qmatch

Did you use Cursor with Gemini and GPT?

I personally believe TF1 was serving the need of its core users. It provided a compileable compute graph with autodiff, and you got very efficient training and inference from it. There was a steep learning curve, but if you got past it, things worked very very well. The distributed TF never really took off—it was buggy, and I think they made some wrong earlier bets in the design for performance reasons that they should have been sacrificed in favor of simplicity.

I believe some years after the TF1 release, they realized the learning curve was too steep, they were losing users to PyTorch. I think also the Cloud team was attempting to sell customers on their amazing DL tech, which was falling flat. So they tried to keep the TF brand while totally changing the product under the hood by introducing imperative programming and gradient tapes. They killed TF1, upsetting those users, while not having a fully functioning TF2, all the while having plenty of documentation pointing to TF1 references that didn’t work. Any new grad student made the simple choice of using a tool that was user-friendly and worked, which was PyTorch. And most old TF1 users hopped on the band wagon.

As a loyal JAX user, I hope they can play catchup. PyTorch has dominated the AI scene since TF1 fumbled the ball at 10th yard line. What Matt Johnson has done turning Autograd into JAX is hopefully going to be worthy of as much praise as what Soumith has received.

I see your point, but not totally buying it. The US innovates for a global population, one that’s still growing.

The best way to infer causality is through experimentation. If regulation does go away, we’ll measure and learn if it actually worked.

How are so confident in causality here?

Similarly curious if anyone has an API, that they ultimately subsidize through some other service or product.

M, N, O, P, …

Need to read the details, but removing the norm can be big. It’s always a pain to make sure that your network is normalized properly when trying new architectures. Likely there will still be other implications of the tanh, since the norm is sometimes solving a conditioning problem, but IMO more alternatives are welcome

This is awesome—nice job! It was so cool to see when I started running and what year I started to get injured :/

What did the line represent? It’s always surprising how predictable it is to estimate people’s race times given what you know about their fitness. That said, about 10% of the time someone surprises you with a time you wouldn’t expect.

Try adjusting the slider! Gives us hope.

Is this similar to the Universal Approximator Theorem?