running wavenet on an FPGA at (almost) 200,000 inferences a second.

https://matpalm.com/blog/wavenet_on_fpga/

https://www.youtube.com/watch?v=-14O8rNdFyQ

The hot mess theory of AI misalignment (+ an experiment!)
https://sohl-dickstein.github.io/2023/03/09/coherence.html

There are two ways an AI could be misaligned. It could monomaniacally pursue the wrong goal (supercoherence), or it could act in ways that don't pursue any consistent goal (hot mess).

Here is your must-read article for the day, a profile of @emilymbender, and her efforts to deflate the ridiculous hype around large language models such as ChatGPT.

It's also about the people who are behind that hype, and about what their way of thinking has the potential to do to us.

It's worth reading all the way to the end.

https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html

Turns out the existence of an "alpha" wolf is a myth.

It was based on old/bad experiments, conducted when unrelated wolves were thrown together in captivity.

In actual wild wolves, packs are just families. There's no dominant alpha battling for the top spot -- and in the rare cases there is, it's an alpha *female* wolf.

Insert obvious implications for the use of this term in human societies.

https://www.scientificamerican.com/article/is-the-alpha-wolf-idea-a-myth/

Do you have any question about #pandas? Few core devs including myself will be answering questions in an AMA (ask me anything) session. Officially scheduled for tomorrow Thursday at 5:30pm UTC, but already open.

https://www.reddit.com/r/Python/comments/11fio85/we_are_the_developers_behind_pandas_currently/