1/2 🧵 @lucamb.bsky.social "To resume our previous conversation, I think you will see things differently if you consider your perception of an apple not just as a copy inside your head, but an actual manifestation of the real apple in front of you @itaisher.bsky.social

No idea what heights you’re chasing or what the next big milestone is, but you’re only halfway up 🤗🎧 open.spotify.com/episode/5k7w...

Dealing with a resubmission? Who gets borderline rates check this convo branch. It's also why I stay in social media, engage with folks as one would do in a lab/conference/etc without physical presence.

📢 FLAIR-HUB paper is published in ISPRS Journal of Photogrammetry and Remote Sensing 2026 !!!! kwnsfk27.r.eu-west-1.awstrack.me/L0/https:%2F...

@lucamb.bsky.social "The biggest argument against general intelligence in my mind is that in the last two years frontier models got enormously better at math and coding and rather worse on writing, having a conversation and understanding sense of humor"

In another episode of (AI) researchers out of the lab 🧵

Seems that @chriswolfvision.bsky.social & I are biking synced 😂

😂 heard that feeling when I did my internship in SF a friend of mine told me, I only date Europeans or Latinas. Dating protocol here is so bizarre (1. Get the number; 2. Text 𝗮𝗳𝘁𝗲𝗿 𝟯-𝟱 𝗱𝗮𝘆𝘀 otherwise ur interest is perceived as desperation; ...) No wonder why AI gf is a biz model in the Bay area 😉

Claude lovers, are you happy that your data is fueling Elon's SpaceX biz? Curious how their moral/ethical stand age

Yes.

Software Engineer position open at U. Copenhagen and @aicentre.dk on the Narratives project, co-advised by @ysiglidis.bsky.social, @mariaa.bsky.social (CU Boulder), and me: candidate.hr-manager.net/ApplicationI...

@andrewgwils.bsky.social "There's something refreshing about reading a pre-2023 paper and being absolutely certain that it was all human generated. As much as I will do research on LLMs, I will never use them to write a word of my papers."

@yoavgo.bsky.social "so now ACL requires also *children* who attend the event with their parents to register as attendees?? at the students rate??? this is truly becoming more and more adversarial, and this time it is also adversarial towards populations they traditionally aim to support."

Quizas es ego o irrespeto... Me importa una verga Si alguien con privilegio dicd qué eres el irrespetuoso en redes sociales, riete en su espalda. Las redes sociales es donde se hace el activismo & periodismo hoy. Los activistas & periodistas de verdad NO estamos al servicio del poder.

Why to follow the rich instead of makers (OSS maintainers & contributors, etc.) & soldiers? Especially as big corp accounts are likely gonna get automated by AI in the short-term

@damienteney.bsky.social"Alternatively, practice self-consistency. ✅Paper accepted? Don’t get too excited, it's partly luck. ❌Paper rejected? Don’t get discouraged, it's partly luck. Just don't anchor your emotions on external validation. Enjoy the process and work on something you'll be proud of."

Which game is the CV/AI community playing? If it's the infinite game, pray that PAMI-TC & ECVA boards think different to him. Western history shows that transparency of institutions is key over abusive institutions (e.g., ruled by benevolent aristocracy/oligarchy).

Rebuttal phase is a weird game. Not a clean “truth correction” phase. More like football + Perudo/Cubilete with reviewers. Sometimes your paper misses the goal because it was weak. Sometimes it hits the post. Sometimes another player blocks it and everyone shouts “unlucky.” 1/2

CHI was had every kind of AI for X, and that made me feel we lack a robust discussion of the limitations of AI as a chimmunity. What are things AI should never be used for, no matter how good the performance?

What % of the NLP papers measure their impact in the real world? This paper proposes an impact evaluation of NLP models or systems for real-world usage, changing the research culture of NLP to focus more on real-world impact and less on SOTA-chasing: doi.org/10.1162/COLI... @ehudreiter.bsky.social

Correction: The correct overleaf link to the #ECCV2026 rebuttal template is www.overleaf.com/read/ctytcjt...

Can't have one without the other

Which is Trump's most successful war campaign? 1. Economic war vs 🇨🇳 2. Wars against 🇮🇷 3. Cultural war against 🇪🇺 4. Cultural war against 🇺🇸 citizens ... youtu.be/UNhL1IJH0kk

(>=6th visit to) Venice: magical city 🇮🇹 Bacaro in "Bacarando Corte dell'Orso": never again. Pls upvote my review 🙏👇 maps.app.goo.gl/nb3hFXLC3QLE... Used to recommend it, until today’s “gentilissimo” bald staff member decided to serve arrogance, humiliation & discriminatory comments alongside ...

🔖🎧... Looking fwd to hear it on my next 🚵🏽

This interview provides an overview of my thoughts on how AI is developing in Europe vs. other parts of the world:

Maybe I should I do an AI horoscope for the new religion: 𝘼𝙄𝙣𝙞𝙨𝙢 In Skynet we trust 🫡🙏🏼

Your AI assistant says a lot about you: ☮️ Claude: tasteful hipster with disposable income 💻 ChatGPT: default human, probably has 47 tabs open 🧔🏽‍♀️ Gemini: Google Docs power user 👵🏼 Copilot: corporate survivor 🤜🏼 Grok: replies “interesting” under Elon posts 📱 Meta AI: accidentally clicked it on WhatsApp

👀💬 Few months ago a similar discussion happened in a Slack group. I was the only supporting to keep submission free. Thus, ppl are OK to pay 🤯 Demographics: 10-20/>50 active/- members. >80% researchers, <30% industry folks, most publishing & submitting onto top-tier venues

AI researchers can be wildly irreverent about ideas, then strangely obedient around incentives. Maybe the real skill is knowing when you’re working with an org vs working for it. Different games. Different risks. Different freedom.

Veneto 🇮🇹 la terra dil prosecco 🍾🥂 You find them for >= 1.75€ even the shitty one sold in 🇬🇧 >= 6€

Some cyborgs are too keyboard wired that customisation is worth even if audio & speech becomes the UI. Reminder to update our shortcuts.

& I/you aren't doing as much (That's what social media leaves onto ppl with a weak mind)

Witch hunting & judging is back. ⁉️ Does it reminds you when we/ppl lost social (tech) battles (e.g., deep learning conquering AI fields, etc.)? Memory pills about unspoken internet rule 1. You don't own a reply to anyone 2. Anyone can broadcast as long as it's NOT flagged as malicious/spam ...

The kind of stuff that it's sticky as polen 🦋 youtu.be/Ykvf3MunGf8

👏🏼💪🏼🇨🇴 Grande!

@iclr-conf.bsky.social "ICLR 2026 is almost there! We have 6 exciting keynotes covering a range of areas from machine learning to robotics, neuroscience and AI for science: Maja Matarić, Max Welling, Percy Liang, Katie Bouman, Karen Adolph, Pablo Arbeláez See you all soon! #ICLR2026"

The AI for Peace Workshop schedule is live! Full program: aiforpeaceworkshop.github.io/schedule/

@egrefen.bsky.social "When looking at deep learning profiles, one of the most obvious tells between a mediocre and a great candidate is whether they prefer chocolate or vanilla ice cream."

What are your all time favorite textbooks? Here are a few of mine.

@ericxw.bsky.social "When looking at faculty profiles, one of the most obvious tells between a mediocre and great candidate is whether they use PowerPoint or Google Slides." Echoing @fchollet (Keras' creator) take between Jax & PyTorch

Heading to @cvprconference.bsky.social? Please reserve your hotel room within the official CVPR hotel block as soon as possible to secure the conference rate. More info: cvpr.thecvf.com/Conferences/...

NeurIPS 2026 is soliciting competition proposals on topics of interest to the NeurIPS community. Read the call for competitions for more information neurips.cc/Conferences/...

Strevens knowledge machine in action! Unifying Popper's penalizing Pendulum; and Kuhn's Paradigmatic Pendulum aligner. www.strevens.org/scientia/

With more people embracing AI agents, students and others will outsource part or all of their writing to LLMs. Golden rule: do not use peers as unpaid labelers for your automation. If AI touched it, proofread it before sharing, and leave edit notes where needed.

NeurIPS encourages and benefits from a diversity of papers and ideas, which can be developed in many different ways. This year, Main Track submissions can select a Contribution Type, including General, Theory, Use-Inspired, Concept & Feasibility, and Negative Results.

@juliekallini.bsky.social: " 1/ "New tokenizer" does not imply "new base model," and "new base model" is not the simplest explanation. There are much simpler explanations that fit Anthropic's public description of Opus 4.7 equally well.

Sometimes advisor/PM/coach helps steer the boat, not as the captain, but as instrument. Other times is to bother the crew: - remember the commitments, - remind that DEI, broad participation, etc. are not grant-writing decorations, - ask the uncomfortable questions: why this, why now, and for whom?

I thought I would do a thread, but honestly the post is so good: kyutai.org/blog/2026-04... It explains "One View Is Enough! Monocular Training for In-the-Wild Novel View Generation" arxiv.org/abs/2603.23488 done in colab with the smart people at kyutai

@csprofkgd.bsky.social "#KostasThoughts: Perhaps we’ve gone too far with requiring ALL authors to be in the reviewer pool in computer vision conferences and instead we should consider a per paper nomination system 🤔"