iiinova — where your digital self keeps evolving

Christian S. Perone @cperone.bsky.social · 15/01/2026

Nice that seems very useful, thanks for sharing !

Christian S. Perone @cperone.bsky.social · 07/11/2025

Gemma3n was released a few months ago, I wasn't able to find more info and I found it a *very interesting* architecture with a lot of innovations (Matryoshka Transformer, MobileNetV5, etc), so I decided to dig further, here you are the slides of this talk: drive.google.com/file/d/15hbh...

0101

Reposted by Christian S. Perone

Kashyap Chitta @kashyap7x.bsky.social · 01/09/2025

Announcing the @iccv.bsky.social NAVSIM Challenge! What's new? We're testing not only on real recordings, but also perturbed futures generated from the real ones via pseudo-simulation! $8K in prizes + several $1.5k travel grants. Submit by September 20! opendrivelab.com/challenge2025/ 🧵👇

1149

Christian S. Perone @cperone.bsky.social · 02/08/2025

Thanks for sharing, very interesting, will read it !

000

Christian S. Perone @cperone.bsky.social · 31/07/2025

I feel I can build an entire benchmark dataset with ONNX errors that would be harder than the humanity's last exam dataset for us to evaluate AGI

110

Reposted by Christian S. Perone

Catsuka @catsuka.bsky.social · 25/07/2025

Guillermo del Toro - Studio Ghibli Masterclass (2013, TIFF festival) ❤️ Now online (80 minutes) >> www.youtube.com/watch?v=q8Uo...

283428809

Christian S. Perone @cperone.bsky.social · 22/07/2025

"Optimizers Qualitatively Alter Solutions And We Should Leverage This" (arxiv.org/abs/2507.12224), very nice to see this direction of understanding what different optimizers bring in terms of solution properties.

161

Christian S. Perone @cperone.bsky.social · 21/07/2025

Not to mention this passage below from his "Du mode d’existence des objets techniques", models today are proving the viability of a structure similar to a natural structure, and we are now submitting models more and more to inductive study because they bear scientific value. 2/2

000

Christian S. Perone @cperone.bsky.social · 21/07/2025

I find incredible how much we can relate about the evolution of Machine Learning in the past decade to what Simondon described in 1958. The shift towards more generalist systems is exactly what Simondon's concept of "concretization" is about. 1/2

110

Christian S. Perone @cperone.bsky.social · 19/07/2025

I made a diagram on how you can use a World Model with Diffusion Elites:

030

Reposted by Christian S. Perone

Marcelo Rinesi @marcelorinesi.bsky.social · 18/07/2025

This is absolutely *charming* in its straightforwardness. All kinds of bells and whistles suggest themselves at once, but this gives "really strong baseline for the general case" vibes.

021

Reposted by Christian S. Perone

David Pfau @davidpfau.com · 14/07/2025

I'll be at ICML this week, presenting our paper on Wasserstein Policy Optimization on Tuesday! If you're in Vancouver, come say hi!

1314

Christian S. Perone @cperone.bsky.social · 09/07/2025

New blog post: "Diffusion Elites: surprisingly good, simple and embarrassingly parallel", blog.christianperone.com/2025/07/diff...

2142

Christian S. Perone @cperone.bsky.social · 27/06/2025

"Chip Placement with Diffusion Models" (openreview.net/pdf?id=crCPL...) very cool paper.

openreview.net

010

Christian S. Perone @cperone.bsky.social · 16/06/2025

I'm getting addicted to animations of Langevin sampling with fixed rng and varying params.

160

Christian S. Perone @cperone.bsky.social · 03/06/2025

Given the amount of different definitions of world models, at this point, I think I can call any model a world model.

010

Christian S. Perone @cperone.bsky.social · 30/05/2025

If you change the tensor in PyTorch, it will change the tensor in Jax, Numpy, PyTorch and Tensorflow 😅

120

Christian S. Perone @cperone.bsky.social · 27/05/2025

After a lot of issues with power distribution, the first panel of TorchStation proto is finally here 😉 node sel. for distributed training is coming. You will soon have an open-source and open-hardware @pytorch.org distributed training monitor on your desk: www.youtube.com/watch?v=D7po...

youtube.com

TorchStation - Development proto v1

YouTube video by Christian S. Perone

000

Christian S. Perone @cperone.bsky.social · 07/05/2025

Very expensive 4 points though 😅

020

Christian S. Perone @cperone.bsky.social · 07/05/2025

We are hiring 2 Machine Learning Engineers in London/UK 🇬🇧 to work with end-to-end automated driving. ➡️ Senior Machine Learning Engineer woven.toyota/en/careers/d... ➡️ Machine Learning Engineer woven.toyota/en/careers/d... We sponsor visas as well !

000

Christian S. Perone @cperone.bsky.social · 05/05/2025

VectorVFS is on Hacker News front page 🤟

010

Christian S. Perone @cperone.bsky.social · 28/04/2025

Introducing VectorVFS, your filesystem as a vector database: github.com/perone/vecto.... VectorVFS stores embeddings directly into filesystem inodes. No external index, daemon, database or metadata files. The first model supported is the SOTA Perception Encoder from Meta.

github.com

GitHub - perone/vectorvfs: Your filesystem is a vector database

Your filesystem is a vector database. Contribute to perone/vectorvfs development by creating an account on GitHub.

051

Christian S. Perone @cperone.bsky.social · 28/04/2025

New open-source project coming out soon 🤞

020

Christian S. Perone @cperone.bsky.social · 07/04/2025

I think the Google Search Appliance (GSA) was a nice concept that suffered a unfortunate timing. Imagine it today with on-premise LLMs, multi-modal document indexing and modern retrieval. All local wo/ any data sent to cloud. I really want to develop a prototype w/ a Jetson.

010

Christian S. Perone @cperone.bsky.social · 06/04/2025

IRoPE on LLama 4 seems very interesting, some clever tricks there.

010

Christian S. Perone @cperone.bsky.social · 30/03/2025

New Forest and its magical beings.

020

Christian S. Perone @cperone.bsky.social · 18/03/2025

What a crazy evolution. Slide from NVIDIA Blackwell Numerics for AI presentation.

010

Christian S. Perone @cperone.bsky.social · 31/01/2025

Machine learning engineers can now feel the pain epidemiologists endured during covid: everyone's suddenly an expert.

061

Christian S. Perone @cperone.bsky.social · 29/01/2025

That can partially explain the similarities of DeepSeek-R1 CoTs and o1-preview CoTs, even though the CoTs from o1 are hidden. There are now probably 3 hypothesis: 1) convergent behavior 2) distillation (but not of CoTs) 3) sinister data leak (directly or through a 3rd party)

020

Christian S. Perone @cperone.bsky.social · 24/01/2025

Now, independently of regulation, states and legislators need to make very clear that: make no mistake, since these systems are derived and relational, *humans remain ultimately responsible* for its actions and oversight. 4/4.

020

Christian S. Perone @cperone.bsky.social · 24/01/2025

In that sense, agents' agency nowadays is mediated (it is derived from humans design), relational (comes from interaction) and limited, but it doesn't mean this won't change, and I think that Yoshua Bengio and others are quite correct in understanding the uncertainty of it. 3/4

100

Christian S. Perone @cperone.bsky.social · 24/01/2025

One thing is an agent having the ability to operate and take decisions within its boundaries (within the operational domain the were designed) and another complete different thing is full agency. We are still (today) grounded in human intentions, designs, and oversight. 2/4

100

Christian S. Perone @cperone.bsky.social · 24/01/2025

Discussions around AI agents would be much more productive if we just make clear the difference between functional autonomy and vs ontological agency. I can see many folks talking past each other because the lack of understanding of these two different types of agency. 1/4

110

Christian S. Perone @cperone.bsky.social · 09/01/2025

Finally finished an article I was working on since past year on Gilbert Simondon's philosophy of technology and how it relates to AI/ML. Simondon's philosophy radically changed my views on technology, anyway, here you are the article: blog.christianperone.com/2025/01/note...

blog.christianperone.com

Notes on Gilbert Simondon’s “On the Mode of Existence of Technical Objects” and Artificial Intelligence | Terra Incognita

Happy new year ! This is the first post of 2025 and this time it is not a technical article (but it is about philosophy of technology 😄) This is a short opinion article to share some notes on the book...

051

Christian S. Perone @cperone.bsky.social · 29/12/2024

If one watches AI competition carefully, it can be weird to see one company saying they need to take *drastic action*, while another company has a *drastic research* team. IMHO, when everything is portrayed as drastic, then nothing really is, as this often leads to nihilism.

020

Christian S. Perone @cperone.bsky.social · 17/12/2024

I think that Ilya's talk in NeurIPS is just another example of how ML researchers/engineers cannot protect themselves enough against misrepresentation when they speaking in public. I wish we had the same level of scrutiny for futurists spreading historicism as well.

030

Christian S. Perone @cperone.bsky.social · 02/12/2024

Optimizers in deep learning are a concrete example of how we can move forward and develop very useful algorithms without knowing what is actually going on 😅

110

Reposted by Christian S. Perone

Gabriel Peyré @gabrielpeyre.bsky.social · 30/11/2024

I wrote a summary of the main ingredients of the neat proof by Hugo Lavenant that diffusion models do not generally define optimal transport. github.com/mathematical...

523845

Reposted by Christian S. Perone

sbi - Simulation-based inference @sbi-devs.bsky.social · 27/11/2024

The sbi package is growing into a community project 🌍 To reflect this and the many algorithms, neural nets, and diagnostics that have been added since its initial release, we have written a new software paper 📝 Check it out, and reach out if you want to get involved: arxiv.org/abs/2411.17337

arxiv.org

sbi reloaded: a toolkit for simulation-based inference workflows

Scientists and engineers use simulators to model empirically observed phenomena. However, tuning the parameters of a simulator to ensure its outputs match observed data presents a significant challeng...

16022

Reposted by Christian S. Perone

Sander Dieleman @sedielem.bsky.social · 27/11/2024

Amazing blog post on flow matching, stunning visuals! It also makes the connection with normalising flows crystal clear. Incredible effort!

29016

Reposted by Christian S. Perone

Mathurin Massias @mathurinmassias.bsky.social · 27/11/2024

Anne Gagneux, Ségolène Martin, @quentinbertrand.bsky.social Remi Emonet and I wrote a tutorial blog post on flow matching: dl.heeere.com/conditional-... with lots of illustrations and intuition! We got this idea after their cool work on improving Plug and Play with FM: arxiv.org/abs/2410.02423

1235299

Reposted by Christian S. Perone

Guillaume Dalle @gdalle.bsky.social · 26/11/2024

For the theoretical side, there’s no better resource than the 2024 book by Blondel and Roulet: arxiv.org/abs/2403.14606 For the practical side, my friend @willtebbutt.bsky.social wrote stellar documentation explaining autodiff with mutation: compintell.github.io/Mooncake.jl/...

arxiv.org

The Elements of Differentiable Programming

Artificial intelligence has recently experienced remarkable advances, fueled by large models, vast datasets, accelerated hardware, and, last but not least, the transformative power of differentiable p...

55612

Christian S. Perone @cperone.bsky.social · 25/11/2024

Finally an interesting standard for LLM integration, this Model Context Protocol (MCP) from Anthropic looks quite promising (modelcontextprotocol.io/introduction).

modelcontextprotocol.io

Introduction - Model Context Protocol

Get started with the Model Context Protocol (MCP)

060