- Querying 3 billion vectors
- 2025 in review
- I want to see the claw
- Walking around the app
- How big are our embeddings now and why?
- Enabling Hugo static site search with Lunr.js
- My favorite use-case for AI is writing logs
- 20 years of YC
- You can just hack on ATProto
- How FastAPI path operations work
- Everything I did in 2024
- Write code with your Alphabet Radio on
- Why are we using LLMs as calculators?
- Dead Internet Souls
- Don't worry about LLMs
- We've been put in the vibe space
- How I search in 2024
- Redis is forked
- Both pyramids are white
- GGUF, the long way around
- What's new with ML in production
- Retro on Viberary
- Why if TYPE_CHECKING?
- Build and keep your context window
- What we don't talk about when we talk about building AI apps
- Naming things
- What should you use ChatGPT for?
- Welcome to the jungle, we got fun and frames
- What I did in 2022
- Everything I learned about accidentally running a successful tech conference
- The cloudy layers of modern-day programming
- Some notes on the Stable Diffusion safety filter
- How I learn machine learning
- On the team as a system
- How to prepare an AWS test image for PyTorch
- Looking back at two years at Automattic and Tumblr
- On owning a software problem
- Git, SQL, CLI
- Migrating to Hugo
- 2021 Work Recap, or the conjoined triangles of success
- The programmer's brain in the lands of exploration and production
- Recsys 2021 Recap
- Doing small, fun projects
- Reaching MLE (machine learning enlightenment)
- The local minima of suckiness
- Writing for distributed teams
- The ritual of the deploy
- The humble hash aggregate
- The ghosts in the data
- Getting machine learning to production
- Securely storing configuration credentials in a Jupyter Notebook
- Lauching a Jupyter notebook from within a virtual environment
- How to make sure your question gets answered online
- A deep dive on Python type hints
- IT runs on Java 8
- Breaking down Apple's iOS Screen Time report
- The attic and the basement of apps
- Data science is different now
- Commit to your lock-in
- The Case of the Broken Lambda
- On Excel and developer productivity
- Good small datasets
- One month with the 2017 MacBook Pro
- It's still hard for beginners to get started with Python
- Building a Twitter art bot with Python, AWS, and socialist realism art
- Working with AWS
- Data ex machina
- Recent podcast and radio appearences
- Alice in Python projectland
- Should Spark have an API for R?
- So, you want to data science
- Setting up SQL for beginners is hard
- Should you replace Hadoop with your laptop?
- Being a woman in programming in the Soviet Union
- What should you think about when using Facebook?
- Strata 2016 and the rise of New Data
- How to parse emails with Python
- Walkthrough of Jupyter Lab Alpha
- My recent experiences using Azure ML
- Why doesn't MySQL have window functions?
- Data strategy in 2016
- How to keep up with the data industry today?
- Building presentations the hard way with reveal.js and Jupyter
- Learning tech by analogy
- Content is dead. Long live filler!
- Starting to dig inside Spark RDDs
- Let's work on the hard stuff
- How non-technical teams work together
- Generating fake Hacker News headlines with Python
- How large should your sample size be?
- Who is doing this to my internet?
- Cumulative totals three different ways in data science
- How many people are actually using your stuff?
- Really easy data concepts that get abused all the time
- Creating readable SQL.
- How to get into big data
- I'm not so sure about Tufte
- Man, do static sites suck.