Christian Wolf's Avatar

Christian Wolf

@chriswolfvision.bsky.social

Principal Scientist at Naver Labs Europe, Lead of Spatial AI team. AI for Robotics, Computer Vision, Machine Learning. Austrian in France. https://chriswolfvision.github.io/www/

6,300 Followers  |  814 Following  |  2,005 Posts  |  Joined: 08.02.2024
Posts Following

Posts by Christian Wolf (@chriswolfvision.bsky.social)

Example: we have several CVPR 2021 papers (so, submitted before ChatGPT) with the name "VQA" in the title which I think would fly much better if I had replaced them with "vision-language" since "VQA" as a task became too specific.

02.03.2026 21:27 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

He absolutely did! He ranted for some time that "China screws over" the West by producing technology that does not work and they sell it to the gullible. He is insane.

02.03.2026 21:24 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I am so confused! Didn't the president of the US recently state in his Davos speech that China didn't actually use their own solar panels and that they only sold them to "stupid" foreign customers because they didn't believe themselves in this ... "scam"? 🀭

02.03.2026 20:50 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 4    πŸ“Œ 0

Yes, Feb 3rd ...

02.03.2026 19:54 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

Pi day is actually Feb 2nd, since 3.14. = 3.2. when you wrap around months modulo 12 πŸ˜‰

02.03.2026 19:50 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

Whenever I come up with paper names I absolute love them. Then one year later I absolute hate them and with hindsight it was a very bad choice. Every single time.

02.03.2026 12:55 β€” πŸ‘ 9    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0
Post image Post image Post image

I have added a new tutorial on discrete diffusion models:
github.com/gpeyre/ot4ml

01.03.2026 14:40 β€” πŸ‘ 51    πŸ” 15    πŸ’¬ 0    πŸ“Œ 0
Post image

Youtube recommendation quality did not yet reach AGI level.

01.03.2026 10:30 β€” πŸ‘ 8    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

This looks like the video from the original densepose (non wifi) from 2018.

01.03.2026 08:01 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 2    πŸ“Œ 0

I was a frequent user of Paracetamol for head aches until we started monitoring the CO2 / O2 levels in our house. Not anymore. And I heard similar stories around me. Get some fresh air πŸ˜€

28.02.2026 15:40 β€” πŸ‘ 9    πŸ” 1    πŸ’¬ 2    πŸ“Œ 0

In your face, the bitter lesson! πŸ˜‚

28.02.2026 10:23 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Bonjour Service Free @free1337.bsky.social, je suis abonné à votre box depuis 20 ans. Si je reçois 2 appels de suite me proposant votre service et je leur dis gentiment que c'est pas le peine de m'appeler de manière répétitive, et puis on me raccroche au nez sans commentaire, trouvez vous ça normal?

27.02.2026 18:13 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

😬

From: "Wired: Wallstreet has AI Psychosis"
www.wired.com/story/wall-s...

27.02.2026 17:56 β€” πŸ‘ 7    πŸ” 1    πŸ’¬ 0    πŸ“Œ 1

My Cousin had an MSX and was VERY unhappy. My uncle had been told by the vendor that it was 'much better'. But there was no software...

27.02.2026 17:11 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

You know strange people!

27.02.2026 15:33 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

A variant of Godwin's law: as a discussion continues with a Gen-X-er, the probability of mentioning a Commodore 64 will approach one.

27.02.2026 15:13 β€” πŸ‘ 2    πŸ” 2    πŸ’¬ 1    πŸ“Œ 0

A variant of Godwin's law: as a discussion continues with a Gen-X-er, the probability of mentioning a Commodore 64 will approach one.

27.02.2026 15:13 β€” πŸ‘ 2    πŸ” 2    πŸ’¬ 1    πŸ“Œ 0
Post image

I am not that old but I learned to code BASIC and then MOS6510 assembly with a dataset drive. But quickly switched to the Commodore 1541 Floppy

Everything was simple enough to understand everything down to the silicon. I loved it.

27.02.2026 15:11 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Post image

The current generation of computer science students is probably the golden one: they are the first be to trained to use AI, and (arguably?) the last to be forced to learn how to code manually, without it.

Like the engineers using slide rules to calculate return trajectories in Apollo 13: love them.

27.02.2026 13:23 β€” πŸ‘ 10    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Post image

For me this was the absolute pinnacle of architecture. From then on, it went all downhill.

Reminds my of the Villa Wagner I by Austrian Otto Wagner, built in 1886 in Vienna.

27.02.2026 12:31 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Video thumbnail

Questions you get get at Austrian immigration:
"When you are in a rush in Austria, what is a common farewell?"

(From the Foil Arms and Hog YT Channel)

27.02.2026 12:21 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

For NeurIPS, you need to opt-in if the paper is rejected, as far as I know?

26.02.2026 18:28 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

It is CVPR but not ICLR, so while it uses openreview, the reviews are not published. The usage of Openreview does not imply that, it is configurable.

Also, even for ICLR (where reviews are public) this would not change anything, as even there reviews are kept public for withdrawn papers.

26.02.2026 18:21 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

Since 2 or 3 conference editions I now see authors withdrawing papers _after_ the decisions have been communicated. Several papers of my batch have been rejected and then later withdrawn by the authors. Why?

26.02.2026 16:39 β€” πŸ‘ 6    πŸ” 1    πŸ’¬ 3    πŸ“Œ 0

You were so ahead of your time.

26.02.2026 06:56 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Deconstructing Nvidia’s Vera Rubin β€” The Successor To Blackwell That’s 10x More Efficient
YouTube video by CNBC Deconstructing Nvidia’s Vera Rubin β€” The Successor To Blackwell That’s 10x More Efficient

The next generation of Nvidia infrastructure "Vera Rubin" will consume 220 kilowatt and weights 2 tonnes ... per rack 😬

But this looks exciting. High density. Watercooling is now an obligation. Easy swapping of components.

www.youtube.com/watch?v=ZkPU...

25.02.2026 23:03 β€” πŸ‘ 2    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0

Cool !

25.02.2026 21:16 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Pour du bon? Tu quittes le Liris ?

25.02.2026 21:05 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

Migration completed! Took about 3 hours, but worked without any trouble.

This account is now hosted in Europe!

24.02.2026 20:17 β€” πŸ‘ 15    πŸ” 2    πŸ’¬ 2    πŸ“Œ 0

OK I will put this on my to do list.

25.02.2026 18:22 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0