Skip to content

Instantly share code, notes, and snippets.

View vunderkind's full-sized avatar
🤖
Becoming turing-complete

mogwai. vunderkind

🤖
Becoming turing-complete
View GitHub Profile
@vunderkind
vunderkind / gist:f74a3dc681dd37edacdd635a021826c9
Created December 21, 2024 15:10
0 to 100: ML Engineer in One Year
If you're reading this, I shared it with you — directly or indirectly — which suggests you're in the general space of people studying to become ML engineers as quickly as possible.
My earliest observation is people accidentally sabotage themselves by thinking it will be too easy, or too hard.
People who think it'll be easy plan to make this transition in 3 months. Six, for some. I recommend a year of conscientious grinding, especially if — like me — your math is completely nonexistent.
If I had to streamline the path, I'd say you need to:
1. Pick up a framework (like Tensorflow or Keras) and get really good at it. This is probably more important than math, when you're starting out.
@vunderkind
vunderkind / gist:ae0096b8391b5cb9337847a72d288db0
Last active September 16, 2024 10:58
Notes from OpenAI's o1 model on endurance as an attribute of intelligence
My query:
Thoughts on conversation?
@xylemic do iq tests test intelligence or mental fatigue? (or does it accidentally test mental fatigue? how many people are just selecting random options after the first n questions?)
@alephile Either way this ends up meaning this is what intelligence is: can you endure problem solving sprints?
@xylemic is that what intelligence is? can a stupid person not endure problem solving sprints? (ie, working at 100% mental capacity to get all problems wrong, without ever breaking a sweat?)
---------
@vunderkind
vunderkind / machine.js
Last active June 30, 2023 20:02
Generated by XState Viz: https://xstate.js.org/viz
async function invokeFetch(context) {
console.log(context)
const { documentUrl } = context;
return await fetch(`https://google.com`)
}
const DocMachine = Machine({
id: 'doc-chat',
let company = { // the same object, compressed for brevity
sales: [{name: 'John', salary: 1000}, {name: 'Alice', salary: 1600 }],
development: {
sites: [{name: 'Peter', salary: 2000}, {name: 'Alex', salary: 1800 }],
internals: [{name: 'Jack', salary: 1300}]