I tried opus, it did what I asked but not without turning the codebase into homework for me to study

This is my new benchmark for AI:

Is it better at the task than me with a gun to my head

If the answer is no then just implement it yourself

And don't ask it to implement things you don't understand yet, take the time to understand them and once you do ask that question again

It's such a shame that nobody is thinking in the 10 year time frame

Were are the parc vibes????

There has never been this much money.... but the timeframe is so small that nothing revolutionary ever gets built

Think in that timeframe

create an organization centered around the idea that a group of people will be tasked with bringing something new to the world in 10 years time!

Arguing with auto complete 🙄

just look at that agency

If I can't see or feel it, then it must be some kind of lie

It's funny

Every single successful person only had to be right just that one time. After that one time the past failures didn't matter and the future was set in stone

We only talk about their one time

compression != intelligence

AI === AGI

Today I'm gonna build an audiobook generator.

#friday

Progress so far
- ui/ux
- load epub
- generate audio with selection (using inworld.ai)
- show audio waveforms

Todo
- let user control voice generation settings
- figure out snippet pagination
- add backend

Spent the last couple hours trying to come up with the perfect UI that just makes everything simple. I tried using a canvas timeline but still it just don't feel right so now back to the drawing board. I am gonna call it a day for now.

The real problem with LLMs is that they don't force you to think simple enough.

Historically, teams were forced to find simple, elegant abstractions because anything overly complex was too expensive to build, maintain, or even reason about. Effort acted as a natural filter: bad ideas died early because they were painful to implement.

LLMs make complexity cheap. Which is a very bad thing.

Effort used to serve as an incentive for simplicity.

Take notion. They came up with a simple block based data model because the other methods required more effort. They were forced in to a corner to think.

LLMs can rob us of this moment of reflection

Hmmm eleven labs got a really well designed audiobook studio

Got rid of the clankers

The exciting thing about real AI is that its truly universal. There are no barriers. Neither compute nor data.

We will be able to teach an AI new languages with very little data.

When you think about it, We can learn a new language from just one book... just by talking to ourselves. Now scale that on a computer...

its fast 🚀

Dealing with all the epub parsing libraries on the frontend was a drag, so I made epub-wasm

I give it epub, it gives me json structure I want

Github | NPM

It's about 800 lines of code in rust.

I could've implemented this with javanoscript... but it just didn't feel right to have chunky code on the frontend.

With epub-wasm I never have to see it again, plus its much faster and I can run it anywhere I want.

I truly never have to look at the code again 🤌

Neurons that fire together wire together

types.ts is the most important part of any project