#tools

https://twitter.com/karpathy/status/1761467904737067456?s=61&t=yGoUuKC9LsNGJxSAKv3Ubg

#career

Her manager was like: sure, just resign.

https://youtu.be/OMI4Wu9wnY0?si=SmlrT4JXkc5cbnX2

#misc

What?

#misc

https://www.reddit.com/r/ImTheMainCharacter/s/T7hAQiqxym

#dev

https://substack.com/redirect/4af1baf3-a4d7-48d1-9592-5ea0971a056a?j=eyJ1IjoiNHowa3gifQ.QnwKDJ1CRSD1ToSPhPzIWMi45g-Rid7OgDj8cqSear0

Sounds good but please DO NOT USE upper cases. It is not only about PEP8 but more about consistency and cognitive load.

We solve this problem by writing down the dimensions in the docstrings and also include the math expressions there. But it is already obvious that writing down the dimensions in the var names makes things much easier.

#ai

Some interesting perspectives on ai and arts

https://www.reddit.com/r/ask/s/t4tLIXqoU5

#fun

This is so cool. Would love to see some results for choice overload in foraging.

https://podcasts.google.com/feed/aHR0cHM6Ly9kYXRhc2tlcHRpYy5saWJzeW4uY29tL3Jzcw/episode/OTViNjcxYWUtMWI4Ny00NDVjLWJmNGEtN2FiODkyNzQ0ZjQ2?ep=14

#data

Never tried dbt but it is definitely popular judging by the amount of people talking about.

I read these discussions on Reddit and I think it is worth sharing.

https://www.reddit.com/r/dataengineering/s/tphxT7p0kI

and

https://www.reddit.com/r/dataengineering/s/YpOXP3y6av

#ai

Scary and exciting at the same time

https://www.figure.ai/

#tool

I liked Fig a lot. It was mostly an install and forget extension. Now Amazon bought the company, and it is part of the infamous codewhisper.

Do you have an alternative for Fig? I don't like Warp because it implements a lot of distractions.

https://fig.io/

#stats

#TIL# I learned about this interesting concept called martingale.

> ... at a particular time, the conditional expectation of the next value in the sequence is equal to the present value, regardless of all prior values.

https://en.wikipedia.org/wiki/Martingale_(probability_theory)

#dl

This repo is really nice.

yuanchenyang/smalldiffusion: Simple and readable code for training and sampling from diffusion models
https://github.com/yuanchenyang/smalldiffusion

#funny

What is wrong with you guys. Neovim is just an editor.

Edit: oh this is probably just chatgpt.

https://www.reddit.com/r/neovim/s/mfq9lZuDsq

#misc

Bayer sold HIV contaminated blood products (while knowing that the products are contaminated) to Asia and LatAm and infected patients....

https://www.tandfonline.com/doi/full/10.1080/08989621.2014.882780

#tools

https://commitizen-tools.github.io/commitizen/

A friend showed me how to use this tool called commitizen to manage commits and releases.

It took me a while to realize that this is a great tool. I haven't exhausted all commands in it but I already feel its power.

I usually don't trust human beings because we all make mistakes. I trust processes because they can miminize all kinds of risks quite fast.
commitizen is a tool that sets up a process. With that, each individual can just follow through. Less manual decisions makes the release process more robust.

Highly recommended! Try it out.

#python

I had the wrong idea for a long time that IDEs treat method/function without return type hint as returning None type.
I was wrong. In PEP484, it says IDE should treat such a method/function as a type that is as general as possible. Ah that just makes sense.

https://peps.python.org/pep-0484/#the-meaning-of-annotations

#data

Interesting read

> we propose DATALORE, a framework that explains data changes between an initial dataset and its augmented version to improves traceability

https://www.amazon.science/publications/datalore-can-a-large-language-model-find-all-lost-scrolls-in-a-data-repository