The advantages of having a single big memory per gpu are not as big in a data center where you can just shard things between machines and use the very fast interconnect, saturating the much faster compute cores of a non Apple GPU from Nvidia or AMD

AI is threatening to remove humans from the whole equation except the very top. AI is an existential threat (not in the Terminator sense).

Especially for art, I'm an AI researcher myself (in bio for health), but I think that ppl are completely understandable for wanting to help artists make a living and want to consume something that someone cared about

This interpretation would have been ok for old generation models without search tools enabled and without reliable tool use and reasoning. Modern LLMs can actually look up the existence of papers with web search, and with reasoning, one can definitely get reasonable results by requiring the model to double check that everything actually exists.

Isn't this mostly a set of citation typos? To me this mostly calls for better bibtex checking, writing and checking bibtex is super annoying

Forgetting authors, misspelling them or the journals, putting a wrong digit etc... could be citation typos. I don't see how you add 5 non-existing authors and put a different—but conceptually plausible—journal in the bibtex.

Besides, I would think most people are using bibliographic managers like Zotero&co..., which will pull metadata through DOIs or such.

The errors look a lot more like what happens when you ask an LLM for some sources on xyz.

If a person usually uses Zotero to manage literature and finds incomplete metadata when exporting BibTeX, and with the submission deadline approaching, they use GPT to complete the metadata, leading to errors, this is indeed lazy and negligent behavior. But is it what many call deceitful and unforgivable?

I believe that once this person realizes the unreliability of using GPT to complete metadata, they will no longer use such methods in the future.

I also look forward to the community's dedicated individuals developing more comprehensive automated export tools, as copying and pasting one by one is inherently tedious and should be automated.

Currently, these individuals used incorrect automated tools and placed excessive trust in them, resulting in errors. This is a profound lesson that must never be repeated.

I think it's not uncommon to ask an LLM for the bibtex for a paper you know about, & it might mess it up, but that doesn't feel like a fireable offense

Architecture difference wrt vanilla transformers and between modern transformers are a tiny part of what makes a model nowadays

I suspect they will keep doing this until they have a substantially better model than the competition. Sharing methods to look good & allow the field to help you keep up with the big guys is easy. I'll be impressed if they keep publishing even when they do beat the big guys soundly.

yes, they turned off all energy economy measures when benchmarking software activity was detected, which completely broke the point of the benchmarks because your phone is useless if it's very fast but the battery lasts one hour

I would assume that huge amount is spent in frontier models just making the models nicer to interact with, as it is likely one of the main things that drives user engagement.

this is such a bad way to evaluate a metric. we want something that can distinguish small percentage changes in affordability for the people who live there

Sidenote, it's interesting how the term "virtue signaling" is arguably objectively an individualistic right-wing dog whistle these days.

I would argue that it is being used all over the media to complain about anyone showing any signs of not being purely individualistic, as if individualism is the only true thing people actually honestly feel. This is obviously incorrect, empathy, professionalism, a desire for a sense of purpose, are all things that people objectively feel in the real world, everyday, everywhere.

I would argue that the expression "virtue signaling" is used systematically in individualistic right wing media by the right about anyone who say, for example, that they care about minorities or less fortunate people or to take action to support them, as if it was false. I would argue that this is harmful.

People do care a good fraction of the time, and they should be recognized for their positive actions, and encouraged. I would argue that we should definitely strive for a culture where individualism is not seen as the only true emotion that people can feel.

So, knowing the negative political and philosophical baggage, I would not use that expression, especially if you don't have actual proof that they don't care about security, professionalism, etc.