Rodri Medina

Cosas veredes amigo Sancho

[Charlie O'Neill @oneill_c]

1/ You can shrink a language model's KV cache by 200×, in a single forward pass, and it still answers correctly.

At 256k context that's 36 GiB of cache down to ~360 MiB, with no change to the base model.

Here's how we did it 👇

[Garry Tan @garrytan]

A 6-person team is building task-specific AI models that are 4-8x faster than anything from OpenAI or Anthropic. 500K downloads on HuggingFace. No hype. Just better engineering winning on the merits.

This is what "make something people want" looks like in the model layer.

https://t.co/nsf8b31xha

[Antonio Lupetti @antoniolupetti]

This is Algebrica. A mathematical knowledge base I’ve been building for 2.5 years.

215+ entries, carefully written and structured.

400k+ views over this time. Not much in absolute terms, but meaningful to me.

No ads.
No courses to sell.
No gamification.
No distractions.

Just essential pages, aiming to explain mathematics as clearly as possible, for a university-level audience.

Built simply for the pleasure of sharing knowledge.

Content licensed under Creative Commons (BY-NC).

Best experienced on desktop.

If it helps even a few people understand something better, it’s worth it.

[Rodri Medina @AlonsoQuixada]

💔

[TBPN @tbpn]

OpenAI Acquires TBPN https://t.co/Sdw32Weo4z

[Tivadar Danka @TivadarDanka]

Just put the knowledge graph explorer online.

Check it out: https://t.co/NGQPYGb4GG

[Rodri Medina @AlonsoQuixada]

The year of FOMO

[el.cine @EHuanglu]

friend in china send me this and said these are basically his employees.. but work 24/7

wild time we are living in

[Bojan Tunguz @tunguz]

Here is a snapshot of a (small) subset of all of my Coding, Data Science and Machine Learning books. This collection would get you close to 98%-99% of all the necessary core skills to be a good Data Scientists. 1/6

[Philip Vollet @philipvollet]

How it started: How it's going:

[Mark Tenenholtz @marktenenholtz]

6 habits of highly successful machine learning modelers:

[Rodri Medina @AlonsoQuixada]

@jessesingal

[Greg Yang @TheGregYang]

2/ The idea is actually really simple: in a special parametrization introduced in https://t.co/vhvvXylq58 called µP, narrow and wide neural networks share the same set of optimal hyperparameters. This works even as width -> ∞.

[Greg Yang @TheGregYang]

1/ You can't train GPT-3 on a single GPU, much less tune its hyperparameters (HPs).

But what if I tell you…

…you *can* tune its HPs on a single GPU thanks to new theoretical advances?

paper https://t.co/urEUY7O3yQ
code https://t.co/5S0YAghCYx
blog https://t.co/QPqinMwOXj

[Joseph Viviano @josephdviviano]

Yoshua Bengio wrote a tutorial on GFlowNets which are interesting for a suprisingly large number of ML problems

https://t.co/C7iQqJyIX6

[C. S. Lewis @CSLewisDaily]

“God, who foresaw your tribulation, has specially armed you to go through it, not without pain but without stain.” - C.S. Lewis

[Kevin Patrick Murphy @sirbayes]

I just got a hardcopy of my own book - it finally feels real :) I'm very happy with the print quality (see attached photo) - but you can still get the pdf for free from https://t.co/dSlKkwYpLr if you prefer.

[Bojan Tunguz @tunguz]

Being able to think in terms of probabilities is a superpower, period. And that includes your own beliefs and opinions. Probabilistic and statistical thinking unfortunately doesn’t come naturally to us. Takes years of training and experience to develop.

[Bojan Tunguz @tunguz]

If you find yourself being asked about this in a DS job interview, or even worse to *code* an ML algorithm from scratch, just end the interview there and then. Trust me on this one.

[C. S. Lewis @CSLewisDaily]

“No man knows how bad he is till he has tried very hard to be good.” - C.S. Lewis