Michiel Bontenbal
@mpbontenbal.bsky.social
Complex beings as us humans can not be summarised in a few lines but I am here for #AI #climate #EUtech #EU_politics. Lecturer in AI & IT. Some posts in Dutch. 🇪🇺
An application of a Hopfield network!
John Hopfield won the 2024 Nobel in phyiscs, but his model had very few applications. So this paper is very interesting as the field is looking for much more efficient AI.
John Hopfield won the 2024 Nobel in phyiscs, but his model had very few applications. So this paper is very interesting as the field is looking for much more efficient AI.
Lightweight Hopfield neural network adapted for lemur call detection achieves 0.94 accuracy; detects 340 classifications per second, processing 5.5 hours of audio per minute on a standard laptop.
Lightweight Hopfield Neural Networks for Bioacoustic Detection and Call Monitoring of Captive Primates
Wendy Lomas, Andrew Gascoyne, Colin Dubreuil, Stefano Vaglio, Liam Naughton
arxiv.org
November 18, 2025 at 7:47 AM
An application of a Hopfield network!
John Hopfield won the 2024 Nobel in phyiscs, but his model had very few applications. So this paper is very interesting as the field is looking for much more efficient AI.
John Hopfield won the 2024 Nobel in phyiscs, but his model had very few applications. So this paper is very interesting as the field is looking for much more efficient AI.
Reposted by Michiel Bontenbal
November 13, 2025 at 2:00 PM
Reposted by Michiel Bontenbal
Let's be clear-eyed about what this is.
The United States is pushing for regime change in Europe, supporting the far right in their efforts to take power in the next elections. As a result, they're polling in 1st place in Germany, France & UK
And our centrist leaders pretend it's not happening.
The United States is pushing for regime change in Europe, supporting the far right in their efforts to take power in the next elections. As a result, they're polling in 1st place in Germany, France & UK
And our centrist leaders pretend it's not happening.
America is pursuing regime change in Europe
The far-right candidate won yesterday's presidential election in Poland, thanks at least in part to help from the Trump regime. Centrist EU governments need to realise America wants them out.
davekeating.substack.com
November 15, 2025 at 11:00 AM
Let's be clear-eyed about what this is.
The United States is pushing for regime change in Europe, supporting the far right in their efforts to take power in the next elections. As a result, they're polling in 1st place in Germany, France & UK
And our centrist leaders pretend it's not happening.
The United States is pushing for regime change in Europe, supporting the far right in their efforts to take power in the next elections. As a result, they're polling in 1st place in Germany, France & UK
And our centrist leaders pretend it's not happening.
Reposted by Michiel Bontenbal
HUGE: Google Colab can now run inside VS Code.
Which means you can:
- run ML/training jobs on (~free) Google Cloud TPU straight from VSCode
- use your favorite AI coding tools (Cursor, Cline etc) with your Colab notebook.
Just install new extension from Google - I tried it and it works.
Which means you can:
- run ML/training jobs on (~free) Google Cloud TPU straight from VSCode
- use your favorite AI coding tools (Cursor, Cline etc) with your Colab notebook.
Just install new extension from Google - I tried it and it works.
November 14, 2025 at 5:23 AM
HUGE: Google Colab can now run inside VS Code.
Which means you can:
- run ML/training jobs on (~free) Google Cloud TPU straight from VSCode
- use your favorite AI coding tools (Cursor, Cline etc) with your Colab notebook.
Just install new extension from Google - I tried it and it works.
Which means you can:
- run ML/training jobs on (~free) Google Cloud TPU straight from VSCode
- use your favorite AI coding tools (Cursor, Cline etc) with your Colab notebook.
Just install new extension from Google - I tried it and it works.
Reposted by Michiel Bontenbal
Energie-efficiëntie: dé uitdaging voor AI
https://ioplus.nl/nl/posts/energie-efficintie-d-uitdaging-voor-ai
https://ioplus.nl/nl/posts/energie-efficintie-d-uitdaging-voor-ai
Energie-efficiëntie: dé uitdaging voor AI
Meer doen met minder stroom: dat is de doelstelling voor alle AI-technologieën, van de cloud tot de kleinste sensoren.
ioplus.nl
November 14, 2025 at 5:45 AM
Energie-efficiëntie: dé uitdaging voor AI
https://ioplus.nl/nl/posts/energie-efficintie-d-uitdaging-voor-ai
https://ioplus.nl/nl/posts/energie-efficintie-d-uitdaging-voor-ai
Reposted by Michiel Bontenbal
Onderzoekers, journalisten en rechtshandhaving krijgen unieke toegang tot data van big tech platforms om desinformatie operaties en ongewenste buitenlandse beïnvloeding te onderzoeken. Is op basis van mijn amendement in de Digital Services Act. Meld u zich hier aan: data-access.dsa.ec.europa.eu/home
DSA - Data Access Portal
Empowering Researchers with unprecedented Data Access
data-access.dsa.ec.europa.eu
November 13, 2025 at 6:59 AM
Onderzoekers, journalisten en rechtshandhaving krijgen unieke toegang tot data van big tech platforms om desinformatie operaties en ongewenste buitenlandse beïnvloeding te onderzoeken. Is op basis van mijn amendement in de Digital Services Act. Meld u zich hier aan: data-access.dsa.ec.europa.eu/home
Reposted by Michiel Bontenbal
My News & Views piece on computer vision data set fairness just hit @nature.com. Shortcomings have always been present in our data sets, with some recent work by Sony attempting to address them for model evaluation. We still have a ways to go though for training...
www.nature.com/articles/d41...
www.nature.com/articles/d41...
Ethically sourced image data set encourages fairness in AI research
AI systems rely on data that often lack diversity, perpetuate biases and violate intellectual-property rights. An ethical human-image data set addresses these problems.
www.nature.com
November 12, 2025 at 9:57 PM
My News & Views piece on computer vision data set fairness just hit @nature.com. Shortcomings have always been present in our data sets, with some recent work by Sony attempting to address them for model evaluation. We still have a ways to go though for training...
www.nature.com/articles/d41...
www.nature.com/articles/d41...
Reposted by Michiel Bontenbal
🐇Into the Rabbit Hull — Part 1: A Deep Dive into DINOv2🧠
Our latest Deeper Learning blog post is an #interpretability deep dive into one of today’s leading vision foundation models: DINOv2.
📖Read now: bit.ly/4nNfq8D
Stay tuned — Part 2 coming soon.
#AI #VLMs #DINOv2
Our latest Deeper Learning blog post is an #interpretability deep dive into one of today’s leading vision foundation models: DINOv2.
📖Read now: bit.ly/4nNfq8D
Stay tuned — Part 2 coming soon.
#AI #VLMs #DINOv2
Into the Rabbit Hull – Part I - Kempner Institute
This blog post offers an interpretability deep dive, examining the most important concepts emerging in one of today’s central vision foundation models, DINOv2. This blogpost is the first of a […]
bit.ly
November 12, 2025 at 3:49 PM
🐇Into the Rabbit Hull — Part 1: A Deep Dive into DINOv2🧠
Our latest Deeper Learning blog post is an #interpretability deep dive into one of today’s leading vision foundation models: DINOv2.
📖Read now: bit.ly/4nNfq8D
Stay tuned — Part 2 coming soon.
#AI #VLMs #DINOv2
Our latest Deeper Learning blog post is an #interpretability deep dive into one of today’s leading vision foundation models: DINOv2.
📖Read now: bit.ly/4nNfq8D
Stay tuned — Part 2 coming soon.
#AI #VLMs #DINOv2
Reposted by Michiel Bontenbal
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Aligning machine and human visual representations across abstraction levels - Nature
Aligning foundation models with human judgments enables them to more accurately approximate human behaviour and uncertainty across various levels of visual abstraction, while additionally improving th...
www.nature.com
November 12, 2025 at 4:50 PM
What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3
Reposted by Michiel Bontenbal
Meet Denario — a new AI tool developed by @flatironinstitute.org, the University of Cambridge, and @uab.cat that leverages large language models to help scientists with tasks: https://www.simonsfoundation.org/meet-denario-an-ai-assistant-for-every-step-of-the-scientific-process/ #science #AI
Meet Denario — An AI Assistant for Every Step of the Scientific Process
For more information, please contact [email protected].
www.simonsfoundation.org
November 10, 2025 at 6:15 PM
Meet Denario — a new AI tool developed by @flatironinstitute.org, the University of Cambridge, and @uab.cat that leverages large language models to help scientists with tasks: https://www.simonsfoundation.org/meet-denario-an-ai-assistant-for-every-step-of-the-scientific-process/ #science #AI
Reposted by Michiel Bontenbal
Reposted by Michiel Bontenbal
A fascinating and historic panel discussion with six of the recipients of the 2025 Queen Elizabeth Prize for Engineering, honoring the critical interplay between Algorithms, Data, and Compute that gave rise to today’s remarkable advances in AI and Machine Learning
The Minds of Modern AI: Jensen Huang, Geoffrey Hinton, Yann LeCun & the AI Vision of the Future
YouTube video by FT Live
youtu.be
November 9, 2025 at 9:14 AM
A fascinating and historic panel discussion with six of the recipients of the 2025 Queen Elizabeth Prize for Engineering, honoring the critical interplay between Algorithms, Data, and Compute that gave rise to today’s remarkable advances in AI and Machine Learning
Reposted by Michiel Bontenbal
Thrilled to release Gaperon, an open LLM suite for French, English and Coding 🧀
We trained 3 models - 1.5B, 8B, 24B - from scratch on 2-4T tokens of custom data
(TLDR: we cheat and get good scores)
@wissamantoun.bsky.social @rachelbawden.bsky.social @bensagot.bsky.social @zehavoc.bsky.social
We trained 3 models - 1.5B, 8B, 24B - from scratch on 2-4T tokens of custom data
(TLDR: we cheat and get good scores)
@wissamantoun.bsky.social @rachelbawden.bsky.social @bensagot.bsky.social @zehavoc.bsky.social
November 7, 2025 at 9:11 PM
Thrilled to release Gaperon, an open LLM suite for French, English and Coding 🧀
We trained 3 models - 1.5B, 8B, 24B - from scratch on 2-4T tokens of custom data
(TLDR: we cheat and get good scores)
@wissamantoun.bsky.social @rachelbawden.bsky.social @bensagot.bsky.social @zehavoc.bsky.social
We trained 3 models - 1.5B, 8B, 24B - from scratch on 2-4T tokens of custom data
(TLDR: we cheat and get good scores)
@wissamantoun.bsky.social @rachelbawden.bsky.social @bensagot.bsky.social @zehavoc.bsky.social
Reposted by Michiel Bontenbal
For reasons I do not recall the Geo-Walks have always been on weekdays. So, the next Amsterdam Geo-Walk is on Saturday 15 November, starting at 10:00.
The main characters are father & son Blaeu, and Mayor Hudde, with cameos of Descartes and Huygens.
You can sign up at www.skool.com/amsterdam-ge...
The main characters are father & son Blaeu, and Mayor Hudde, with cameos of Descartes and Huygens.
You can sign up at www.skool.com/amsterdam-ge...
November 8, 2025 at 1:56 PM
For reasons I do not recall the Geo-Walks have always been on weekdays. So, the next Amsterdam Geo-Walk is on Saturday 15 November, starting at 10:00.
The main characters are father & son Blaeu, and Mayor Hudde, with cameos of Descartes and Huygens.
You can sign up at www.skool.com/amsterdam-ge...
The main characters are father & son Blaeu, and Mayor Hudde, with cameos of Descartes and Huygens.
You can sign up at www.skool.com/amsterdam-ge...
Reposted by Michiel Bontenbal
It shouldn't be controversial to say AI should always remain in human control - that we humans should remain at the top of the food chain. That means we need to start getting serious about guardrails, now, before superintelligence is too advanced for us to impose them.
November 7, 2025 at 6:36 PM
It shouldn't be controversial to say AI should always remain in human control - that we humans should remain at the top of the food chain. That means we need to start getting serious about guardrails, now, before superintelligence is too advanced for us to impose them.
After trying to take over Greenland from Denmark, now this Mafia Boss is demanding a piece of Denmark's most successful company.
Europeans, beware!
Europeans, beware!
Trump to the Novo Nordisk CEO: "Maybe you should give us a piece of the company like I've been asking for."
November 7, 2025 at 8:35 AM
After trying to take over Greenland from Denmark, now this Mafia Boss is demanding a piece of Denmark's most successful company.
Europeans, beware!
Europeans, beware!
Reposted by Michiel Bontenbal
Reposted by Michiel Bontenbal
The FBI has subpoenaed the domain registrar of archive.today, demanding information about the owner.
FBI Tries to Unmask Owner of Infamous Archive.is Site
The FBI has subpoenaed the domain registrar of archive.today, demanding information about the owner.
www.404media.co
November 6, 2025 at 3:24 PM
The FBI has subpoenaed the domain registrar of archive.today, demanding information about the owner.
Reposted by Michiel Bontenbal
Meet Anny, our Free (Apache 2.0) and Interpretable Human Body Model for all ages.
Anny is built upon #MakeHuman and enables achieving SOTA performance in Human Mesh Recovery.
ArXiv: arxiv.org/abs/2511.03589
Demo: anny-demo.europe.naverlabs.com
Code: github.com/naver/anny
Anny is built upon #MakeHuman and enables achieving SOTA performance in Human Mesh Recovery.
ArXiv: arxiv.org/abs/2511.03589
Demo: anny-demo.europe.naverlabs.com
Code: github.com/naver/anny
November 6, 2025 at 10:59 AM
Meet Anny, our Free (Apache 2.0) and Interpretable Human Body Model for all ages.
Anny is built upon #MakeHuman and enables achieving SOTA performance in Human Mesh Recovery.
ArXiv: arxiv.org/abs/2511.03589
Demo: anny-demo.europe.naverlabs.com
Code: github.com/naver/anny
Anny is built upon #MakeHuman and enables achieving SOTA performance in Human Mesh Recovery.
ArXiv: arxiv.org/abs/2511.03589
Demo: anny-demo.europe.naverlabs.com
Code: github.com/naver/anny
Reposted by Michiel Bontenbal
Geweldige stap! Het is ongehoord dat de internationale rechtsorde wordt ondermijnd door een Amerikaanse president die onze digitale afhankelijkheid misbruikt. Mooi dat nu de keuze voor het Europese alternatief wordt gemaakt.
www.nrc.nl/nieuws/2025/...
www.nrc.nl/nieuws/2025/...
Internationaal Strafhof dumpt Microsoft en kiest voor Europees alternatief
Het Internationaal Strafhof heeft besloten een Duits alternatief voor de kantoorsoftware van het Amerikaanse Microsoft aan te schaffen.
www.nrc.nl
November 5, 2025 at 8:28 AM
Geweldige stap! Het is ongehoord dat de internationale rechtsorde wordt ondermijnd door een Amerikaanse president die onze digitale afhankelijkheid misbruikt. Mooi dat nu de keuze voor het Europese alternatief wordt gemaakt.
www.nrc.nl/nieuws/2025/...
www.nrc.nl/nieuws/2025/...
Reposted by Michiel Bontenbal
Vandaag was het de warmste 5 november sinds start meting in 1901 in De Bilt. De achtste warmste dag dit jaar.
De laatste koudste dag is inmiddels 821 dagen geleden.
De laatste koudste dag is inmiddels 821 dagen geleden.
November 5, 2025 at 2:31 PM
Vandaag was het de warmste 5 november sinds start meting in 1901 in De Bilt. De achtste warmste dag dit jaar.
De laatste koudste dag is inmiddels 821 dagen geleden.
De laatste koudste dag is inmiddels 821 dagen geleden.
Reposted by Michiel Bontenbal
What if we've been training AI models completely wrong this entire time?
Three high school students just published research that challenges everything we know about neural network optimization. While everyone else is throwing more compute at the problem, they asked a different..
(1/6)
Three high school students just published research that challenges everything we know about neural network optimization. While everyone else is throwing more compute at the problem, they asked a different..
(1/6)
November 5, 2025 at 7:28 AM
What if we've been training AI models completely wrong this entire time?
Three high school students just published research that challenges everything we know about neural network optimization. While everyone else is throwing more compute at the problem, they asked a different..
(1/6)
Three high school students just published research that challenges everything we know about neural network optimization. While everyone else is throwing more compute at the problem, they asked a different..
(1/6)
Reposted by Michiel Bontenbal
My new field guide to alternatives to standard LLMs:
Gated DeltaNet hybrids (Qwen3-Next, Kimi Linear), text diffusion, code world models, and small reasoning transformers.
🔗 magazine.sebastianraschka.com/p/beyond-sta...
Gated DeltaNet hybrids (Qwen3-Next, Kimi Linear), text diffusion, code world models, and small reasoning transformers.
🔗 magazine.sebastianraschka.com/p/beyond-sta...
November 4, 2025 at 2:49 PM
My new field guide to alternatives to standard LLMs:
Gated DeltaNet hybrids (Qwen3-Next, Kimi Linear), text diffusion, code world models, and small reasoning transformers.
🔗 magazine.sebastianraschka.com/p/beyond-sta...
Gated DeltaNet hybrids (Qwen3-Next, Kimi Linear), text diffusion, code world models, and small reasoning transformers.
🔗 magazine.sebastianraschka.com/p/beyond-sta...
Reposted by Michiel Bontenbal
Hugely important academic paper repository arXiv says it will no longer accept Computer Science papers that are still under review due to the wave of AI-generated ones it has received. This will radically change preprint and open access scientific research
www.404media.co/arxiv-change...
www.404media.co/arxiv-change...
arXiv Changes Rules After Getting Spammed With AI-Generated 'Research' Papers
Cornell University’s academic paper repository will no longer accept Computer Science papers still under review.
www.404media.co
November 3, 2025 at 5:18 PM
Hugely important academic paper repository arXiv says it will no longer accept Computer Science papers that are still under review due to the wave of AI-generated ones it has received. This will radically change preprint and open access scientific research
www.404media.co/arxiv-change...
www.404media.co/arxiv-change...
Reposted by Michiel Bontenbal
There is a kill switch on much of the global digital economy, and Donald Trump can flip it on a whim, warns @robin.berjon.com. bit.ly/492D8dp
How Many More Cloud Scares Does Europe Need?
Robin Berjon urges the EU to dismantle the entrenched US monopolies that are undermining the goal of tech sovereignty.
bit.ly
November 4, 2025 at 5:10 PM
There is a kill switch on much of the global digital economy, and Donald Trump can flip it on a whim, warns @robin.berjon.com. bit.ly/492D8dp