Nicolas Pugeault's Avatar

Nicolas Pugeault

@pugeault.bsky.social

Teaching blind robots to see... I am a scientist and Reader at the University of Glasgow, interested in AI, Computer Vision & Robotics.

1,024 Followers  |  1,436 Following  |  16 Posts  |  Joined: 14.08.2024  |  2.9432

Latest posts by pugeault.bsky.social on Bluesky

Post image

This is probably the strongest I've put it, but I think it's perfectly accurate.

Europe (pick your institutional configuration) is at war with Russia without a reliable superpower partner, and needs to act like it.

28.11.2025 17:37 โ€” ๐Ÿ‘ 14    ๐Ÿ” 3    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 0

And MacDonalds 2 millions. Not sure those numbers mean anything

27.11.2025 18:50 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Preview
'BBC Bias' Memo Was Authored by Lobbyist Tied to Pro-Trump Tech Giants The leaked memo that fuelled Trumpโ€™s attack on the BBC was written by a lobbyist at a firm paid by US tech giants tied to the President

๐ŸงตOn our revelations today about the BBC Coup and the Prescott Dossier.

The leaked โ€˜BBC Biasโ€™ memo Trump used to attack the BBC was authored by Michael Prescott โ€” a Hanover lobbyist paid by US tech/media giants tied to Trump. Full story: bylinetimes.com/2025/11/11/b... 1/12

11.11.2025 21:19 โ€” ๐Ÿ‘ 819    ๐Ÿ” 610    ๐Ÿ’ฌ 33    ๐Ÿ“Œ 90

Lots of stuff here that you won't read in most of the legacy media which seems completely uninterested in the backgrounds, associations and possible motivations of those behind the BBC crisis.

Why was the BBC having its homework marked by a lobbyist who hasn't worked as a journalist for 20 years?

11.11.2025 21:25 โ€” ๐Ÿ‘ 400    ๐Ÿ” 194    ๐Ÿ’ฌ 6    ๐Ÿ“Œ 3

Just absolutely classic BBC. You've just broadcast the most successful programme of the year, uniting Gen Z kids online and Boomers on broadcast in a return to appointment television. And instead of celebrating, your DG reigns due to a made-up right-wing scandal.

09.11.2025 20:10 โ€” ๐Ÿ‘ 1839    ๐Ÿ” 361    ๐Ÿ’ฌ 50    ๐Ÿ“Œ 12
Post image Post image

Updated thread on The Lost King film:
Iโ€™m digging over old ground but I have to challenge Steve Coogan's line. He claims itโ€™s all true (as has Philippa Langley): Langley will be celebrated for her achievement, Richard Taylor & his personal gripes will be forgotten
So whatโ€™s the real story? 1/15

31.10.2025 10:53 โ€” ๐Ÿ‘ 124    ๐Ÿ” 49    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 14

A very good thread, which also makes the connection between the conspiratorial / antiestablishment viewpoint of Philippa Langley and Steve Coogan and the assault on science coming from RFK Jr and the White House.

Itโ€™s the same syndrome manifesting in different ways.

04.11.2025 16:39 โ€” ๐Ÿ‘ 17    ๐Ÿ” 3    ๐Ÿ’ฌ 2    ๐Ÿ“Œ 0

bsky.app/profile/holy...

09.10.2025 15:52 โ€” ๐Ÿ‘ 0    ๐Ÿ” 1    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Je suis un peu... dรฉpitรฉe de l'idรฉe " faut surtout pas culpabiliser les abstentionnistes."

D'un cotรฉ, je vois l'idรฉe : effectivement, ce ne sont PAS les premiers responsables de cette situation et oui, voter blanc, c'est un vote comme un autre.

08.10.2025 12:52 โ€” ๐Ÿ‘ 22    ๐Ÿ” 4    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 2

We can huddle up and lob bombs between camps, calling each other techbros and luddites, if we want... but there's acres and acres of middle ground on this stuff as well.

I do think the vibe shift around tech broadly makes this particular middle ground almost impossible to occupy though.

07.10.2025 18:18 โ€” ๐Ÿ‘ 73    ๐Ÿ” 5    ๐Ÿ’ฌ 5    ๐Ÿ“Œ 1
Post image Post image Post image Post image

Benchmarking Egocentric Visual-Inertial SLAM at City Scale

Anusha Krishnan, Shaohui Liu, @pesarlin.bsky.social , Oscar Gentilhomme, David Caruso, Maurizio Monge, Richard Newcombe, Jakob Engel, @marcpollefeys.bsky.social
tl;dr: in title. Aria-SLAM rules, ORB-SLAM3 2nd.
arxiv.org/abs/2509.26639
1/2

01.10.2025 10:45 โ€” ๐Ÿ‘ 3    ๐Ÿ” 1    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 0

My family came here 50 years ago. They've worked, paid taxes, raised families, been good citizens.

My many colleagues and friends who were either born elsewhere, or their parents were, the same.

It's so offensive to be told you're here on sufferance.

29.09.2025 09:20 โ€” ๐Ÿ‘ 1211    ๐Ÿ” 339    ๐Ÿ’ฌ 39    ๐Ÿ“Œ 11

An exceptionally tasteless headline from a publication that should really know better, especially at a time like this. But I suppose it resonates with a sensationalist Anglo crowd and their deeply rooted prejudices.

26.09.2025 18:53 โ€” ๐Ÿ‘ 242    ๐Ÿ” 42    ๐Ÿ’ฌ 23    ๐Ÿ“Œ 2
Jimmy Kimmel is Back!
YouTube video by Jimmy Kimmel Live Jimmy Kimmel is Back!

Full Jimmy Kimmel opening monologue from last night here:

www.youtube.com/watch?v=c1tj...

24.09.2025 07:04 โ€” ๐Ÿ‘ 437    ๐Ÿ” 134    ๐Ÿ’ฌ 18    ๐Ÿ“Œ 15
Post image Post image Post image

ProDyG: Progressive Dynamic Scene Reconstruction via Gaussian Splatting from Monocular Videos

Shi Chen, Erik Sandstrรถm, Sandro Lombardi, Siyuan Li, @martin-r-oswald.bsky.social

tl;dr: Splat-SLAM+SAM2->tracking; Motion Scaffolds+GS->dynamic

arxiv.org/abs/2509.17864

23.09.2025 19:12 โ€” ๐Ÿ‘ 3    ๐Ÿ” 3    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

- "Les chercheurs sont dรฉconnectรฉs, enfermรฉs dans leur tour d'ivoire !"

Chercheurs : descendent dans la citรฉ.

- "Oui mais pas comme รงa !"

ร€ chaque fois ๐Ÿ™ƒ

23.09.2025 06:12 โ€” ๐Ÿ‘ 217    ๐Ÿ” 57    ๐Ÿ’ฌ 7    ๐Ÿ“Œ 1

You can be a โ€œcar guyโ€ and not like car dependency. You can be a โ€œcar guyโ€ and get that too many cars in cities is bad for everyone, including drivers. You can be a โ€œcar guyโ€ and be tired of lies & manipulations like โ€œthe war on cars.โ€ You can be a โ€œcar guyโ€ and know more choice means more freedom.

22.09.2025 20:26 โ€” ๐Ÿ‘ 1161    ๐Ÿ” 193    ๐Ÿ’ฌ 73    ๐Ÿ“Œ 42
Post image Post image Post image Post image

Tendency: render non-visual data/parameters into images fed to neural networks.

Joint angles: GENIMA genima-robot.github.io
Trajectories: VINT arxiv.org/abs/2306.14846
Object pose: MFOS arxiv.org/abs/2310.01897
Waypoints: PIVOT arxiv.org/abs/2402.07872

(Repost+update of a tweet from 2024)

22.09.2025 14:07 โ€” ๐Ÿ‘ 11    ๐Ÿ” 3    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 0

It was obvious when EU citizens had the terms of their settlement here retrospectively revoked that a precedent was being set and that, in time, the argument would move on to other categories of migrant, always tending towards a purity test of who is โ€˜indigenousโ€™. Unless resisted.

22.09.2025 09:17 โ€” ๐Ÿ‘ 375    ๐Ÿ” 130    ๐Ÿ’ฌ 16    ๐Ÿ“Œ 8

I've been thinking a bit about the history of cyber punk and how the vision of a virtual reality turned out to be true, but in a very different way. The assumption was that there would still be a concensus, reference reality but what has actually happened is reality has become virtual to individuals

21.09.2025 16:26 โ€” ๐Ÿ‘ 6    ๐Ÿ” 2    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 0

Annnnnd it's a reject!

Scale is a religion and if you go against it, you're a heretic and you should burn, "despite [the reviewers] final ratings".

But scale is still not necessary!

Side note: First time swinging reviews up (from 2,2,4,4 to 2,4,4,5) does not get the paper accepted. Strange days.

18.09.2025 17:04 โ€” ๐Ÿ‘ 18    ๐Ÿ” 3    ๐Ÿ’ฌ 4    ๐Ÿ“Œ 0
Preview
'Staggering' childcare costs swallowing three-quarters of mum's salary Bernie Carranza is among those calling for a new funding model that would cap childcare costs at 5% of a family's income.

This is such a good article about childcare - the high cost and postcode lottery of availability is holding back many families in Edinburgh South West.

news.stv.tv/east-central...

11.09.2025 08:01 โ€” ๐Ÿ‘ 0    ๐Ÿ” 1    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Abstract: Under the banner of progress, products have been uncritically adopted or
even imposed on users โ€” in past centuries with tobacco and combustion engines, and in
the 21st with social media. For these collective blunders, we now regret our involvement or
apathy as scientists, and society struggles to put the genie back in the bottle. Currently, we
are similarly entangled with artificial intelligence (AI) technology. For example, software updates are rolled out seamlessly and non-consensually, Microsoft Office is bundled with chatbots, and we, our students, and our employers have had no say, as it is not
considered a valid position to reject AI technologies in our teaching and research. This
is why in June 2025, we co-authored an Open Letter calling on our employers to reverse
and rethink their stance on uncritically adopting AI technologies. In this position piece,
we expound on why universities must take their role seriously toa) counter the technology
industryโ€™s marketing, hype, and harm; and to b) safeguard higher education, critical
thinking, expertise, academic freedom, and scientific integrity. We include pointers to
relevant work to further inform our colleagues.

Abstract: Under the banner of progress, products have been uncritically adopted or even imposed on users โ€” in past centuries with tobacco and combustion engines, and in the 21st with social media. For these collective blunders, we now regret our involvement or apathy as scientists, and society struggles to put the genie back in the bottle. Currently, we are similarly entangled with artificial intelligence (AI) technology. For example, software updates are rolled out seamlessly and non-consensually, Microsoft Office is bundled with chatbots, and we, our students, and our employers have had no say, as it is not considered a valid position to reject AI technologies in our teaching and research. This is why in June 2025, we co-authored an Open Letter calling on our employers to reverse and rethink their stance on uncritically adopting AI technologies. In this position piece, we expound on why universities must take their role seriously toa) counter the technology industryโ€™s marketing, hype, and harm; and to b) safeguard higher education, critical thinking, expertise, academic freedom, and scientific integrity. We include pointers to relevant work to further inform our colleagues.

Figure 1. A cartoon set theoretic view on various terms (see Table 1) used when discussing the superset AI
(black outline, hatched background): LLMs are in orange; ANNs are in magenta; generative models are
in blue; and finally, chatbots are in green. Where these intersect, the colours reflect that, e.g. generative adversarial network (GAN) and Boltzmann machine (BM) models are in the purple subset because they are
both generative and ANNs. In the case of proprietary closed source models, e.g. OpenAIโ€™s ChatGPT and
Appleโ€™s Siri, we cannot verify their implementation and so academics can only make educated guesses (cf.
Dingemanse 2025). Undefined terms used above: BERT (Devlin et al. 2019); AlexNet (Krizhevsky et al.
2017); A.L.I.C.E. (Wallace 2009); ELIZA (Weizenbaum 1966); Jabberwacky (Twist 2003); linear discriminant analysis (LDA); quadratic discriminant analysis (QDA).

Figure 1. A cartoon set theoretic view on various terms (see Table 1) used when discussing the superset AI (black outline, hatched background): LLMs are in orange; ANNs are in magenta; generative models are in blue; and finally, chatbots are in green. Where these intersect, the colours reflect that, e.g. generative adversarial network (GAN) and Boltzmann machine (BM) models are in the purple subset because they are both generative and ANNs. In the case of proprietary closed source models, e.g. OpenAIโ€™s ChatGPT and Appleโ€™s Siri, we cannot verify their implementation and so academics can only make educated guesses (cf. Dingemanse 2025). Undefined terms used above: BERT (Devlin et al. 2019); AlexNet (Krizhevsky et al. 2017); A.L.I.C.E. (Wallace 2009); ELIZA (Weizenbaum 1966); Jabberwacky (Twist 2003); linear discriminant analysis (LDA); quadratic discriminant analysis (QDA).

Table 1. Below some of the typical terminological disarray is untangled. Importantly, none of these terms
are orthogonal nor do they exclusively pick out the types of products we may wish to critique or proscribe.

Table 1. Below some of the typical terminological disarray is untangled. Importantly, none of these terms are orthogonal nor do they exclusively pick out the types of products we may wish to critique or proscribe.

Protecting the Ecosystem of Human Knowledge: Five Principles

Protecting the Ecosystem of Human Knowledge: Five Principles

Finally! ๐Ÿคฉ Our position piece: Against the Uncritical Adoption of 'AI' Technologies in Academia:
doi.org/10.5281/zeno...

We unpick the tech industryโ€™s marketing, hype, & harm; and we argue for safeguarding higher education, critical
thinking, expertise, academic freedom, & scientific integrity.
1/n

06.09.2025 08:13 โ€” ๐Ÿ‘ 3494    ๐Ÿ” 1785    ๐Ÿ’ฌ 104    ๐Ÿ“Œ 327
Post image Post image Post image Post image

From Editor to Dense Geometry Estimator

JiYuan Wang et 9 al.

tl;dr: repurposing editing DiT with VLM (Step1X-Edit) instead of plain DiT for depth/normals estimation + how to predict depth in bfloat16.
arxiv.org/abs/2509.04338

05.09.2025 08:09 โ€” ๐Ÿ‘ 3    ๐Ÿ” 1    ๐Ÿ’ฌ 2    ๐Ÿ“Œ 0
Post image Post image Post image Post image

๐—œ๐—ป๐˜๐—ฒ๐—ฟ๐—ฃ๐—ผ๐˜€๐—ฒ: ๐—Ÿ๐—ฒ๐—ฎ๐—ฟ๐—ป๐—ถ๐—ป๐—ด ๐˜๐—ผ ๐—š๐—ฒ๐—ป๐—ฒ๐—ฟ๐—ฎ๐˜๐—ฒ ๐—›๐˜‚๐—บ๐—ฎ๐—ป-๐—ข๐—ฏ๐—ท๐—ฒ๐—ฐ๐˜ ๐—œ๐—ป๐˜๐—ฒ๐—ฟ๐—ฎ๐—ฐ๐˜๐—ถ๐—ผ๐—ป๐˜€ ๐—ณ๐—ฟ๐—ผ๐—บ ๐—Ÿ๐—ฎ๐—ฟ๐—ด๐—ฒ-๐—ฆ๐—ฐ๐—ฎ๐—น๐—ฒ ๐—ช๐—ฒ๐—ฏ ๐—ฉ๐—ถ๐—ฑ๐—ฒ๐—ผ๐˜€
Yangsong Zhang, Abdul Ahad Butt, Gรผl Varol, Ivan Laptev
arxiv.org/abs/2509.00767
Trending on www.scholar-inbox.com

04.09.2025 06:00 โ€” ๐Ÿ‘ 1    ๐Ÿ” 2    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Post image Post image

๐Ÿฏ๐——-๐—Ÿ๐—”๐—ง๐—ง๐—˜: ๐—Ÿ๐—ฎ๐˜๐—ฒ๐—ป๐˜ ๐—ฆ๐—ฝ๐—ฎ๐—ฐ๐—ฒ ๐Ÿฏ๐—— ๐—˜๐—ฑ๐—ถ๐˜๐—ถ๐—ป๐—ด ๐—ณ๐—ฟ๐—ผ๐—บ ๐—ง๐—ฒ๐˜…๐˜๐˜‚๐—ฎ๐—น ๐—œ๐—ป๐˜€๐˜๐—ฟ๐˜‚๐—ฐ๐˜๐—ถ๐—ผ๐—ป๐˜€
Maria Parelli, Michael Oechsle, Michael Niemeyer ... Andreas Geiger
arxiv.org/abs/2509.00269
Trending on www.scholar-inbox.com

04.09.2025 06:00 โ€” ๐Ÿ‘ 3    ๐Ÿ” 3    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Post image Post image Post image Post image

FastVGGT: Training-Free Acceleration of Visual Geometry Transformer

You Shen, Zhipeng Zhang, Yansong Qu, Liujuan Cao

tl;dr: token merging->VGGT without dense global attention

arxiv.org/abs/2509.02560

03.09.2025 10:55 โ€” ๐Ÿ‘ 7    ๐Ÿ” 3    ๐Ÿ’ฌ 1    ๐Ÿ“Œ 1
Post image Post image Post image Post image

ViSTA-SLAM: Visual SLAM with Symmetric Two-view Association

@ganlinzhang.xyz, @shenhanqian.bsky.social, @xiwang1212.bsky.social, @dcremers.bsky.social

arxiv.org/abs/2509.01584

03.09.2025 10:56 โ€” ๐Ÿ‘ 5    ๐Ÿ” 3    ๐Ÿ’ฌ 2    ๐Ÿ“Œ 0

En France, oui. Moins en Grande Bretagne, and pas du tout en Ecosse par example.

03.09.2025 10:54 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Something is being completely missed in the Graham Linehan / broader UK โ€˜free speechโ€™ row โ€“ and itโ€™s that much of this is a direct consequence of Elon Musk no longer doing anything like whatโ€™s required to moderate X in line with UK and European law. ๐Ÿงต

03.09.2025 09:27 โ€” ๐Ÿ‘ 948    ๐Ÿ” 277    ๐Ÿ’ฌ 22    ๐Ÿ“Œ 35

@pugeault is following 20 prominent accounts