J Daniel

Economist & Machine Learning Engineer. Pain is temporary, success lasts forever.

[Sam Elliott @samelliottsam]

Early anthropic employees commuting to work post IPO

[J Daniel @JDanielFD]

@NegroPichardo @TecnoSinHumo No es claro que lo ve, con la diferencia de velocidad de los coches es posible que ni siquiera se haya dado cuenta que venía el Porsche atrás.

[J Daniel @JDanielFD]

@NegroPichardo @TecnoSinHumo Coincido

[Andrej Karpathy @karpathy]

I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then:

- the human iterates on the prompt (.md)
- the AI agent iterates on the training code (.py)

The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc.

https://t.co/YCvOwwjOzF
Part code, part sci-fi, and a pinch of psychosis :)

[J Daniel @JDanielFD]

@ecommartinez 4.6

[J Daniel @JDanielFD]

@ElMalaguero Argentino: penaaal para Argentina!

[J Daniel @JDanielFD]

@wholesome_X_ That’s in Pittsburgh

[J Daniel @JDanielFD]

@F1 Vaaaaaaaamooooos!

[J Daniel @JDanielFD]

@_onlyscott Di María

[J Daniel @JDanielFD]

@robin_j_brooks What do you think of the following: The dotted lines are parallel (blue and red)
(Roughly) Seems like the trend of Mexico’s’ growth is the same before and after 2018, but Mexico did not recover that well from the COVID shock

[MisterChip (Alexis) @2010MisterChip]

¿Qué le vais a pedir que gane ahora? ¿Quizá algún torneo intergaláctico contra extraterrestres de cuatro brazos y seis piernas?

EL PRIMER ENTRENADOR EN LA HISTORIA DEL FÚTBOL QUE GANA DOS TRIPLETES ES...

[J Daniel @JDanielFD]

@I_Yellowlot @_Ramtin_ @FDataAnalysis This is interesting. So, why @FDataAnalysis says in other tweet, medium-hards was the optimal strategy?

[J Daniel @JDanielFD]

@Leo_P999 @FDataAnalysis I agree. I want to see what’s the data behind that statement.

[J Daniel @JDanielFD]

@FDataAnalysis Was medium- hard the best strategy m? What makes you said so? What data?
Was the beginning difficult on the mediums because of tyre degradation?

[J Daniel @JDanielFD]

@SChecoPerez Venga! Por el liderato de pilotos!

[J Daniel @JDanielFD]

Oleeeé oleeé olé olé Poleee Poleee
Vaaaamos ! @SChecoPerez

[José Antonio Meade🇲🇽 @JoseAMeadeK]

A Seneca le hubiera gustado darle seguimiento a la #F1.
Le dio al clavo cuando dijo que “La suerte es lo que sucede cuando la preparación se encuentra con la oportunidad”.
Así el @SChecoPerez hoy! Seguro también mañana.
#GPMiami.

[J Daniel @JDanielFD]

@F1Tornello Y si perjudicó el SC a Max? Salió de pits en 7 y tenía tráfico adelante. Con el SC todos los de adelante entraron a pits antes de lo planeado y se le liberó camino a Max. Únicamente le quedó LEC a quien pasó en un par de curvas. Max hubiera perdido tiempo en tráfico sin el SC.

[J Daniel @JDanielFD]

@honestlhfan @WasGarbage @F1 Lol NO

[J Daniel @JDanielFD]

@dhavalkhatadia @F1 @SChecoPerez Totally nauseating