Simone
@simone_z
- 80 Followers
- 73 Following
- 661 Posts
He/Him|MSc|PhD
#Researcher|#MachineLearning|#Bioinformatics|#AI. #Geek, I enjoy #games (board|video|role|retro), (web) #comics, #anime, while hoping for a better world.
Marius Fortuna_3
Manish VijVery funny AI-denier troll (original post is a genuine Monet) https://xcancel.com/shl0ms/status/2054280631807316329
Rachel BurchOaks in mist along a footpath near Sheepstor, I hope to see them again soon. Taken April 2024, #Dartmoor #Devon #photography #oaks #mosstodon
LogicalErzorRE: https://mastodon.social/@geerlingguy/115945025024114949
this and https://github.com/kyutai-labs/pocket-tts are both scary to clone anyone’s voice. can easily be used for nefarious purposes unfortunately
on a slightly positive note, audiobooks will be easy to generate and people can easily read while doing something else at the same time. plus more voices for legally blind people
Stephen TurnerIt's kind of wild how quick and easy it is to clone a voice with ~5 seconds of audio. github.com/kyutai-labs/...
uvx pocket-tts generate --voice "./my_sample.wav" --text "whatever you want it to say"
GitHub - kyutai-labs/pocket-tt...
GitHub - kyutai-labs/pocket-tt...
Mike @ Misaligned Markets(11/12) How does the bubble unwind? Already is. Watch the pricing games and the shift to charging for tokens. AI companies have been operating on Loony Toons physics, but gravity is catching up.
As the bubble pops, expect the house of cards around data center financing to implode. This is as much a “Data Center bubble” as it is an “AI bubble.”
https://misaligned.markets/what-to-expect-ai-bubble/#1-implosion-of-the-ai-startup-ecosystem
« A small number of samples can poison LLMs of any size » \ Anthropic
As little as 500 documents can poison (hijack the hyper trajectory that determines the sequence of generated contents) to make a LLM as large as 13B hyper parameters output what the attacker wants it to. Small reminder that no, #EU cannot rely on "open" models made under/by dictatorial or aspiring dictatorial governments.
https://www.anthropic.com/research/small-samples-poison?ref=misaligned.markets
As little as 500 documents can poison (hijack the hyper trajectory that determines the sequence of generated contents) to make a LLM as large as 13B hyper parameters output what the attacker wants it to. Small reminder that no, #EU cannot rely on "open" models made under/by dictatorial or aspiring dictatorial governments.
https://www.anthropic.com/research/small-samples-poison?ref=misaligned.markets
Roonie the RabbitHow to deal with expectations. | Early bird Kickstarter deal live - https://www.kickstarter.com/projects/portermason/cute-animals-for-flawed-people?ref=49i8hi
nixCraft 🐧Google Chrome silently installs a 4 GB AI model on your device https://www.thatprivacyguy.com/blog/chrome-silent-nano-install/
> No consent dialog. No opt-out UI. Re-installs itself if the user removes it manually.
That is the true definition of malware.
Google Chrome silently installs a 4 GB AI model on your device without consent. At a billion-device scale the climate costs are insane. — That Privacy Guy!
Google Chrome is downloading a 4 GB Gemini Nano model onto users' machines without consent, with no opt-in, no opt-out short of enterprise tooling, and an automatic re-download every time the user deletes it. The pattern is identical to the Anthropic Claude Desktop case I wrote about last month, but the scale is between two and three orders of magnitude larger. This article does the legal analysis and, for the first time, the environmental analysis. The numbers are not small.
Brian BìnhThe English suffix "-ard" (as in "drunkard", "laggard", "braggard", and "sluggard") is used to make adjectives into deprecating nouns.
"He's drunk" is just a description.
"He's a drunkard" is a moral judgment that finds the subject lacking.
The same process turned "he's wise" into "he's a wizard".
"Wizard" was the medieval equivalent of "smartass".