Bálint Gyevnár
@gbalint.bsky.social
Postdoc @ CMU • Metascience of AI: safeguarding scientific integrity in the age of AI • Explainable multi-agent systems @ University of Edinburgh • gbalint.me • 🇭🇺🏴
Pinned
Bálint Gyevnár
@gbalint.bsky.social
· Apr 17
'AI Safety for Everyone' is out now in @natmachintell.nature.com! Through an analysis of 383 papers, we find a rich landscape of methods that cover a much larger domain than mainstream notions of AI safety. Our takeaway: Epistemic inclusivity is important, the knowledge is there, we only need use it
Reposted by Bálint Gyevnár
And now the propaganda rag "Magyar Nemzet" (along with other pro-Orban media) shared - then removed - a map containing the personal data, including the address, of the opposition supporters and activists whose data was stolen. Clear and brazen intimidation tactics.
444.hu/2025/11/07/n...
444.hu/2025/11/07/n...
November 8, 2025 at 2:01 AM
And now the propaganda rag "Magyar Nemzet" (along with other pro-Orban media) shared - then removed - a map containing the personal data, including the address, of the opposition supporters and activists whose data was stolen. Clear and brazen intimidation tactics.
444.hu/2025/11/07/n...
444.hu/2025/11/07/n...
Reposted by Bálint Gyevnár
Do you teach #rstats? Do your students complain about how lame and old-fashioned dplyr is? Don't worry: I have the solution for you: github.com/hadley/genzp....
genzplyr is dplyr, but bussin fr fr no cap.
genzplyr is dplyr, but bussin fr fr no cap.
GitHub - hadley/genzplyr: dplyr but make it bussin fr fr no cap
dplyr but make it bussin fr fr no cap. Contribute to hadley/genzplyr development by creating an account on GitHub.
github.com
November 6, 2025 at 11:25 PM
Do you teach #rstats? Do your students complain about how lame and old-fashioned dplyr is? Don't worry: I have the solution for you: github.com/hadley/genzp....
genzplyr is dplyr, but bussin fr fr no cap.
genzplyr is dplyr, but bussin fr fr no cap.
Reposted by Bálint Gyevnár
I had a long conversation with MIT Tech Review’s @willdouglasheaven.bsky.social about AGI; it fed into this very good read on the rise of AGI orthodoxy
How AGI became the most consequential conspiracy theory of our time
The idea that machines will be as smart as—or smarter than—humans has hijacked an entire industry. But look closely and you’ll see it’s a myth that persists for many of the same reasons conspiracies d...
www.technologyreview.com
November 2, 2025 at 2:19 AM
I had a long conversation with MIT Tech Review’s @willdouglasheaven.bsky.social about AGI; it fed into this very good read on the rise of AGI orthodoxy
Reposted by Bálint Gyevnár
Arxiv published an enforcement change where position papers & surveys need to be accepted to conferences before they can be uploaded to arxiv.
This is the wrong decision. What Arxiv is in practice versus what it is in reality is very different.
open.substack.com/pub/natolamb...
This is the wrong decision. What Arxiv is in practice versus what it is in reality is very different.
open.substack.com/pub/natolamb...
Contra Arxiv Moderation
Opposing a recent Arxiv change.
open.substack.com
November 1, 2025 at 4:17 PM
Arxiv published an enforcement change where position papers & surveys need to be accepted to conferences before they can be uploaded to arxiv.
This is the wrong decision. What Arxiv is in practice versus what it is in reality is very different.
open.substack.com/pub/natolamb...
This is the wrong decision. What Arxiv is in practice versus what it is in reality is very different.
open.substack.com/pub/natolamb...
I have been doing ML wrong my whole life: arxiv.org/abs/2309.08632
Pretraining on the Test Set Is All You Need
Inspired by recent work demonstrating the promise of smaller Transformer-based language models pretrained on carefully curated data, we supercharge such approaches by investing heavily in curating a n...
arxiv.org
November 1, 2025 at 9:54 PM
I have been doing ML wrong my whole life: arxiv.org/abs/2309.08632
Cycling in Pittsburgh is great: the city is beautiful in the autumn colours, I never need to do cardio training again, and one may experience the "Pittsburgh left" which is when oncoming cars cut ahead of you at an unprotected left turn.
en.wikipedia.org/wiki/Pittsbu...
en.wikipedia.org/wiki/Pittsbu...
Pittsburgh left - Wikipedia
en.wikipedia.org
October 26, 2025 at 5:09 PM
Cycling in Pittsburgh is great: the city is beautiful in the autumn colours, I never need to do cardio training again, and one may experience the "Pittsburgh left" which is when oncoming cars cut ahead of you at an unprotected left turn.
en.wikipedia.org/wiki/Pittsbu...
en.wikipedia.org/wiki/Pittsbu...
Reposted by Bálint Gyevnár
new blog post: Vibe coding in the 90's
Vibe coding in the 90's
Hey folks. You know how programming is hard in 1994. You need books, help files, manuals, references to learn about a certain feature. You read all those. Then you go through endless iteration of tria...
ssg.dev
October 24, 2025 at 11:52 PM
new blog post: Vibe coding in the 90's
Reposted by Bálint Gyevnár
Oh my god some fucking tech dingdong posted this on Twitter with the caption "AI games are going to be amazing" totally seriously, you have to watch it. You have to. In full screen.
October 24, 2025 at 3:02 PM
Oh my god some fucking tech dingdong posted this on Twitter with the caption "AI games are going to be amazing" totally seriously, you have to watch it. You have to. In full screen.
TIL that the expression: constant == expression, is called a Yoda condition and it makes so much sense.
docs.astral.sh/ruff/rules/y...
docs.astral.sh/ruff/rules/y...
yoda-conditions (SIM300) | Ruff
Checks for conditions that position a constant on the left-hand side of the
docs.astral.sh
October 23, 2025 at 10:19 PM
TIL that the expression: constant == expression, is called a Yoda condition and it makes so much sense.
docs.astral.sh/ruff/rules/y...
docs.astral.sh/ruff/rules/y...
I just saw a recent @nature.com paper that designs parachutes by literally cutting hundreds of holes into a disc.
Inspired by kirigami, which is like origami but instead of folds you make cuts, if you cut a disc the right way, then you might just be able to jump out of a plane and survive.
Inspired by kirigami, which is like origami but instead of folds you make cuts, if you cut a disc the right way, then you might just be able to jump out of a plane and survive.
October 21, 2025 at 8:10 PM
I just saw a recent @nature.com paper that designs parachutes by literally cutting hundreds of holes into a disc.
Inspired by kirigami, which is like origami but instead of folds you make cuts, if you cut a disc the right way, then you might just be able to jump out of a plane and survive.
Inspired by kirigami, which is like origami but instead of folds you make cuts, if you cut a disc the right way, then you might just be able to jump out of a plane and survive.
Throw back to a cloud inversion on Beinn Eighe a year ago. If you have the chance, explore the Scottish Highlands!
October 14, 2025 at 3:27 PM
Throw back to a cloud inversion on Beinn Eighe a year ago. If you have the chance, explore the Scottish Highlands!
The Hungarian government often mails people “National Consultations” paid for by taxpayers, asking violently condescending questions:
“Do you wanna raise taxes?”
A: “Yes, I like paying more to fund those who don’t work.”
B: “No.”
Then they use the results to fake support for their agenda…
“Do you wanna raise taxes?”
A: “Yes, I like paying more to fund those who don’t work.”
B: “No.”
Then they use the results to fake support for their agenda…
October 12, 2025 at 4:48 PM
The Hungarian government often mails people “National Consultations” paid for by taxpayers, asking violently condescending questions:
“Do you wanna raise taxes?”
A: “Yes, I like paying more to fund those who don’t work.”
B: “No.”
Then they use the results to fake support for their agenda…
“Do you wanna raise taxes?”
A: “Yes, I like paying more to fund those who don’t work.”
B: “No.”
Then they use the results to fake support for their agenda…
Just came across this mesmerising art by Estampa of the components of GenAI interacting with humanity: cartography-of-generative-ai.net
Reminds me a little of Kate Crawford and Vladan Joler's Anatomy of an AI System: anatomyof.ai
Reminds me a little of Kate Crawford and Vladan Joler's Anatomy of an AI System: anatomyof.ai
Cartography of generative AI
cartography-of-generative-ai.net
October 11, 2025 at 6:29 PM
Just came across this mesmerising art by Estampa of the components of GenAI interacting with humanity: cartography-of-generative-ai.net
Reminds me a little of Kate Crawford and Vladan Joler's Anatomy of an AI System: anatomyof.ai
Reminds me a little of Kate Crawford and Vladan Joler's Anatomy of an AI System: anatomyof.ai
Reposted by Bálint Gyevnár
Please share!
We have a number of fully funded PhD studentships in "Designing Responsible Natural Language Processing". I'm a possible supervisor & I'd be keen to support projects on sociolinguistics-AI, e.g., accent bias in AI, language+gender/sexuality+AI.
www.responsiblenlp.org
We have a number of fully funded PhD studentships in "Designing Responsible Natural Language Processing". I'm a possible supervisor & I'd be keen to support projects on sociolinguistics-AI, e.g., accent bias in AI, language+gender/sexuality+AI.
www.responsiblenlp.org
Our CDT is based in the Edinburgh Futures Institute – the University of Edinburgh’s brand new hub for research, innovation and teaching focused on socially just artificial intelligence and data.
www.responsiblenlp.org
October 10, 2025 at 3:03 PM
Please share!
We have a number of fully funded PhD studentships in "Designing Responsible Natural Language Processing". I'm a possible supervisor & I'd be keen to support projects on sociolinguistics-AI, e.g., accent bias in AI, language+gender/sexuality+AI.
www.responsiblenlp.org
We have a number of fully funded PhD studentships in "Designing Responsible Natural Language Processing". I'm a possible supervisor & I'd be keen to support projects on sociolinguistics-AI, e.g., accent bias in AI, language+gender/sexuality+AI.
www.responsiblenlp.org
Ironically, the “AI scientist”-written paper that was accepted to an ICLR workshop was about LSTMs and the paper didn’t cite Schmidhuber. AI scientists should have a Schmidhuber agent that verifies these important details in the future.
October 7, 2025 at 9:23 PM
Ironically, the “AI scientist”-written paper that was accepted to an ICLR workshop was about LSTMs and the paper didn’t cite Schmidhuber. AI scientists should have a Schmidhuber agent that verifies these important details in the future.
Reposted by Bálint Gyevnár
Never ask a man his age, a woman her salary, or GPT-5 whether a seahorse emoji exists
September 6, 2025 at 1:08 PM
Never ask a man his age, a woman her salary, or GPT-5 whether a seahorse emoji exists
Reposted by Bálint Gyevnár
Adam Raine, 16, died from suicide in April after months on ChatGPT discussing plans to end his life. His parents have filed the first known case against OpenAI for wrongful death.
Overwhelming at times to work on this story, but here it is. My latest on AI chatbots: www.nytimes.com/2025/08/26/t...
Overwhelming at times to work on this story, but here it is. My latest on AI chatbots: www.nytimes.com/2025/08/26/t...
A Teen Was Suicidal. ChatGPT Was the Friend He Confided In.
www.nytimes.com
August 26, 2025 at 1:01 PM
Adam Raine, 16, died from suicide in April after months on ChatGPT discussing plans to end his life. His parents have filed the first known case against OpenAI for wrongful death.
Overwhelming at times to work on this story, but here it is. My latest on AI chatbots: www.nytimes.com/2025/08/26/t...
Overwhelming at times to work on this story, but here it is. My latest on AI chatbots: www.nytimes.com/2025/08/26/t...
"I have wait a long time for this moment, my little red friend." - Emperor Palpatine, probably
Hollow Knight: Silksong will be available September 4 on all platforms and day one on Xbox Game Pass!
Watch the release trailer: youtu.be/6XGeJwsUP9c
Watch the release trailer: youtu.be/6XGeJwsUP9c
Title: Hollow Knight: Silksong - Release Trailer
YouTube video by Team Cherry
youtu.be
August 25, 2025 at 12:10 PM
"I have wait a long time for this moment, my little red friend." - Emperor Palpatine, probably
Having been playing The Art of Fauna while procrastinating, and it is the most relaxing game with beautiful illustrations of animals. I cannot recommend it enough: theartof.app/fauna/
The Art of Fauna: Cozy Puzzles
Discover the wonders of nature with this cozy puzzle game. Download Now!
theartof.app
August 25, 2025 at 12:06 PM
Having been playing The Art of Fauna while procrastinating, and it is the most relaxing game with beautiful illustrations of animals. I cannot recommend it enough: theartof.app/fauna/
For a week, we visited @iyadrahwan.bsky.social's Center for Humans and Machines in Berlin, where we have met an incredible array of interdisciplinary researchers.
Special thanks to @neeleengelmann.bsky.social for hosting us and @alice-ross.bsky.social for organising the trip from the start!
Special thanks to @neeleengelmann.bsky.social for hosting us and @alice-ross.bsky.social for organising the trip from the start!
August 16, 2025 at 8:24 PM
For a week, we visited @iyadrahwan.bsky.social's Center for Humans and Machines in Berlin, where we have met an incredible array of interdisciplinary researchers.
Special thanks to @neeleengelmann.bsky.social for hosting us and @alice-ross.bsky.social for organising the trip from the start!
Special thanks to @neeleengelmann.bsky.social for hosting us and @alice-ross.bsky.social for organising the trip from the start!
Have been neglecting Bluesky recently, so I am happy to share a big update 🎉
I will join @cmu.edu as a postdoc in September working with the incomparable @atoosakz.bsky.social and Nihar Shah on understanding risks from LLM co-scientists. If you are in Pittsburgh, I would love to connect!
I will join @cmu.edu as a postdoc in September working with the incomparable @atoosakz.bsky.social and Nihar Shah on understanding risks from LLM co-scientists. If you are in Pittsburgh, I would love to connect!
August 16, 2025 at 8:12 PM
Have been neglecting Bluesky recently, so I am happy to share a big update 🎉
I will join @cmu.edu as a postdoc in September working with the incomparable @atoosakz.bsky.social and Nihar Shah on understanding risks from LLM co-scientists. If you are in Pittsburgh, I would love to connect!
I will join @cmu.edu as a postdoc in September working with the incomparable @atoosakz.bsky.social and Nihar Shah on understanding risks from LLM co-scientists. If you are in Pittsburgh, I would love to connect!
Reposted by Bálint Gyevnár
This thread is incredible.
Jason ✨👾SaaStr.Ai✨ Lemkin (@jasonlk)
.@Replit goes rogue during a code freeze and shutdown and deletes our entire database
xcancel.com
July 20, 2025 at 3:01 PM
This thread is incredible.
VCs: oh yeah, developers with AI go brrrrr...
METR: Uhm, actually they are 19% slower with AI
metr.org/blog/2025-07...
METR: Uhm, actually they are 19% slower with AI
metr.org/blog/2025-07...
Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
We conduct a randomized controlled trial to understand how early-2025 AI tools affect the productivity of experienced open-source developers working on their own repositories. Surprisingly, we find th...
metr.org
July 11, 2025 at 12:21 PM
VCs: oh yeah, developers with AI go brrrrr...
METR: Uhm, actually they are 19% slower with AI
metr.org/blog/2025-07...
METR: Uhm, actually they are 19% slower with AI
metr.org/blog/2025-07...
The Edinburgh RL & Agents reading group is back in action after a hiatus. Previous speakers come for all across the world, including DeepMind, CMU, Oxford, NUS, etc. Sign up for some great discussions about cutting-edge RL and agents research.
Hello world! This is the RL & Agents Reading Group
We organise regular meetings to discuss recent papers in Reinforcement Learning (RL), Multi-Agent RL and related areas (open-ended learning, LLM agents, robotics, etc).
Meetings take place online and are open to everyone 😊
We organise regular meetings to discuss recent papers in Reinforcement Learning (RL), Multi-Agent RL and related areas (open-ended learning, LLM agents, robotics, etc).
Meetings take place online and are open to everyone 😊
July 10, 2025 at 2:09 PM
The Edinburgh RL & Agents reading group is back in action after a hiatus. Previous speakers come for all across the world, including DeepMind, CMU, Oxford, NUS, etc. Sign up for some great discussions about cutting-edge RL and agents research.