Did you know elephants are helping shape robot design?

Their trunks' precise movements have been inspiring new tech innovations for 20 years. Researchers are using the insights to create robotic limbs with precise grips that can work to pick up objects in the unpredictable real world:
theconversation.com/elephants-…
#tech #research #wildlife

A friend asked me to demo the sound effects I've made with #ElevenLabs. This ended up longer than I thought, so have almost 20 minutes of random noises, music and utterances that make absolutely no sense. Enjoy.
onj.me/media/Eleven_Labs_Sound…

reshared this

Ein schlechter Scherz? Leider nein. Statt das 49-Euro-Ticket zu sichern, will die Regierung das Dienstwagenprivileg ausbauen.

Wir fordern zusammen mit vielen weiteren Verbänden, den Haushalt an sozialer Gerechtigkeit und Klimaschutz zu orientieren!

49-Euro-Ticket für alle statt Luxus-Dienstwagen für wenige! Dafür machen wir Druck! Willst du uns dabei helfen? Dann unterschreibe jetzt unsere Petition!
vcd.org/macht-das-49euro-ticke…

#Deutschlandticket #Dienstwagenprivileg

This entry was edited (1 year ago)

Wer unsere Undercover-Recherche zu Lobbyismus im Regierungsviertel verpasst hat: Die ZDF-Doku gibt's jetzt auch auf Youtube:
youtube.com/watch?v=rjjZQAuy1T…

We have a small gift for the #Fediverse today: over the past 8 months, we've been building an icon library for various decentralized social platforms and protocols within the space. It's an ongoing effort, and we're happy to release our initial version today. It's like Font Awesome, but for the Social Web!

wedistribute.org/2024/08/decen…

reshared this

TIL Philadelphia Cream Cheese is produced in New York, and always has been since it was invented by William Lawrence in 1872. He named his company Philadelphia to market off of the city's reputation for high quality dairy products.

en.wikipedia.org/wiki/Philadel…
#til #todayilearned
reddit.com/r/todayilearned/com…

Y'know, for a project that really likes to dogpile on bloggers for not citing them, LuaJIT's docs are incredibly sparse.

I say this as a person trying to learn about its internals.

Below, the closest thing to docs or papers I can find on their site. Yes, it's instructing you to google basic compiler terminology, plus a 2009 mailing list post.

The amount of interpretive labor this asks of its reader is huge. I'm good, I'll read about other JITs.

luajit.org/faq.html

This entry was edited (1 year ago)
in reply to Lars Marowsky-Brée 😷

Spelling is a known thing that LLMs are bad at, and it doesn't reflect on their performance on other problems.

LLMs don't see language like we do, all input is first tokenized (AKA turned into subwords) before passed into an LLM.

Most common words are just one token, long and unusual words are two or three. Common combinations of letters, like "ing", "able", "anti" etc also form tokens, so a word like "antiwiral" might be passed as ["anti", "wir", "al"].

LLMs have never really seen letters, you can imagine them as extremely intellectually sophisticated people who enjoy all content through audiobooks and have never actually seen written text.