michael_feffer's profile picture. great value chidi anagonye. more seriously Societal Computing PhD student @S3DatCMU. not just an ML account.

Michael Feffer

@michael_feffer

great value chidi anagonye. more seriously Societal Computing PhD student @S3DatCMU. not just an ML account.

Michael Feffer reposted

Hello! We are researchers from Carnegie Mellon University. We are looking for mental health professionals or end users who have sought mental health support from LLM-based products to participate in a survey and workshop study. (Details in flyer) Sign-up: forms.gle/AHniGREb1PuXng…

Ningjing_Tang's tweet image. Hello! We are researchers from Carnegie Mellon University. We are looking for mental health professionals or end users who have sought mental health support from LLM-based products to participate in a survey and workshop study. (Details in flyer)
Sign-up: forms.gle/AHniGREb1PuXng…

Michael Feffer reposted

Excited to share our beta release of Music Arena, a live evaluation platform for state-of-the-art AI music generation models! 🎧 Listen to the latest models and 🗳️ vote for your favorite ⚔️ music-arena.org ⭐️ github.com/gclef-cmu/musi… 📜 arxiv.org/abs/2507.20900

chrisdonahuey's tweet image. Excited to share our beta release of Music Arena, a live evaluation platform for state-of-the-art AI music generation models!

🎧 Listen to the latest models and 🗳️ vote for your favorite

⚔️ music-arena.org 
⭐️ github.com/gclef-cmu/musi…
📜 arxiv.org/abs/2507.20900

Michael Feffer reposted

Excited to announce 🎵Magenta RealTime, the first open weights music generation model capable of real-time audio generation with real-time control. 👋 **Try Magenta RT on Colab TPUs**: colab.research.google.com/github/magenta… 👀 Blog post: g.co/magenta/rt 🧵 below


Michael Feffer reposted

Our study highlighting plagiarism concerns in AI-generated research is now accepted to ACL (main conference): arxiv.org/abs/2502.16487. Effort led by amazing @tarungupta360. Will share other accepted papers soon. Stay tuned 🙂

Remember this study about how LLM generated research ideas were rated to be more novel than expert-written ones? We find a large fraction of such LLM generated proposals (≥ 24%) to be skillfully plagiarized, bypassing inbuilt plagiarism checks and unsuspecting experts. A 🧵



Michael Feffer reposted

🏳️‍🌈🎨💻📢 Not on here much anymore, but happy to share our workshop study on queer artists’ experiences critically engaging with GenAI Looking forward to presenting this work at #FAccT2025 and you can read a pre-print here: arxiv.org/abs/2503.09805

nprandchill's tweet image. 🏳️‍🌈🎨💻📢 Not on here much anymore, but happy to share our workshop study on queer artists’ experiences critically engaging with GenAI  

Looking forward to presenting this work at #FAccT2025 and you can read a pre-print here: arxiv.org/abs/2503.09805

Michael Feffer reposted

In a preview of EMPIRE OF AI for @TheAtlantic, I share new behind-the-scenes details on what happened when OpenAI’s board fired Sam Altman—as well as the core argument of my book and why it matters to everyone. Gift link. theatlantic.com/technology/arc…


Michael Feffer reposted

Five hours after our AI research story published — we reached out to all the AI labs 1-2 weeks beforehand — OpenAI pledged it would release safety test results more often & launched a webpage showing how its models score on tests for harmful content. cnbc.com/2025/05/14/met…


Michael Feffer reposted

Q: What do school buses, desktop computers, and AI have in common? A: The same decades-old laws and processes apply when governments go to purchase them. Our new #FAccT2025 paper examines how these legacy public procurement norms apply to AI. 🧵

narijohnson's tweet image. Q: What do school buses, desktop computers, and AI have in common?
A: The same decades-old laws and processes apply when governments go to purchase them.

Our new #FAccT2025 paper examines how these legacy public procurement norms apply to AI. 🧵

Michael Feffer reposted

"Copyright is outdated and should be abolished to make way for AI.... unless the US Copyright Office changes its mind on AI, in which case I would like some copyright please." - AI bros


Michael Feffer reposted

You don't need to attempt to revitalize NFTs when companies can just NOT steal.


Michael Feffer reposted

"[AI agents] are threatening to break the blood-brain barrier between the application layer and the OS layer." Signal President Meredith Whittaker (@mer__edith) warns of "real danger" in agentic AI hype. To be "magic genie bots" (concert booking etc.), they need root access to…


Michael Feffer reposted

I don't think people realize how incredible it is that a movie rendered in nearly real-time in free software on a single computer just won the Oscar for best animated film. If there was ever any proof you don't need AI, this is it.

Flow was made entirely with Blender and rendered in real time with EEVEE. We didn't use a renderfarm. The final render was done on my PC. There was no compositing, all the colors were tweaked and adjusted using shaders.



Michael Feffer reposted

With the market & AI mythos reeling post-DeepSeek, it seems like a good time to reup this year-old paper, on the perils of the "bigger is better" approach to AI, coauthored with @SashaMTL + @GaelVaroquaux arxiv.org/abs/2409.14160

mer__edith's tweet image. With the market & AI mythos reeling post-DeepSeek, it seems like a good time to reup this year-old paper, on the perils of the "bigger is better" approach to AI, coauthored with @SashaMTL + @GaelVaroquaux 

arxiv.org/abs/2409.14160

Michael Feffer reposted

In every paper on LLM attacks/jailbreaks, there are subtly different experimental setups. s.t. each one basically has its own unique operationalization of "jailbreak" and "ASR". People are usually much too quick to make poor apples-to-apples comparisons between papers.


Michael Feffer reposted

It's objectively funny to see DeepSeek — an open source, high performing AI model trained by a Chinese startup for less than $6 million — wipe $1 trillion off tech stocks only days after Trump, OpenAI, and SoftBank announce their $500 billion Stargate boondoggle.


Michael Feffer reposted

AI agents are coming. But who's in control? @mmitchell_ai, one of the best researchers in AI ethics, makes a critical point about autonomy: fully autonomous systems carry unknowable risks because they operate on computer logic rather than human logic. The solution? Build…


Michael Feffer reposted

Deregulating nuclear power plants? What's the worst that could happen, it's just bureaucracy. Surely nothing bad ever happened in one.

HeidyKhlaaf's tweet image. Deregulating nuclear power plants? What's the worst that could happen, it's just bureaucracy. Surely nothing bad ever happened in one.

Michael Feffer reposted

“the green new deal is expensive” ok you know what else is expensive the entire country burning the fuck down


Michael Feffer reposted

Hey remember 10-15 years ago when they used to throw people in jail if they pirated films to make an example out of them? Turns out, totally fine for "AI" companies.

New allegations: Meta didn't just download "pirated copyrighted material... for use in its commercially available AI models", they even shared it with peers once they were done downloading (seeding). 🏴‍☠️ Why would they do that? Out of courtesy? 🙉 wired.com/story/new-docu…

wired.com

Meta Secretly Trained Its AI on a Notorious Piracy Database, Newly Unredacted Court Docs Reveal

One of the most important AI copyright legal battles just took a major turn.



Loading...

Something went wrong.


Something went wrong.