I didn’t have STEM background.
I didn’t know what pip install was.
Heck, I didn’t speak a word of English.
I grew up with my grandma in the rural part of Kyrgyzstan.
My first job was selling potatoes and onions on a sidewalk at a local bazar in Bishkek.
Now, I live in San Francisco.
Every day, I meet the smartest and the most Influential people in Tech.
My self-doubt and imposter syndrome have never been this high.
99% of the time, I am the dumbest person in the room.
This slowly was getting under my skin.
It was making me feel small and insignificant.
I wouldn’t ask questions because I didn’t want to sound foolish. This made me stuck in one place.
Until I realized, if I keep being quiet, I won’t ever grow and learn.
So now, the first thing I do when I wake up is leetcode problems.
I go to deeply technical meetups.
I make friends with excellent engineers and scientists.
And I am finally okay with looking like a fool and learning as much as I can. As fearlessly as I can.
This is a reminder, to never let your environment dictate how you should feel about yourself.
Instead use it as a source of inspiration to propel you forward.
#leetcode #adhd #datascience #motivation #inspiration #impostersyndrome
I just released Beads, a drop-in cognitive upgrade for your coding agent of choice. https://t.co/d94hykIk7y
In a nutshell, it is a magical 4-dimensional graph-based git-backed fairy-dusted issue-tracker database, designed to let coding agents track all your work and never get lost again. Beads replaces that disgusting pile of half-eaten markdown files in your plans/ directory. You know you have one.
Install the Beads binary, tell your agent in https://t.co/q9aIg3VyQZ to stop using Markdown and run `bd quickstart`, and your agents will spontaneously get better at everything, particularly long-horizon planning and keeping track of newly discovered work.
If you are looking for a long-ass but occasionally funny article about how I discovered it while burning a 350k line code base to the ground: https://t.co/DQ0ZZIJhRT
Or you can, you know, just start using it. It's MIT-licensed, written in Go, and super lightweight and portable. Let me know how you like it.
Enjoy!
@benhylak Ben, dude. This was an insightful read.
Evals are becoming so meta.
You mentioned you have trained models that “pluck” those edge cases.
Can you elaborate more?
And how can we define what is “bad” to look out for?
Claude Code: no evals
[well known code agent company]: no evals
[well known code agent company 2]: kinda halfassed evals
[leading vibe coding company]: no evals
[ceo of company selling you evals]: mmmmm yess all my top customers do evals, you should do evals
[vc's in love with ceo of evals company]: mmmmm yes all my top founders do evals, must do evals
(NOTE: i -do- also think that evals are impt, but the eval pilled ai engineers have also noticed that it is not a strict requirement for success and, at least for 0-to-1 stage, may even be anticorrelated, think thru why)
- create custom evals
- implement guardrails
- have a human in a loop
- write tests before code
- ditch Streamlit, learn Typescript
And you are golden, baby.
#AIengineer
Please, stop building RAG projects!
If you want to succeed as an AI engineer, focus on this instead:
- solve a real problem
- build a simple agent
- scrape the web
- retrieve your data
- call another API