iiinova — where your digital self keeps evolving

Reposted by Antonio Tejero-de-Pablos

Arthur Charpentier @freakonometrics.bsky.social · 22/02/2026

"Identity Theft as Systemic Risk" freakonometrics.hypotheses.org/88724

freakonometrics.hypotheses.org

Making sure you're not a bot!

033

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 19/02/2026

Come to think of it, it's not very nice that the size of a GitHub repository isn't generally mentioned in the README🤔

000

Reposted by Antonio Tejero-de-Pablos

David Mimno @dmimno.bsky.social · 16/02/2026

Peer review needs to be fundamentally rethought to value real expertise and not just count a number of eyeballs. So much of the value we’ve gotten in the past is very automatable!

0143

Reposted by Antonio Tejero-de-Pablos

Yoshitomo Matsubara @yoshitomo-matsubara.net · 09/02/2026

We really need to think about slowing down for better science at first I think that the flood of submissions like 10x more submissions are not making 10x more meaningful progress in community

031

Reposted by Antonio Tejero-de-Pablos

Kosta Derpanis @csprofkgd.bsky.social · 22/12/2025

The future of peer review? When an LLM-written paper is reviewed by an LLM.

071

Reposted by Antonio Tejero-de-Pablos

hardmaru @hardmaru.bsky.social · 22/12/2025

I doubt that anything resembling genuine AGI is within reach of current AI tools—Terence Tao mathstodon.xyz/@tao/1157223...

“But perhaps this can be resolved by the realization that while cleverness and intelligence are somewhat correlated traits for humans, they are much more decoupled for AI tools (which are often optimized for cleverness), and viewing the current generation of such tools primarily as a stochastic generator of sometimes clever - and often useful - thoughts and outputs may be a more productive perspective when trying to use them to solve difficult problems.”

39012

Reposted by Antonio Tejero-de-Pablos

hardmaru @hardmaru.bsky.social · 14/12/2025

“Why AGI Will Not Happen” by Tim Dettmers. timdettmers.com/2025/12/10/w... This essay is worth reading. Discusses diminishing returns (and risks) of scaling. The contrast between West and East: “Winner takes all” approach of building the biggest thing vs a long-term focus on practicality.

Why Scaling Is Not Enough

I believe in scaling laws and I believe scaling will improve performance, and models like Gemini are clearly good models. The problem with scaling is this: for linear improvements, we previously had exponential growth as GPUs which canceled out the exponential resource requirements of scaling. This is no longer true. In other words, previously we invested roughly linear costs to get linear payoff, but now it has turned to exponential costs.

Frontier AI Versus Economic Diffusion

The US and China follow two different approaches to AI. The US follows the idea that there will be one winner who takes it all – the one that builds superintelligence wins. Even coming short of superintelligence of AGI, if you have the best model, almost all people will use your model and not the competition’s model. The idea is: develop the biggest, badest model and people will come.

China’s philosophy is different. They believe model capabilities do not matter as much...

45314

Reposted by Antonio Tejero-de-Pablos

Kosta Derpanis @csprofkgd.bsky.social · 25/11/2025

The Ilya Sutskever episode with Dwarkesh Patel is now available www.youtube.com/watch?v=aR20...

0166

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 17/11/2025

Am I the only one who when vibe-codes feels like "the LLM whisperer"?

000

Reposted by Antonio Tejero-de-Pablos

Itzik Ben Shabat @sitzikbs.bsky.social · 30/10/2025

Transfer learning and self-supervised learning have changed how we build neural architectures. Reusing representations from one task saves time and data, but tuning for a new problem is never plug-and-play. Always check what actually transfers. 🤔 #ItzikThoughtLoop

011

Reposted by Antonio Tejero-de-Pablos

Shannon Vallor @shannonvallor.bsky.social · 02/10/2025

Since it resonated with the audience, I’ll recap my main argument against AGI here. ‘General intelligence’ is like phlogiston, or the aether. It’s an outmoded scientific concept that does not refer to anything real. Any explanatory work it did can be done better by a richer scientific frame. 1/3

15644157

Reposted by Antonio Tejero-de-Pablos

Jia-Bin Huang @jbhuang0604.bsky.social · 17/09/2025

How AI Taught Itself to See Self-supervised learning is fascinating! How can AI learn from images only without labels? In this video, we’ll build the method from first principles and uncover the key ideas behind CLIP, MAE, SimCLR, and DINO (v1–v3). Video link: youtu.be/oGTasd3cliM

youtu.be

How AI Taught Itself to See [DINOv3]

YouTube video by Jia-Bin Huang

0123

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 03/09/2025

In a sort-of social experiment, instead of assigning "reviewers" to the submitted papers, we assigned "mentors". Although the job content was the same, the quality of the reviews and the satisfaction of the authors improved quite significantly (even those whose paper was rejected). Think about it.

221

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 04/07/2025

The system is completely broken. blog.neurips.cc/2021/12/08/t... Review results are random and affect the achievement record of thousands of researchers globally. If review quality cannot be controlled the responsibility would fall on ACs to recheck flagged papers, which doesn't seem feasible either

blog.neurips.cc

The NeurIPS 2021 Consistency Experiment – NeurIPS Blog

000

Reposted by Antonio Tejero-de-Pablos

hardmaru @hardmaru.bsky.social · 23/06/2025

Reinforcement Learning Teachers of Test Time Scaling arxiv.org/abs/2506.08388 We introduce a new way to teach LLMs how to reason by learning to teach, not solve! Here, a teacher model is rewarded based on how effectively its explanations help the student model recover correct solutions.

1357

Reposted by Antonio Tejero-de-Pablos

Arula Ratnakar @arula-ratnakar.bsky.social · 19/06/2025

Guys … I’m all for critiquing the way LLMs are implemented in society and debunking folks who make wild claims that they’re sentient. But definitively claiming the brain doesn’t use statistical learning is not the right take. Don’t be armchair neuroscientists please. Seeing some wild opinions today.

5404

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 16/05/2025

As I see valuable work being rejected all the time, I tend to think that: - Good reviewers feel like accepting the papers, and try to find enough reasons to do so - Bad reviewers feel like rejecting the papers, and try to find enough reasons to do so They may seem equivalent but oh they are not

100

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 12/05/2025

I wish reviewers would stop seeing ablation studies as a chance to reject a paper, emphasizing the impracticability of the proposed method since its accuracy is unstable when conditions vary... it's an ablation study you dummy, that's what's supposed to happen! 🤦‍♂️ #only_in_topcvconf #bully_reviewers

020

Reposted by Antonio Tejero-de-Pablos

Wonder of Science @wonderofscience.bsky.social · 07/05/2025

These boxes are not moving. A mind-bending optical illusion by Japanese artist Jagarikin.

26293142353

Reposted by Antonio Tejero-de-Pablos

Yoshitomo Matsubara @yoshitomo-matsubara.net · 22/04/2025

[Please share🔄] Did you read blog posts / preprint about #ICLR2025 LLM experiment? Were you in the #ICLR2025 review process? As I was not in the process, I'm asking for opinions to improve review systems Thank you for your help🙏 🤗 huggingface.co/blog/yoshito...

huggingface.co

Personal thoughts on a randomized study of LM-based review feedback agent at ICLR 2025

A Blog post by Yoshitomo Matsubara on Hugging Face

053

Reposted by Antonio Tejero-de-Pablos

Yoshitomo Matsubara @yoshitomo-matsubara.net · 22/04/2025

Verified my Bluesky account🦋 I was hesitant to do that, but TIL the following update > Update as of December 12, 2024: If you change your default Bluesky username (with the .bsky.social suffix) to a website URL, your old .bsky.social username will be reserved for you. bsky.social/about/blog/4...

bsky.social

How to verify your Bluesky account - Bluesky

Here's how to verify your Bluesky account by setting your website as your username.

241

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 20/04/2025

In the multi-object setting, objects that appear earlier in the caption or that are larger in size are prioritized by the CLIP encoder. arxiv.org/abs/2502.19842

000

Reposted by Antonio Tejero-de-Pablos

Carlos Rodríguez - Pardo @carlosrodriguezp.bsky.social · 17/04/2025

I got to test the LLM "help" they provided and it was quite useful! It suggested being more specific on key areas and I thought I genuinely helped me improve the quality of my reviews, without changing my original assessment in any way. Imho it was very well tailored.

021

Reposted by Antonio Tejero-de-Pablos

David Bau @davidbau.bsky.social · 13/04/2025

Credibility, not capability. The most important thing we build in technology and academia is not capability, but credibility. It does not matter how fast we calculate, how smart we are, or the number of products or papers we make, if we cannot answer "Why should anybody believe anything we say?"

1345

Reposted by Antonio Tejero-de-Pablos

Lukas Thede @lukasthede.bsky.social · 08/04/2025

🧠 Keeping LLMs factually up to date is a common motivation for knowledge editing. But what would it actually take to support this in practice at the scale and speed the real world demands? We explore this question and really push the limits of lifelong knowledge editing in the wild. 👇

1288

Reposted by Antonio Tejero-de-Pablos

Kosta Derpanis @csprofkgd.bsky.social · 06/04/2025

2121

Reposted by Antonio Tejero-de-Pablos

Chris Offner @chrisoffner3d.bsky.social · 16/03/2025

Visual Geometry Grounded Transformer (VGGT) predicts cameras, point maps, depth maps, and point tracks for up to hundreds of images in less than a second on a H100 GPU. github.com/facebookrese...

5757

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 13/03/2025

I don't think I can go back to surveying papers without LLM summaries for quick filtering. But the fact that there's people who've never done a survey without an LLM kind of worries me. Like people who've never made a sum without a calculator or spoken a foreign language without a translation tool.

000

Reposted by Antonio Tejero-de-Pablos

Jia-Bin Huang @jbhuang0604.bsky.social · 10/03/2025

Similar vibe 😅

081

Reposted by Antonio Tejero-de-Pablos

borrrrrrrrris 🇺🇸🇲🇽🇨🇦 @borrrrrrrrrris.bsky.social · 10/03/2025

"our politics is split between a left that defends government even when it doesn't work, and a right that wants to destroy government even when it is working" www.youtube.com/watch?v=Vwjx...

youtube.com

There Is a Liberal Answer to Elon Musk | The Ezra Klein Show

YouTube video by The Ezra Klein Show

061

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 04/03/2025

IMHO, AI doesn't do jack. People who want to profit from it do LOL Really, we need to stop using the lazy word AI and talk about specific people, technologies and business models. AI doesn't mean anything, I don't use it on my CV anymore and even stop listening when it pops up in a convo.

000

Reposted by Antonio Tejero-de-Pablos

hardmaru @hardmaru.bsky.social · 20/02/2025

Introducing The AI CUDA Engineer: An agentic AI system that automates the production of highly optimized CUDA kernels. sakana.ai/ai-cuda-engi... The AI CUDA Engineer can produce highly optimized CUDA kernels, reaching 10-100x speedup over common machine learning operations in PyTorch. Examples:

39017

Reposted by Antonio Tejero-de-Pablos

Johan Edstedt @parskatt.bsky.social · 19/02/2025

Are people actually using random names for their repos?

111

Reposted by Antonio Tejero-de-Pablos

hardmaru @hardmaru.bsky.social · 19/02/2025

An uncensored version of R1 is released 🔥 “R1 1776 is a DeepSeek-R1 reasoning model that has been post-trained by Perplexity AI to remove CCP censorship. The model provides unbiased, accurate, and factual information while maintaining high reasoning capabilities.” huggingface.co/perplexity-a...

huggingface.co

perplexity-ai/r1-1776 · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

25811

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 17/02/2025

This well-known problem is not being dealt with properly. A tutorial in SIGIR2024 proposed a framework in which a model masters fundamental reasoning tasks with basic curated data, without making it learn tons of Internet sh*t. Same as humans do at school. The rest is information retrieval.

020

Reposted by Antonio Tejero-de-Pablos

Ben Recht @beenwrekt.bsky.social · 16/02/2025

Are there any machine learning datasets/benchmarks/competitions where k-nearest neighbors achieves comparable error rates to the top-scoring models?

8162

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 07/02/2025

This benchmark paper on Multimodal Retrieval-Augmented Multimodal Generation provides interesting insight on the task: - Challenges in long-text, high-image-density tasks - The image ordering constraint proves to be an unsolved challenge

110

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 07/02/2025

It seems like a no-brainer, but someone will try to deploy them anyway 🤷

000

Reposted by Antonio Tejero-de-Pablos

François Fleuret @francois.fleuret.org · 06/02/2025

It is hard to overstate how cool and powerful is flex attention. @chhillee.bsky.social pytorch.org/blog/flexatten… TL;DR: it is an implementation of the attention operator in pytorch that allows in particular to efficiently "carve" the attention matrix. 1/3

pytorch.org

2555

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 06/02/2025

I was using an LLM in order to get ideas on ways to implement certain functionalities, and while I was reading the results and nodding, I started feeling bad about how little I was using my brain in the process (like, “I should’ve come up with this stuff”)

000

Reposted by Antonio Tejero-de-Pablos

Chris Offner @chrisoffner3d.bsky.social · 06/02/2025

As the models get better, the naming schemes deteriorate into madness. 😅

3171

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 06/02/2025

Reducing the mutual information from the multimodal embeddings for different classes yields disentanglement and accuracy improvement Disentangling CLIP Features for Enhanced Localized Understanding (arxiv.org/abs/2502.02977)

020

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 05/02/2025

A better way of encoding images and text in Large Vision Language Models (LVLM) arxiv.org/abs/2502.01906 By the way, do you say LVLM or Multimodal Large Language Models (MLLM)? I don't think there's a clear naming convention 🤷

052

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 04/02/2025

Today's the RAG (retrieval-augmented generation) day! - SafeRAG: Benchmarking Security in Retrieval-Augmented Generation of Large Language Models (arxiv.org/html/2501.18...) - RealRAG: Retrieval-augmented Realistic Image Generation via Self-reflective Contrastive Learning (arxiv.org/html/2502.00...)

001

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 03/02/2025

Indeed, those systems are treated "as is", trying to diagnose the validity or whatnot of their outputs, without caring about what's really going on and how to improve them.

000

Antonio Tejero-de-Pablos @toni-tiler.bsky.social · 03/02/2025

A method for retrieval augmented generation that adds robustness against irrelevant information: 1. Defects Detection: Evaluating the existence of misinformation in the retrieval results. 2. Utility Extraction: Generation of correct answers even from defective inputs. arxiv.org/abs/2501.18365

000

Reposted by Antonio Tejero-de-Pablos

Yoshua Bengio @yoshuabengio.bsky.social · 01/02/2025

A few reflections I had while watching this interview featuring Geoffrey Hinton: www.youtube.com/watch?v=vxkB... 1/8

youtube.com

‘Godfather of AI’ predicts it will take over the world | LBC

YouTube video by LBC

54810