Data Science Collective is a Medium publication with articles on machine learning, AI agents, code analysis, and Python data tooling.
HN commenters mostly link their own articles there, such as an analysis predicting Apple's Kuzu acquisition and a piece arguing AI outputs all sound alike.
Over my ten-year tenure in Big Tech, I’ve witnessed conflicts that drove exceptional people out, hollowed out entire teams, and hardened rifts between massive organizations long after any business rationale, if there ever was one, had faded.
The conflicts I explore here are not about strategy, conflicts of interest, misaligned incentives, or structural failures. Nor are they about money, power, or other familiar human vices.
They are about identity. We shape and reinforce it over a lifetime. It becomes our strongest armor — and, just as often, our hardest cage.
My two previous reddits in the Tech Bro Saga series:
- Why Big Tech Turns Everything Into a Knife Fight - a noir-toned piece on how pressure, ambiguity, and internal competition turn routine decisions into zero-sum battles (https://medium.com/data-science-collective/why-big-tech-turns-everything-into-a-knife-fight-42e221944ec8)
- Big Tech Performance Review: How to Gaslight Employees at Scale - a sardonic look at why formal review systems often substitute process for real leadership and honest feedback (https://medium.com/data-science-collective/big-tech-performance-review-01fff2c5924d)
No prescriptions or grand theory. Just an attempt to give structure to a feeling many of us recognize but rarely articulate.
I wrote a long practical guide on image augmentation based on ~10 years of training computer vision models and ~7 years maintaining Albumentations.
Despite augmentation being used everywhere, most discussions are still very surface-level (“flip, rotate, color jitter”).
In this article I tried to go deeper and explain:
• The *two regimes of augmentation*:
– in-distribution augmentation (simulate real variation)
– out-of-distribution augmentation (regularization)
• Why *unrealistic augmentations can actually improve generalization*
• How augmentation relates to the *manifold hypothesis*
• When and why *Test-Time Augmentation (TTA)* helps
• Common *failure modes* (label corruption, over-augmentation)
• How to design a *baseline augmentation policy that actually works*
The guide is long but very practical — it includes concrete pipelines, examples, and debugging strategies.
Would love feedback from people working on real CV systems.
Link:
https://medium.com/data-science-collective/what-is-image-augmentation-4d31dcb3e1cc
In Feb this year, when I analyzed KuzuDB's source code, I predicted Apple's reasoning to buy them was to introduce Siri with cross-app personal context. "..WWDC 2026 or 2027 introduces any “contextual intelligence” features in Siri that require cross-app relationship reasoning." https://medium.com/data-science-collective/i-analyzed-163k-lines-of-kuzus-codebase-here-s-why-apple-wanted-it-12294a7035fa?sk=6edea2d92a8e00f05dc38a376833c665. There is no confirmation on which tech is being used to achive that though
Have a look at the diagram in the accompanying blog post https://medium.com/data-science-collective/pipask-know-what-youre-installing-before-it-s-too-late-2a6afce80987#7ae4 , it explains how the process works.
In short, you can get metadata from pyproject.toml, but (a) it can still involve executing code due to PEP 517 hooks, and (b) a malicious package would use the legacy setup.py to get their code executed.
Cool! I built a similar bot a year ago, mostly for adding subtitles for short videos within telegram, to follow news from around the world.
Blog post here:
https://medium.com/data-science-collective/breaking-language-barriers-how-i-built-a-video-translation-telegram-bot-in-one-hour-db739db429f4
Check it out here: https://t.me/ktuvitbot
They all are. This is proven in research. https://medium.com/data-science-collective/the-ai-hivemind-problem-why-all-ai-sounds-the-same-9273360376cd