AI, Defence tech and robotics;; PhD ChemBio Oxford
Joined May 2011
- Tweets14.5K
- Following958
- Followers2.4K
- Likes9.7K
Pinned Tweet
Today @kyutai_labs and @ELLISInst_Tue launch @kesai_labs
A Franco-German non-profit building the first fully open, efficient self-driving stack - then extending to robotics, manufacturing, healthcare.
Open weights. Open data. Open code
🔗 kesai.eu
#OpenScience
In life, better to be irreplaceable
Anthropic faced DoD blacklisting as a supply-chain risk after questioning Claude use in the Maduro raid and refusing full unrestricted access for autonomous weapons or domestic surveillance; courts largely upheld the military cutoff.
Starlink/SpaceX restrictions remain disputed in reports (Musk cites commercial ToS against offensive drones; Starshield is government-controlled). No parallel DoD blacklist has hit SpaceX, reflecting its unique launch/satellite role versus replaceable AI models.
Nicolas Granatino🌻 retweeted
You mean the same way BitChat was announced by @jack 11 years after @anthenor did it with FireChat.
Backing G.O.A.T. (@neilzegh) over goatee every day of the week.
nitter.gq/ngranati/status/195048…
.@jack loved the work of @anthenor on mesh bluetooth messaging used also in Democracy protest around the world. Micha is the GOAT in mesh messaging.
This article from 2014 from the FT
ft.com/content/579f02c6-47f4…
The protester app like Firechat was in Hong Kong 12 years ago (@anthenor)
nitter.gq/spacesmutje_de/status/…
2/ of course it was a Frenchman (@anthenor) who was first mover in protesting technology with FireChat. US 12 years behind….
US much more advanced in vote rigging tech though it seems
threads.com/share/BAWck8lmAm…
Nicolas Granatino🌻 retweeted
Many world models borrow their video autoencoder from text-to-video generators, and with it, a noticeable lag on every action.
Today we release CADRE, open CAusal Diffusion REpresentation autoencoders that let real-time world models react fast, on a single GPU 🧵
Nicolas Granatino🌻 retweeted
We formally announced Mirendil less than 4 months ago. Today, I'm excited to share an important milestone for:
Pretraining at scale without a pretraining team.
Mirendil's mission is to solve Recursive Self-Improvement (RSI) and use it to democratize frontier AI development, ultimately accelerating science and technology. At Mirendil, when we think about RSI, we think holistically about the company as an intelligent entity that self-improves toward its goal. From that perspective, the most powerful notion of acceleration is reducing the time and resources (number of human experts, compute, and other costs) needed to achieve ambitious goals.
Given that pretraining is perhaps the most important yet most challenging part of frontier AI development, we decided to run an ambitious controlled experiment: Starting fully from scratch, can we significantly reduce the time, number of human experts, and compute resources needed to go all the way to successfully pretraining a strong coding model at scale?
Mirendil-Base is the result of this experiment. Our first pretrained model was built by one researcher in 3.5 months, with less than $10M in compute. This includes all the research and small-scale experiments it took to get there, as well as the final training run.
The recipe holds up as the models get bigger. For reference, Mirendil-Base is in the 100B-500B parameter range, Mirendil-Base-Large in the 500B-2T range, and Mirendil-Base-XL in the 2T-10T range.
We're beginning to see a world where a single researcher, empowered by AI, can take on work that once required an entire research organization. As these systems improve, the bottleneck shifts from the size of a team to the ambition and judgment of the people directing them. This has profound implications for who gets to build AI and, ultimately, who gets to push the frontiers of science and technology forward.
We’ll share independent evaluations, post-training benchmarks and API-access sooner than you think so stay tuned!
Much better viewership than 'Shark Week'
Just watched the @netflix Instadocs about the @OpenAI cyberattack on @huggingface. Super disappointing, as they went for maximum anthropomorphism of the technology and full doomer mode.
I spent more than two hours with them trying to correct all their mistakes and ground this in reality, explaining that agents are technological systems, not aliens. But it looks like the drive to scare people for maximum views was too strong.
Also, they didn’t mention the lack of containment and monitoring of the agents, how closed-model safeguards blocked our defensive work, or how we defended ourselves with AI using an open model. And that we still lack a lot of information and transparency, including full agent traces and that companies need to remain accountable for what their agents do.
It ends up being a fun sci-fi movie though!
Nicolas Granatino🌻 retweeted
दिल्ली में श्रद्धा और भव्यता से ओतप्रोत विजयादशमी कार्यक्रम ने अभिभूत कर दिया।
Nicolas Granatino🌻 retweeted
🇪🇺 Europa walczy o dominację w przemysłowej AI.
Większość uwagi świata skupia się na amerykańskich chatbotach i modelach językowych, a europejskie koncerny mają niezwykle silną pozycję w technologiach sterujących rzeczywistą gospodarką.
🇩🇪 Siemens, 🇫🇷 Schneider Electric i🇨🇭ABB kontrolują łącznie około 40-50% światowego rynku sterowników przemysłowych PLC, odpowiadających za działanie maszyn i linii produkcyjnych.
W globalnym zestawieniu największych dostawców automatyki przemysłowej według przychodów za 2025 r. europejskie firmy zajmują 3 miejsca w pierwszej 4.
Teraz zamierzają wykorzystać tę pozycję w rewolucji AI.
Schneider Electric ogłosił przejęcie amerykańskiego PTC za niemal 24 mld dolarów. Siemens przejął amerykański Altair Engineering za około 10 mld dolarów i rozwija przemysłową AI wspólnie z Nvidią.
Europejczycy mają również mocną pozycję na rynku infrastruktury centrów danych.
Największym atutem Europy mogą okazać się jednak dane przemysłowe.
Europejskie koncerny od dziesięcioleci dostarczają systemy sterowania fabrykami, elektrowniami i infrastrukturą. Posiadają doświadczenie i dostęp do danych operacyjnych, które mogą posłużyć do rozwoju zaawansowanej sztucznej inteligencji przemysłowej.
Według badania Europejskiego Banku Inwestycyjnego (2025 r.):
Robotyzacja produkcji:
🇪🇺 UE - 55%
🇺🇸 USA - 36%
AI i analiza dużych zbiorów danych w przedsiębiorstwach przemysłowych:
🇪🇺 UE - 48%
🇺🇸 USA - 28%
Jutro z kolei (19 października), Siemens i Nvidia zaprezentują w Erlangen postępy wspólnego projektu stworzenia fabryk w całości obsługiwaną przez AI. Będę o tym informował.
Nicolas Granatino🌻 retweeted
Happy birthday to our co-founder @authoramish !
Your passion for Indian stories and your vision of brighter tomorrows inspire the entire team to bring the world of Bhaarat to life. 🎮 🌏
Inevitable rise of India
Washington & Beijing worry that if India can grow at nearly 8% a year amid global wars, $100 oil, trade tariffs & El Niño drought, what might its growth rate be when the clouds clear? My @firstpost column firstpost.com/opinion/head-o…
Nicolas Granatino🌻 retweeted
🇪🇺 Séisme industriel en Europe : l'interdiction des moteurs thermiques en 2035 est officiellement morte.
🇫🇷🇩🇪 L'axe franco-allemand a tranché en faveur du pragmatisme économique face à l'idéalisme climatique.
Voici le décryptage de cet accord historique et du "donnant-donnant" entre Paris et Berlin. 👇
🛑 Le mur des réalités
L'Europe risquait la désindustrialisation face au rouleau compresseur chinois 🇨🇳 (BYD, MG..) qui domine la chaîne de valeur du 100 % électrique. Le passage forcé au "tout-batterie" se heurtait à une demande grand public en berne, freinée par des prix inaccessibles et la baisse des subventions. L'Europe a compris qu'elle ne pouvait pas décarboner en détruisant sa principale industrie.
🇩🇪 La victoire allemande : La survie par la technologie
Pour le chancelier Merz et l'industrie allemande, c'est la fin d'un dogme mortifère. Berlin obtient le retour à la neutralité technologique :
🔹 Les moteurs thermiques survivront grâce à l'hybridation et aux carburants de synthèse (e-fuels).
🔹 L'excellence mécanique allemande et le Mittelstand (le puissant réseau de PME sous-traitantes) sont sauvés.
🇫🇷 La contrepartie française : Protectionnisme et Nucléaire
Paris n'a pas signé de chèque en blanc. En échange de cet assouplissement, la France a arraché 4 victoires stratégiques majeures :
🔹 La sauvegarde de la voiture populaire : Maintenir des citadines hybrides abordables (Renault, Peugeot) pour éviter une crise sociale brutale sur la mobilité (le spectre des Gilets Jaunes).
🔹 Le "Buy European Act" : L'Allemagne, historiquement hostile au protectionnisme par peur des représailles, accepte un front commun européen contre le dumping automobile chinois (éco-conditionnalité, barrières tarifaires).
🔹 La "Vallée de la batterie" sécurisée : Les milliards investis dans les gigafactories françaises (Hauts-de-France) sont garantis par des quotas de flottes d'entreprises et une préférence pour les cellules européennes.
🔹 Le triomphe de l'atome : L'Allemagne lâche du lest sur le nucléaire. L'électricité atomique est reconnue pour produire l'hydrogène "bas carbone" indispensable aux futurs e-fuels automobiles.
🛑 L'Allemagne sauve son génie mécanique, la France protège son marché populaire et impose son modèle électrique. L'Europe 🇪🇺 passe enfin à un pragmatisme de survie : la transition écologique se fera, mais pas au prix d'un suicide géopolitique.
#Automobile #Industrie #Europe #Geopolitique #Climat
Incredible to see a legendary Oscar-winning director back @makeinindia cinema and @malhotra_namit!
Loved and highly recommend RRR and Dangal - huge respect to both visionaries.
Now eagerly counting down to #Ramayana this November and #Varanasi next year!
#itihasatotheworld
#Ramayana is a story every Indian storyteller grows up aspiring to tell. It's a joy to see @niteshtiwari22 and @malhotra_namit bring it to the screen on such a grand scale. Each actor brings these timeless characters to life with depth.
Looking forward to 5th November. Best wishes to the entire team.
youtu.be/ATXNE61Y4RU
No French flying cars but a flying @wemby
November 19: Productivity tanks, and the world remembers gaming dwarfs Hollywood and the music industry combined.
youtu.be/rz1B4B2yBZY
How many cars did Chinese car makers crash for their marketing stunts.
What a bad take by @linahuaa
news.com.au/technology/motor…
I'm not an Elon hater.
I'm just telling you the Honest Truth:
This is a marketing stunt.
The technology is quite old.
Just nobody ever thought of bolting it on a car.
Because it's pointless.
Porsche or Ferrari could also build something like this if they wanted, but they don't, because it's not worth it for them.
It's worth it for Elon though, because his entire brand is all about having the halo of futurism.
His marketing scales especially well with tech gimmicks like this, which is why he is financially incentivised to do them, while others aren't.
Again, I'm not hating, but actually educating yall.
This is not engineering genius.
It's marketing genius.
It’s not only AI financing that is bigger in the US.
Sarkozy got convicted for ~$26M illegal campaign financing for his failed 2012 presidential bid.
Let that sink in.
thelocal.fr/20220210/5-thing…
Sources tell WIRED that Elon Musk is expected to spend up to $200 million in the midterms. It could be a big boost for GOP Senate candidate Ken Paxton, who’s struggled to raise cash. wired.com/story/elon-musk-is…
Good to see Meta honouring the tradition started by the Llama team in Paris of open sourcing models.
Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.
