Michael Fitzpatrick

StoryReel, Previously at Apple, PullString, ConnectSolutions, Adobe, Macromedia, Vignette

[Michael Fitzpatrick @mitrick2]

Of course, the longest jams are still spread across 1.0-3.0 eras, but I’m not going to get into that with the LLM today

[Michael Fitzpatrick @mitrick2]

As recent @phish jams increase in length (eg Simple, Tweezer), I wondered if there’s a relationship between the number of pedals @treyanastasio is using and jam length. A simple linear relationship is too narrow, so I turned to @AnthropicAI to discuss:

[Michael Fitzpatrick @mitrick2]

To determine the best model, we would need to:

1. Collect data on jam lengths and pedal numbers across multiple tours
2. Plot this data to visualize the relationship
3. Test different models and use statistical methods (like R-squared or AIC) to determine which fits best

[Michael Fitzpatrick @mitrick2]

LLM generations are becoming faster than some of our Postgres queries…@GroqInc

[Michael Fitzpatrick @mitrick2]

@EHuanglu Mind sharing your workflow file? Nice work!!!

[Michael Fitzpatrick @mitrick2]

Llama in a maze finding a monolith

@phish @runwayml

[Michael Fitzpatrick @mitrick2]

@zzyzx That’s more Buddy Guy than Jerry

[Michael Fitzpatrick @mitrick2]

@drewhouston or Dropbox eng: plz consider adding option to OCR and automatically name newly scanned docs within Dropbox iOS app and I’ll spend more $.

[Michael Fitzpatrick @mitrick2]

Rivian R2 LoRA for Stable Diffusion XL for any other #rivian fans out there:

https://t.co/TIwZsr9HV0

[Mason Riney @caliking49er17]

@TheSFNiners @AdamSchefter 2024/2025 Third downs just got a lot easier. 🔥🔥🔥

[Michael Fitzpatrick @mitrick2]

@RootsBelow Rutherford DaBrave as well

[Michael Fitzpatrick @mitrick2]

@bryanmuntzer They’ve never been the same since the Niners game

[Michael Fitzpatrick @mitrick2]

@brahmin_noodles Looks more like tequila 🤔

[Michael Fitzpatrick @mitrick2]

@MikeHamad Excellent work sir! Please continue

[Michael Fitzpatrick @mitrick2]

Magic Brush from @runwayml is a great example of a scaled annotation system where both creators and the company benefit from its use. Creators get more control over their video production, while the company gets vast amounts of high quality annotation data