These are experiments. They're not products, and they're definitely not businesses. They're prototypes that convey an idea or explore something we're curious about.
WHAT
Generative AI is everywhere—except in most businesses. In the Census Bureau's Business Trends and Outlook Survey, a bi-weekly study with 1.2M participants, only 5% of respondents say they have used AI in the last two weeks.
It's not that most businesses are run by Luddites. It’s that we, the AI developers, haven’t made truly useful tools for businesses. We build chatbots that can write personalized horoscopes for house plants, but we overlook the less flashy, yet potentially more valuable, capabilities of LLMs: classic NLP tasks like summarization, classification, and entity extraction, which previously required specialized models but are now accessible to every developer with an API key.
GreenTape.us is a project that explores how those classic NLP tasks can be used to build something useful for businesses. TL;DR: The U.S. federal government sets aside contract opportunities under $250K for small businesses, but these listings can be hard to find. GreenTape.us tries to help.
WHAT
GPT vs. Gemini is a demo that pits OpenAI’s generative AI model, GPT, against Google’s model, Gemini, in a 10-question game of trivia.
The demo explores how LLMs themselves can be used for automated evals. The questions and assessments are all AI-generated.
It's LLMs grading answers given by LLMs to questions written by LLMs. Lots of caveats apply (read the FAQ!), but I had fun building it.
WHAT
In 1961 — after working on the Manhattan Project and attending the Atoms for Peace conference, after spending a sabbatical in Brazil, and after an ex-girlfriend absconded with his Albert Einstein medal — Richard Feynman took over the introductory physics course at Caltech.
He began his first lecture with the following question: "If, in some cataclysm, all of scientific knowledge were to be destroyed, and only one sentence passed on to the next generation of creatures, what statement would contain the most information in the fewest words?"
His own answer was this: "All things are made of atoms — little particles that move around in perpetual motion, attracting each other when they are a little distance apart, but repelling upon being squeezed into one another."
This site collects the answers of 152 other people, curated from a dataset with many more.
WHAT
A newsletter and Instagram account about the history of technology and the humans who use it. The project went from 0 to 1.5k followers in 3 months.