savant_penguin

1. Incorrect. The output of the decoder LLM is the probability distribution of the next token given the input text. Temperature=0 means that the output distribution is not pushed to be closer to a uniform distribution. The randomness comes from the sampling of the next token according to the output distribution to generate text. If you want determinism you always get the argmax of the distribution.

+ optimization

It looks fantastic for the compute budget

Could you provide one example of communism that did not end up an authoritarian regime?

Communism is just not compatible with freedom. When free to choose, societies do not become communist, they embrace production and trade. You cannot achieve what communists want without extensive use of force

"Some jobs will go away, that's fine."

To whom is that fine? To the people going from $2/hour to zero it certainly isn't.

I honestly cannot grasp how some people can _forcibly_ remove other people's _options_ then pat themselves on the back as some sort of armchair savior.

Yes, for some people a bad option could still be their best option. What sort of moral superpower is that that enables you to forbid a contract between two consenting adults A and B such that A wants to work for X and B that's willing to pay X, but you as a C that just won't let it happen and is willing to use force to stop it from happening

It's usually people who _really_ want to decide what can be said

Maybe I'm oversimplifying but the question is: "does an artist _own_ a style?"

To me the answer is no.

This is hilarious

Now feed it Doom haha

"Competition is a bitch" - OpenAI, probably

In this case I believe python is faster by a few months.

Jokes aside this is pytorch so this is compiled to C++ or cuda, the problem likely comes from the different functions that are called for += vs +

If the disruption of supply chains is the actual culprit for inflation, shouldn't we observe gradual deflation as they recover?

If on the other hand it's printing money (as I'd expect from the trillions of free money in recent months) it should only get worse. What I observe is more consistent with money printing than disruption of supply chains

"In some cases, however, potential harms to the populations studied may outweigh the benefit of publication."

I'm sure our demigods will gladly enlighten us with their wisdom to distinguish between misthoughts and correct thoughts.

So that when journalists claim to be following the science there will be no dissenting voice in research to disagree.

Those who defend opposing ideas will have no leg to stand on and will rightfully be labeled science deniers

Just to comment in the "little bit of free code" there are projects like python and numpy that would take ages to do properly and depending on the company would never ship (or function as reliably as it does right now)

Sometimes they're buying decades worth of debugging/testing