Thanks for the shout-out! I'm glad you liked the post :D
25.02.2026 09:45
👍 1
🔁 0
💬 0
📌 0
Thanks for the shout-out! I'm glad you liked the post :D
I'm almost finished my new book, titled "50 ML projects to understand LLMs".
I call this figure the "mitten plot."
How to live a happy life, according to the most elegant equation in humanity. (Hint: it's the math version of "omit needless words.")
mikexcohen.substack.com/p/eulers-ide...
Thanks :) Your paper is great! But more sophisticated than what I'll show in part 2. I'll show how to manipulate embeddings vectors as they pass through the residual stream to normalize the bias. So it's not actually fixing the bias, it's correcting one example. And learning about mech interp.
In my latest substack post, I show how to measure gender bias in open-source LLMs. The post comes with detailed explanations and a Python notebook file that you can easily adapt to any other biases.
Curious? Come check it out :)
mikexcohen.substack.com/p/gender-bia...
And thank you, Paolo!
No comment.
ps, if you liked that joke, then you will absolutely love my recordings of the two-day workshop I taught on EEG synchronization in time, frequency, and space.
Here's part 1: mikexcohen.substack.com/p/eeg-synchr...
Statistician: Someone who has their head in an oven, their feet in an ice bucket, and says "on average I feel fine."
If I had a million euros (that's 1.170686 million USD), I would do Daily Data Challenges to master machine-learning. Like this one, which involves running a t-test in Python to compare two random samples.
Act like a millionaire. Solve Daily Data Challenges.
mikexcohen.substack.com/p/ddc-15-two...
“King – man + woman = queen” sounds smart until you actually check the math.
Spoiler: it’s fake news.
mikexcohen.substack.com/p/king-man-woman-queen-is-fake-news
I don't know what's going on in this AI image.
But I do know this: Daily Data Challenge #25 is short and very practical for data science. You'll get to create a symmetric matrix and extract its unique elements.
Can you solve it in less than 6 minutes?
mikexcohen.substack.com/p/ddc-25-uni...
Awesome.
Here’s a picture of Thalorion The Destroyer reading my calculus textbook. I hope he likes it! Otherwise he might unleash doom on humanity.
www.amazon.com/dp/B0DX6BPPZ8
Interesting perspective. I think peer-review is imperfect but does more good than harm. I've published >100 papers and reviewed... dunno, many 100s of manuscripts. My thinking is that the incentive structure is the root of the problem, although I could be wrong. I agree with your suggestions though.
Fair point. I suppose the question is whether the entire system needs to be revamped (I'd find that unlikely to happen, given that too many people are too vested in it) or whether relatively small changes can have meaningful impact.
I'd love to see the evaluations in academia based on one paper every five years. Not a perfect solution, but it would increase scientific publishing signal-to-noise-ratio by encouraging researchers to put more effort into a smaller number of papers.
Hehe, yep, although I'm not so active here. I'm mostly building up content (tutorial posts on data science and LLM mechanisms) on mikexcohen.substack.com.
Nice to hear from you, Luiz!
That was definitely part of my decision, although I wouldn't call it a "problem." More that I was (and still am) in an extremely fortunate situation, and decided to spend my time in a career that had higher positive impact for more people, and that was also more personally satisfying.
Thank you for taking the time to share you lucid and honest opinion.
Thank you kindly, Thomas. My feelings, reactions, and motivations are definitely not unique, but ex-academics have an unfortunate habit of reticence.
Karl Gauss (yeah, *that* Gauss) sent me a fax this morning to say that he really enjoyed working through today's Daily Data Challenge.
mikexcohen.substack.com/p/ddc-23-dis...
You earned them.
That is a fantastic title and topic. Congrats!
Why I left academia and neuroscience.
This post on Substack has gained a lot of traction. I think many people identify with it.
(Most of my posts are technical tutorials on machine-learning and LLM-mechanisms.)
mikexcohen.substack.com/p/why-i-left...
Thank you, Felipe :)
Daily Data Challenge 8: Test for Pythagorean triplets
A data challenge a day helps you master machine learning.
More exercises and solutions at mikexcohen.substack.com
youtube.com/shorts/LlqS2...
Bootstrapping: because sometimes asking “what if?” 10,000 times is a valid math strategy.
open.substack.com/pub/mikexcoh...
genAI never fails to deliver the laughs.
