Americans who like my posts at 5:00 am Eastern: Who hurt you 😂
Claas Voelcker
@cvoelcker.bsky.social
For professional, see https://cvoelcker.de If I seem very angry, check if I have been watered in the last 24 hours. Now 🇺🇸 flavoured, previously available in 🇨🇦 and 🇩🇪
Tell me there are no queer people in your life without telling me there are no queer people in your life... My right to marry in my home country is younger than my relationship with my husband (this is obviously a dunk on the screenshot person, not the bsky person, just to be safe)
Revan@revan.eurosky.social · 19h agoSo this is basically one of the entry points to becoming a reactionary
It's only research when it comes from the NeurIPS region of AAA conferences, otherwise it's just sparkling failure
Noone will like this, but I just got a bunch of negative reviews for a paper and I'm NOT mad! The critique was reasonable, the improvement suggestions feasible, and everybody seems to have understood what we did correctly. Peer review can work!
Abolish awards! Recognize opportunity and priviledge what they are and built resilient structures that favor collaboration and shared recognition!
Dr. Katja Thieme (she/they)@katjathieme.bsky.social · yesterdayA humble academic proposal: let's not have stars. Everyone hired as an academic at a university is there to teach younger generations, to make new knowledge, to have that knowledge critically evaluated by others, and to act on collective insights. We do not need to designate some people as stars.
Every time I land in Germany, I’m reminded again that I’m comparatively tiny here 😂
There are so many major problems with the "Go big" advice, and I would love it if senior people started addressing them: - "Go big" is close to impossible without practising in the small first - Not publishing frequently locks you out from almost all networking to find good collaborators
Clément Canonne@ccanonne.github.io · 4d ago"We're all worried," as what it means to do research (in my field, Theoretical CS) seems to be shifting, and shifting fast. What to do? Senior researchers must lead by example, knowing that not everything will pan out. What I'm suggesting below may not work everywhere, but here's my own advice: 1/
"Write dumb rebuttals, get dumb replies" I think the LLM era has made me appreciate brevity sooooooooooo much more. I get it now, I apologize, everything should just be a single line.
Fun thing about bluesky reach: everybody complains that bluesky has no reach compared to twitter, but we are at 1/3 of the likes here compared to over there... Given the relative relevance of the platforms, I count that as a win for us here.
Claas Voelcker@cvoelcker.bsky.social · 6d agoThis is a robot failing to grasp a ball. Almost every robot lab produces clips like this daily… and almost all of them get thrown away. This is the most abundant but underused resource in robot learning. We’re collecting all of it now as “OopsieData”, please join us at oopsie-data.com! (1/13)
Big shoutout to Maria Attarian in @igilitschenski.bsky.social's lab for producing the nicest oopsie data point we have so far in terms of resolution :)
Claas Voelcker@cvoelcker.bsky.social · 6d agoThis is a robot failing to grasp a ball. Almost every robot lab produces clips like this daily… and almost all of them get thrown away. This is the most abundant but underused resource in robot learning. We’re collecting all of it now as “OopsieData”, please join us at oopsie-data.com! (1/13)
Turn your robot errors into a paper!
Claas Voelcker@cvoelcker.bsky.social · 6d agoThis is a robot failing to grasp a ball. Almost every robot lab produces clips like this daily… and almost all of them get thrown away. This is the most abundant but underused resource in robot learning. We’re collecting all of it now as “OopsieData”, please join us at oopsie-data.com! (1/13)
Ah, this is great. I love it. I love the Tolstoy quote too. I like every part of this
Claas Voelcker@cvoelcker.bsky.social · 6d agoThis is a robot failing to grasp a ball. Almost every robot lab produces clips like this daily… and almost all of them get thrown away. This is the most abundant but underused resource in robot learning. We’re collecting all of it now as “OopsieData”, please join us at oopsie-data.com! (1/13)
This is a robot failing to grasp a ball. Almost every robot lab produces clips like this daily… and almost all of them get thrown away. This is the most abundant but underused resource in robot learning. We’re collecting all of it now as “OopsieData”, please join us at oopsie-data.com! (1/13)
Und selbst wenn es wirklich ein Fall von islamistischem Terror ist ist das wieder nur extrem convenient um damit dann weiteren Rassismus und „Ausländer“hass zu schüren anstatt sich darüber Gedanken zu machen wie man auch hier den Boden für weiteren Hass selbst ebnet.
OK, maybe I missed the day and it will be tomorrow 😂 Sorry, multi-lab coordination is a fickle beast!
Claas Voelcker@cvoelcker.bsky.social · last wk.Huge news dropping tomorrow #VaguePostingAIStyle
Random writing pet peeve (and, yes, I admit I'm super opinionated): learned cannot be used in the passive instead of trained. Model are trained, models are not learned.
A new method from @utoronto.ca computer scientists trains AI faster and more reliably using less data and computing power. This could help bring smarter robots and autonomous systems into everyday life. Learn more: uoft.me/cyu @cvoelcker.bsky.social @igilitschenski.bsky.social
Has anybody studied where the weird viby language of LLMs comes from? Why do they always sound like that one stoned friend at 2am who has seen the time knife?
Billy Bishop has a bunch of really cheap flights to East Coast locations going for it, I'm not quite sure that "only really wealthy people use it" is a solid take...
John Michael McGrath@jm-mcgrath.bsky.social · last wk.re: "fringe group" — the vast majority of GTA residents do not, and likely never will, fly out of Billy Bishop, this was always special pleading from a narrow interest group, they just happen to be extremely wealthy
After #neurips2026, can we agree forced review was a bad experiment?
OK, seeing the other reviews for papers I reviewed for NeurIPS makes me offended for the authors sake...
OK, maybe I should stop clowning on safety research that is mostly competitive sci-fi story writing. The sci-fi stories have escaped containment.
Doing something different in your code than what you wrote in your paper should come with a much bigger generational aura loss! github.com/galilai-grou...
LeJEPA loss computed on MLP-projections vs embeddings themselves · Issue #17 · galilai-group/lejepa
Hi! I was going over the code in your MINIMAL.md notes, and noticed that the LeJEPA loss is computed on non-linear projections of a ViT backbone's output, but the probe uses the actual output of th...
github.com
Reciprocal reviewing needs to be tuned to the amount of papers you submit. Currently I think I fall into the most punished category: I submit roughly 1-2 papers per conference, and I always review. If I submitted 4x the amount, I would still have to review the same amount.
Be me, don’t use agents, still don’t understand what the paper is about, be pretty sure the authors are making it up, reread it thrice and realize you were an idiot!
Vilém Zouhar@zouharvi.bsky.social · 2w agoSeveral reviewers with no technical background told me they reviewed for NeurIPS using agents, stating they didn't understand what the work is about. Asked how they could trust the review. They said when the paper hits the socials, the people will catch the errors.
Today for some reason A LOT of people are failing the basic empathy test...
Claas Voelcker@cvoelcker.bsky.social · 3w agoIf we wanted more people to like and embrace AI, maybe we shouldn't have left a bunch of the worst possible people in charge? And maybe we should all collectively show a tiny modicum of empathy to people worried about their livelihoods and meaning in life... I'm really just spit-balling...