresearch repo specifically. Agents can in principle easily juggle and collaborate on thousands of commits across arbitrary branch structures. Existing abstractions will accumulate stress as intelligence, attention and tenacity cease to be bottlenecks.
