Top all-time · #4187
Piotr Migdał
Personal blog · p.migdal.pl · on HN as staredaidatamath
Piotr Migdał's personal blog at p.migdal.pl covers ai, data and math. Since 2015, 18 different Hacker News readers have linked it 197 times, most often “Learning Deep Learning with Keras – An Overview for Beginners”. It is still writing: the last post was on 2026-10-07, and you can follow it by feed.
Too few sampled comments (2) to say how often HN recommends it.?
Why HN links it
Posts on visual exploration of vector embeddings and a 'Dating for Nerds' guide covering gender differences. The comments mostly reference them in passing rather than judging the writing.
Summarised by Claude from a sample of 2 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Piotr Migdał - blog posts and stuff
Most-linked posts?
- Learning Deep Learning with Keras – An Overview for Beginners3 people · 11 mentions · first linked 2017-04-30
- What I do or: science to data science3 people · 11 mentions · first linked 2015-12-15
- I Learned to Stop Worrying and Love the Types and Tests2 people · 11 mentions · first linked 2020-03-02
- Data science intro for math/phys background2 people · 10 mentions · first linked 2016-03-16
- Quantum mechanics for high-school students2 people · 10 mentions · first linked 2016-08-18
- Show HN: Interactive Machine Learning, Deep Learning and Statistics Websites2 people · 8 mentions · first linked 2018-08-12
- Dating for nerds (part 2): gender differences2 people · 2 mentions · first linked 2019-04-08
- Don't use cosine similarity carelessly2 people · 2 mentions · first linked 2025-01-14
What Hacker News says?
http://p.migdal.pl/2017/04/30/teaching-deep-learning.html -> "What mathematicians thing I do" (Full disclaimer - I did theoretical physics, so understand both sides. :))
Comment linking /2017/04/30/teaching-deep-learning.html
on “Machine Learning Crash Course”Submitted Don't use cosine similarity carelessly as a story · 439 points
HN discussionSadly, innovation in teaching is neither a priority in academia (since it is not research) nor in the industry (since it is this kind of activity that ). I wouldn't say there is no innovation though. There are science-based games ( https://github.com/stared/science-based-games-list; it was here some time ago: https://news.ycombinator.com/item?id=14661813 ). I started collecting interactive machine learning, stats an…
Comment linking /interactive-machine-learning-list/
on “Immersive Linear Algebra (2016)”Submitted If it is worth keeping, save it in Markdown as a story · 338 points
HN discussionI wrote a blog post exactly on that: http://p.migdal.pl/2016/03/15/data-science-intro-for-math-ph... (including the data science part). I strongly advice for: - using Python in the interactive environment Jupyter Notebook, - starting with classical machine learning (scikit-learn), NOT from deep learning; first learn logistic regression (a prerequisite for any neural network), kNN, PCA, Random Forest, t-SNE; concepts…
Comment linking /2016/03/15/data-science-intro-for-math-phys-background.html
on “Ask HN: How to get started with machine learning?”Submitted King – man + woman is queen; but why? as a story · 263 points
HN discussionI created a list of Interactive Machine Learning, Deep Learning, and Statistics websites ( https://p.migdal.pl/interactive-machine-learning-list/ ), which includes Immersive Linear Algebra, among other things. For a broader range of subjects, there is a list of Explorables ( https://explorabl.es/ ).
Comment linking /interactive-machine-learning-list/
on “Immersive Linear Algebra (2015)”Submitted Learning Deep Learning with Keras – An Overview for Beginners as a story · 74 points
HN discussionTerm 'Artificial Intelligence' is vague. "Deep learning is a name for machine learning techniques using many-layered artificial neural networks. Occasionally people use the term artificial intelligence, but unless you want to sound sci-fi, it is reserved for problems that are currently considered “too hard for machines” - a frontier that keeps moving rapidly." from http://p.migdal.pl/2017/04/30/teaching-deep-learnin…
Comment linking /2017/04/30/teaching-deep-learning.html
on “AI Playbook”Submitted Show HN: Weltschmerz by age as a story · 72 points
HN discussionOn a lighter tone, here are some lessons I've learnt at !!con (a conference by Recurse Center aka Hacker School alumni): http://p.migdal.pl/2017/08/14/bangbangcon.html tl;dr: you can get great inclusivity primarily by being welcoming, accommodating for needs (in an open-ended way, not restricting it to a few axes) and, crucially, by not shaming or bulling otherwise good-willed transgressors Sure, not all of these th…
Comment linking /2017/08/14/bangbangcon.html
on “I’m an Ex-Google Woman Tech Leader and I’m Sick of Our Approach to Diversity”Submitted Science-based games and explorable explanations as a story · 68 points
HN discussionIn 2018, I co-wrote a blog post with the inflammatory title “Don’t use TensorFlow, try PyTorch instead” ( https://news.ycombinator.com/item?id=17415321 ). As it gained traction here, it was changed to “Keras vs PyTorch” (some edgy things that work for a private blog are not good for a corporate one). Yet the initial title stuck, and you can see it resonated well with the crowd. TensorFlow (while a huge step on top o…
Comment linking /blog/2017/04/teaching-deep-learning/
on “Leaving Meta and PyTorch”Submitted Dating for nerds (part 2): gender differences as a story · 42 points
HN discussionWell, there are languages better at incorporating the Zen of Python than Python itself, vide (full disclosure: my blog post) https://p.migdal.pl/blog/2020/03/types-test-typescript/ . When it comes to package managers, as you noted, the situation is even more ironic, as nicely depicted in https://xkcd.com/1987/ .
Comment linking /blog/2020/03/types-test-typescript/
on “Uv: Python packaging in Rust”Submitted Dating for nerds (part 1): problem diagnosis as a story · 31 points
HN discussionOn conference deadlines and application times, there are some plots: https://p.migdal.pl/2017/11/15/after-pydata-warsaw-2017.html... It is interesting when there are a few deadlines (e.g. early bird, regular, last call).
Comment linking /2017/11/15/after-pydata-warsaw-2017.html
on “Suspicious Discontinuities”Submitted AI won’t make artists redundant – thanks to information theory as a story · 17 points
HN discussionNow I don't track offers (I get contracts through recommendations/networking), so I may be not up-to date. My background is different (PhD in quantum physics), so for me stats/data/ML is simple, but software architecture, algorithms - not as much. When 3 years ago I was looking for data science internships, most of interview were strictly in software engineering. (I got into a more data-analysis oriented.) Even when…
Comment linking /2015/12/14/sci-to-data-sci.html
on “Google's Free Deep Learning Course”Submitted Data science intro for math/phys background as a story · 13 points
HN discussionFor webcomics and ML, see this footnote: http://p.migdal.pl/2017/04/30/teaching-deep-learning.html#fn... "It made a few episodes of webcomics obsolete: xkcd: Tasks (totally, by Park or Bird?), xkcd: Game AI) (partially, by AlphaGo), PHD Comics: If TV Science was more like REAL Science (not exactly, but still it’s cool, by LapSRN)"
Comment linking /2017/04/30/teaching-deep-learning.html
on “Machine Learning”Submitted I Learned to Stop Worrying and Love the Types and Tests as a story · 7 points
HN discussionHow to know if an article would fit there? For example, I was thinking about adjusting my http://p.migdal.pl/2017/01/06/king-man-woman-queen-why.html (already with some interactive components) or writing about RoI polling ( https://deepsense.io/region-of-interest-pooling-explained/ - by my colleague, but more interactive). Would it be on-topic? (After changing style accordingly.)
Comment linking /2017/01/06/king-man-woman-queen-why.html
on “Why Momentum Works”Submitted Show HN: Interactive Machine Learning, Deep Learning and Statistics Websites as a story · 7 points
HN discussionAs a side note, I have a half-finished project of turning such city-month temperature maps into vectors (with an interactive matrix factorization/PCA). Source: https://github.com/stared/matrix-decomposition-viz Live: http://p.migdal.pl/matrix-decomposition-viz/
Comment linking /matrix-decomposition-viz/
on “The Well-Tempered Traveler”Submitted Data science intro for math/phys background as a story · 6 points
HN discussion...until it fails. I am a person who's perseverance towards my life goals caused depression, anxiety, and panic disorders. And almost my life. I wanted to become a top scientist. Yet, the academic path is full of disillusionment (mainly with the system). Perseverance gave me the power to go through sacrifices one after another and to carry on. However, only after I let it go (and turning to data science / deep learn…
Comment linking /2015/12/14/sci-to-data-sci.html
on “Perseverance Toward Life Goals Can Fend Off Depression, Anxiety, Panic Disorders”Submitted What I do or: science to data science as a story · 6 points
HN discussionDeep learning consulting ("I teach people to teach machines"). Finished my PhD in quantum optics in 2014, but immediately moved to data science. Why not physics? Full version: http://p.migdal.pl/2015/12/14/sci-to-data-sci.html tl:dr: I wanted a fast-pasted field with more freedom. Physics is now stale (no fundamental changes in the last decades, compared to each year in deep learning; cf. physics in 1900-1920), and…
Comment linking /2015/12/14/sci-to-data-sci.html
on “Ask HN: Physicists of HN, what are you working on these days?”Submitted If it is worth keeping, save it in Markdown as a story · 5 points
HN discussion
