Sitemap - 2024 - pnote
Hash Code: AI from DeepMind beats top human programmers
Record yourself: working on my spoken French with voice recognition AI
Rest in pieces: end-to-end vs modular deep learning
Prompt injection is not about segmentation
The distractible LLM: prompt injection explained
Connecting the dots: cross-referencing Steinbeck's diary with Atomic Habits using NotebookLM
"Yes, and..." is not the point: what I learned about improv in a corn field retreat
We need a better copy-paste for the LLM era
20 times denser: making my own festival program page for Off Avignon
Theater curation: screening a festival program 484 pages thick
Watch it learn: a neural network training on Shakespeare
Speak to me in Shakespeare: building a tiny language model
One year of writing every week
No more forms: LLM interface for calendar events
FTBL is not football: decoding Invader locations
Off the rails: "Talk to me, human" and the scriptless dialogue
Less can be more: squeezing more points from the Titanic passenger data
Top of the cake: a neural net for Titanic predictions in under 1 hour
Gatsby in the haystack: LLM Inception
Tuned for safety: climbing Mt Fuji in winter
Future, tense: the downside of a dopaminergic mind
The simplest ML model: reading my own handwriting
Basic ideas are forever, everything else is tweaks
Seeing like a neural network: chihuahua or blueberry muffin
Seeing like a machine: finding the edge of Space Invader
Hidden features: how movie recommendation systems work
Simple lasts longer: making a map stateful using scotch tape and Web Storage API