Archive - pnote

Prompt injection is not about segmentation

🚧 Delimiters don't save us from pirate jokes; the separation of instructions and data doesn’t exist; LLMs are hard to steer

Sep 8 •

🗓️ Life happens one week at a time; pushing my life-tracking obsession a bit further; I made a new tool to create annotated life calendars

Sep 1 •

August 2024

The distractible LLM: pirate jokes may be the least of our problems

🏴‍☠️ We ask to review an email draft and hear a pirate joke instead; LLMs are distractible; a trusted prompt with untrusted data equals trouble

Aug 25 •

Connecting the dots: cross-referencing Steinbeck's diary with Atomic Habits using NotebookLM

✍️ Inspiration is a daughter of memory; cross-referencing is hard work; NotebookLM is my new staff librarian

Aug 18 •

"Yes, and..." is not the point: what I learned about improv in a corn field retreat

🌽 An improv about wedding planning escalates to absurdity; what to do instead of thinking what to say; learning life skills in a corn field theater

Aug 11 •

Think before: chain-of-thought prompting for creative brainstorming

🧩 Putting a viral reel to test; creative brainstorming vs math puzzles; LLMs need tokens to "think"

Aug 4 •

July 2024

We need a better copy-paste for the LLM era

🖼️ The system clipboard contains multitudes

Jul 28 •

20 times denser: making our own festival program page for Off Avignon

🔧 Feeling stuck in a low information-density infinite scroll? Fine, let's make our own and make it 20x denser

Jul 21 •

Theater curation: screening a festival program 484 pages thick

Off Avignon, where theater blossoms and the completionist spirits come to die; The infinite website scroll vs the useful beauty of a single JSON file…

Jul 14 •

The search for tiles: adventures in mapmaking with Stamen Toner, OSM and the German Federal Agency for Cartography

All I want is a tileset that's free to use and grayscale

Jul 7 •

June 2024

Everything all at once: how more compute power coupled with research breakthroughs enabled the explosion of generative AI

💫 Decoding the explosion of progress in generative AI; the adoption/innovation flywheel of neural networks; Yann LeCun was right after all

Jun 30 •

Going forward, going back: how all neural networks learn

💫 We open a neural network and see how it learns; we talk about the loss and backpropagation; Paris celebrates fête de la musique

Jun 23 •

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts