Founder, designer, and engineer.
Joined October 2025
- Tweets 757
- Following 287
- Followers 242
- Likes 2,699
60 Photos and videos
one of the biggest misconceptions about design is that it's just about pretty colors. but you can't make sensible design decisions without a deep sense of strategic understanding, positioning, and signaling.
I was talking to a founder earlier who wanted product feedback, and was asking my thoughts on his design. I could give the simple answer of a few aesthetic nits like button positions or font choices. and most designers would have probably given that answer. but the position of a button means nothing if the problem frame is wrong to begin with, or if the product's utility isn't being clearly communicated in the first place.
so it spawned a whole brainstorming session about who his competitors are, what his position really is relative to others, the product category — what is this thing? is this a data product? why couldn't I just use Claude?
at one point he mentioned a $250,000 deal that he lost because he misread the org topology and a power player blocked the deal, and the visceral human relatability of that story that finally made his product click — that he was building something to understand the enterprise buyer room. and at that point we could iterate on how to communicate that value to his customers, and design the product around this core message.
design is an iceberg. there's nothing more important than understanding how deep it goes.
2
267
Wispr Flow charges $15/month to send your voice to a server so they can harvest your data.
You can have this for free. Fully private.
parrot is free, open source, runs on-device, and is ultrafast thanks to Apple Neural Engine (does require silicon).
digimata.github.io/parrot/
2
6
339
this is the way education should be done for everything now
We're hiring four recent graduates (or folks looking to make a career change) for the world's best VC training program.
Tuition: $0
You'll work 60 hours a week, get paid what you would have spent on your MBA tuition and learn more than you can imagine -- unless you quit because you can't handle the pace.
You start as a researcher
Then you become an analyst
... and 1.5 in three make it to associate
You need to:
1. Be a learning machine
2. Have high executive function
3. Be extremely focused and curious
4. Be able to work 12 hours a day for our founders
We accept < 1% of applicants, we don't care about how fancy your degree is, we love folks from @UTAustin, and we love folks with a chip on their shoulders
Four slots, program starts in May/June... email a cover letter on why you want to be a venture capitalist, what skills you currently have and give us some examples of your work ethic...
... because 80% of VCs are lazy AF, and you can beat them by simply showing up for work and doing 10-12 hours a day and checking your email on the weekends.
Not kidding... these VCs are all calling in rich and skiing in Japan. They don't really do much work.
researchers@launch.co
[ samurai/jedi only ]
1
148
our credentials system is fundamentally broken, there are no real signals for surfacing talent early in our society anymore.
universities are broken, useless, and corrupt.
YC lost its alpha - it produces pure beta at this point.
whoever solves this will be rich beyond measure
114
SaaS is like pizza now in that we're going to have a lot more of it. which means we need good, reusable open-source building blocks.
got tired of wiring up auth, billing, workspaces, and entitlements from scratch every time, so I'm building a template that handles all of it.
what's included:
- email/password OAuth sign-in
- workspace creation & switching
- role-based membership & invitations
- Stripe billing with plan entitlements
- slug-based multi-tenant routing
powered by:
- @better_auth
- @resend
- @stripe
2
21
3,382
someone’s gotta build the gist for agents - md idea files instant shareable with clean API
Apr 4
Wow, this tweet went very viral!
I wanted share a possibly slightly improved version of the tweet in an "idea file". The idea of the idea file is that in this era of LLM agents, there is less of a point/need of sharing the specific code/app, you just share the idea, then the other person's agent customizes & builds it for your specific needs.
So here's the idea in a gist format: gist.github.com/karpathy/442…
You can give this to your agent and it can build you your own LLM wiki and guide you on how to use it etc. It's intentionally kept a little bit abstract/vague because there are so many directions to take this in. And ofc, people can adjust the idea or contribute their own in the Discussion which is cool.
2
211
i don't understand this debate at all.
the comparison is entirely wrong. it makes no sense to rebuild something like slack in house, but it also makes no sense to pay 220k / year for it.
the obvious evolution here is that slack gets replaced by a smaller company that doesn't have to charge 220k / year
you don't want to have to deal with in house maintenance on every tool you use, but the cost of software is plummeting anyways
Mar 30
This is a great report on the state of software and AI by @Redpoint - thank you, @loganbartlett!
Where I disagree is the build vs. buy slide:
1) I'm not sure if it takes ~12 engineers to build/maintain a Slack clone for 1 customer. As AI keeps getting better at not only code gen but all software engineering tasks I think you'll be able to do it with a smaller team. Doesn't mean you should spend engineering time on it because I expect...
2) ... there will be agencies who specialize in this kind of work (e.g. build a Slack clone and sell customized versions of it).
3) ... there will be lots of cheap, (more or less) good enough Slack clones
4) ... there will be AI-native startups that rethink the category.
All of these factors, I think, will contribute to pricing pressure for Slack and other traditional SaaS companies ... which they will only be able to defend against if they get a share of the agentic revenue enabled by their products.
3
253
apparently @mercury offers the first, and now has personal accounts:
mercury.com/personal-banking
1
2
103