My 30 best Twitter threads

Covering from machine learning and CS theory to philosophy, writing, and more.

Not too long ago, there was a time when Twitter was the best place online to connect with like-minded people and learn from each other. I grew up, intellectually, in that environment, and into the now-forgotten tradition of sharing long threads on technical topics. That was my first encounter with technical writing beyond academia, and I loved it while it lasted. It was also where I made most of my online friends, who are now avid readers of this substack, as I’m of theirs.

So, yes, I’m feeling a bit nostalgic, partly because it seems those days are gone and may never return. Maybe a bit over my head as well. In any case, in celebration of the great old tech Twitter, it’s my pleasure to share with you what I consider my best technical threads. I’ve grouped them into categories for easier browsing, and sorted them more or less in the order I believe it’s most informative if you’re getting started with some of these topics.

A small caveat before moving on. These Twitter threads are, by their very nature, broad and superficial, in some sense the epitome of canned content. Please take them as nothing but a motivational intro to their respective topics, and don’t hesitate to dig further.

---

On Machine Learning

This list of short threads introduces the most relevant concepts in machine learning. They are meant to be read in order and cover the most basic ideas, from what is machine learning problem to different modeling paradigms and how to evaluate them.

1. What is Machine Learning?

2. Types of machine learning problems.

3. Parametric vs. Non-parametric models.

4. Generative vs. Discriminative models.

5. Evaluation metrics.

And these are four introductory threads to other relevant machine-learning topics. These don’t have a specific order.

• Language models (before GPT).

• Generative Adversarial Networks.

• Automated Machine Learning (AutoML).

• 10 practical AutoML tools (a bit outdated).

On Computer Science Theory

These are 5 related but separate threads on some fundamental aspects of Computer Science theory.

• Algorithmic complexity.

• NP-Completeness.

• The P vs. NP problem.

• The Halting Problem.

• Why recursion matters?

On the Philosophy of Computer Science

These are my favorite threads of all time. They touch on specific philosophical questions always from a computational perspective.

• What is intelligence?

• Are we living in a simulation?

• Can we answer all questions?

• How do you know that what you know is true?

• Is there an objective comparison metric?

• Should self-driving cars decide who to kill?

• Is superintelligence necessarily benevolent?

• What is experience?

• Who are you?

On Technical Writing (and other stuff)

And these are threads on practical matters for researchers and students, mostly focusing on technical writing but also touching on related topics.

• Presenting an idea with the Why-What-How framework.

• The Vision-Steps-News-Contribution framework for presenting ideas.

• Three lessons on the interaction between academia and the industry.

• Tips to improve your technical writing skills.

• Three health habits to survive grad school.

• Doing a literature review that doesn’t suck.

• What is there to gain from Twitter threads (yeah, meta!)

---

Many of these threads are initial drafts for longer posts I’ll write in the following months. If there is one in specific you’d like me to prioritize, please let me know in the comments. I hope you find them all useful, or at least as fun to read as it was for me to write them.