I don't care if you use DSPy, I care that you build AI software properly, especially software that rides the wave of AI progress and doesn't fall prey to the bitter lesson. If you understand the five principles below (or some simplified version of them; WIP), you will build good

Make art so epic it will raise the birthrate.

*taps the sign*

learning about the streisand effect the hard way

i listen to so much on 2x it's a strong signal of its value, we should have ai just send us a short summary

george hotz tried to fix twitter search and decided it would be easier to take down nvidia

i don't think i'd have gone to college if we had our level ai then. i enjoyed college but didn't need it to learn cs i was already pretty good

webllm running dspy in the browser. ax has come far i'm excited to build more, axflows, optimization, etc. models like gemma 3n make edge ai a very real thing.

google jules is back on the task.

google search ai is my goto for 90% of my questions.

The DSPy community in Toronto is getting together tomorrow at New Stadium! Come by for a remote workshop with Maxime Rivest 🧙‍♂️🦙🐧, spacy and hang with local DSPy experts like Prashanth Rao and Robbie Pasquale. An evening of building elegant AI systems and good pizza.

very impressive but also interesting that our photos of the past will contain predicted elements

Looks like LLMs can learn by themselves if you just check whether updates to their knowledge text reduces surprise

Looks like Toronto's "PROGRAM11: Advanced DSPy" is now out on YouTube. Featuring Maxime Rivest 🧙‍♂️🦙🐧 and spacy. As well as Prashanth Rao and Robbie Pasquale! Link below. cc Tarun Sachdeva

Program 11: Advanced DSPY at New Stadium

just use grok in x it has none of the weird stuff

killer idea i'd switch in an instant

uv might bring me to do more python