How will AI be good for me?
AI is getting good. Every new model release is another step up the benchmarks. It seems like every business - not just buzzy startups - is trying to figure out how to use it.
The chatbot setup is pretty helpful. ChatGPT and Claude are “the new Google” in that I turn to them for basic questions about the world, and they do a good job of that.
But AI is really good at lots more stuff, not just general knowledge about the world. And, compared to how much we’re investing in it, it’s only marginally better than old Google for those kinds of questions.
Making a whole reasoning AI that can only do what Google did 15 years ago is… weird overkill? A huge waste? It can be zo much more.
It’s really good at coding. It can type so fast. A lot of software eng is about knowing thousands of bits of trivia about how computers and software works, and these LLMs can fit so much trivia in their weird exobrains. There’s tons of ways they are still bad at software dev, but they’re getting really good. I had a ton of fun writing code before AI, and I have more fun now, with AI.
But… most people don’t code.
Coding, programming, hosting software, and making apps is cool and I love it. But most people don’t want to spend their days doing it. The computer should work for you.
Computers are really powerful and growing more powerful all the time. But most of the way we use computers in our lives is kinda crappy. They’re little tvs, or we play games. The dream of the computer, “a bicycle for the mind” has mostly been lost to brainrot and email.
It will be a big loss for humanity if great AI comes along and all it does for most people is make brainrot and email better. I know there’s lots of software people and businessers who are getting a lot of enterprise value from the business use cases. Generating code quickly, scanning vulnerabilities, writing investment analysis reports, yea yea yea.
But that’s still not the world I want. I want computers to help us live better, be better friends, happier, healthier. I don’t want “apps”. I don’t want passwords and logins and 2fa and unsubscribe from marketing drip emails forever.
I want my computer to be MINE. And I want it to be fun and cool and make me a better friend and husband and brother. I want AI to be good for me, to make me smarter and funnier and cooler and better looking.
All the efforts on alignment are sort of about this. Making Claude good and wise and a friend is a start! But for AI to make a real difference in how we live and how we relate to computers, we need the personal computing revolution to really happen.
—
I’ve spent a lot of time teaching software. It’s a nice path to a nice job for a lot of awesome people who want to get into it. But software’s also something that I love, and, in teaching, I get to share that love. Writing your own software feels sooo good.
When my career pendulum arced back into building software, natural questions are “what to build” and “how does it connect” and “what’s the throughline connecting these”. Making people good at using computers powerfully and making computers good at being used powerfully are how I’ve been explaining the two sides of the coin to myself.
All that’s to say: I recently joined Zo Computer. We’re making a personal computer in the cloud, where AI helps you to make software for yourself. Right now, it has files and SMS and email, and you can build and host software with it (even if you don’t really know what those things mean, at the start). We’re adding more all the time.