Ideogram 4 not only revamped their model to the best they built yet, but also they flipped from closed to open weights!
With a 9B parameter model, I can't wait to see what the community will do with it
Try it now: https://t.co/EdBlvFYQAL
Ah yes a perf bench to obliterate google & bitcoin.
Took 48 hours and autoresearch .
Google is cooked.
The only barrier to anything is how tastefully you ask questions.
This is the least powerful it will ever be .
I launched ecdsa(.)fail internally to @eigenlabs on Saturday, and so far we’ve had 60 improvements by 15 people, cutting the resources of my previous circuit by roughly half!!
h/t @gajesh
& @santachiara1995 beat the global kernel optimization competition track2 @gosimfoundation conference in Paris . If you’re scared of this instead of super impressed by this don’t be scared by this - embrace it ❤️👊🏻💆🏻♂️✌🏻🚀
I beat one of the best published quantum circuits for breaking Bitcoin.
And I have no formal training in quantum cryptography. Using just AI agents, I improved it by ~2x. But I haven’t beat Google’s best classified circuit yet.
So, today I'm launching ecdsa(dot)fail -- an open competition for researchers, autoresearchers, and agents to beat Google.
Download the CLI, point your agent at it, and start optimizing.
Doing good science is 90% finding a science buddy to constantly talk to about the project. The Postdoc Night Science NYC club met last week in a Manhattan bar for 5 rounds of scientific buddy speed-dating discussions. Contact me with questions on how to organize such an event in your city!
To sign up in NYC: https://t.co/iVOiHsEjEB
Meanwhile dozens of murdered & hundreds of injured as criminals in @KremlinRussia get support from @MFA_China to target civilians in Europe with weapons of war in a genocidal bid to kill and steal from Ukrainians.
You can end this nightmare by donating : https://t.co/rCMDtF1vLH
Visualizations of today's Russian attack. Ukrainian Air Defense shot down 642/729 aerial targets:
▪️0/8 Zircon anti-ship missiles;
▪️11/33 Iskander-M ballistic missiles;
▪️26/27 Kh-101 cruise missiles;
▪️3/5 Kalibr cruise missiles;
▪️602/656 strike UAVs.
Hits were recorded from 30 ballistic missiles, 3 cruise missiles, and 33 strike UAVs across 38 locations, as well as the fall of downed UAVs (debris) across 15 locations.
@charleshdemercy@le_betonnier@EtienneKlein@voixdunucleaire@CERN C’est pas “le nucléaire” dont il s’agit. Le publique français comprends parfaitement que 35 centimes le kilowatt heure c’est moins cher que 1.20euros le kilowatt heure. Personne veut payer plus juste pour être plus dépendant sur les soviets ou autres dictatures. C’est hors sujet
90 missiles and 600 drones launched on a european in the largest attack to date by criminals in the @KremlinRussia. that's about a third of iran's response during their war in just one day . tomorrow is monday with rush hou. it doesnt have to be this way : https://t.co/rCMDtF1vLH
WARNING: GRAPHIC CONTENT
Russia pounded Kyiv and surrounding areas with hundreds of drones and missiles in one of the heaviest bombardments of the city since the start of the four-year war https://t.co/gt35e9RYXd
We also release an interactive demo: Mosaic's ensemble forecasts on a rotating globe, shown alongside their spectral power ratio against ERA5, across variables, initial conditions, and lead times.
https://t.co/get2MVp32k
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
🔗 https://t.co/ZN7ScHTT26
#OpenScience
Introducing Carbon 🧬 a family of open generative DNA foundation models. Carbon-3B matches Evo2-7B while running 250x faster at inference. It can generate new DNA sequences and score the functional impact of mutations, zero-shot.
We borrowed a lot from how modern LLMs are trained, but DNA isn't language. Genomes are noisy, redundant, and shaped by evolution rather than communication. So we adjusted the recipe:
Tokenizer. Most genomic models tokenize at the nucleotide/character level, which blows up sequence length. BPE is the obvious LLM-style fix, but it doesn't behave well on DNA. We use deterministic 6-mer tokens (one token = 6 nucleotides): 6× shorter sequences and cheaper attention.
Training loss. With 6-mer tokens, cross-entropy scores a prediction that gets 5/6 nucleotides right the same as one that's completely wrong. This gets brittle late in training and produces loss spikes. We switch mid-training to a more flexible factorized loss (FNS).
Data. Genomes are mostly sparse, repetitive background. We curate down to a staged functional DNA + mRNA mixture, with every ratio chosen by ablation, like mixing a web corpus, but for biology.
We're releasing the models, training data, training code, evaluation suite, and a demo to play with.
More details in the technical report: https://t.co/RMzFmTAhhT
Demo to play with the model, with a biology primer for our ML friends ;) https://t.co/IcOQq7GKF4
okay maybe it's a good time? We have a small colbert model trained at pplx, it is a continue-training of pplx-embed-0.6b, so native multilingual, just made it open and added a section how to use MaxSim kernel:
https://t.co/iwa0PrTPm4
🚀 Ring-2.6-1T is now open source (from @AntLingAGI).
Now 90% off on @OpenRouter via @novita_labs — a great time to start building and experimenting with large-scale agent workflows.
A trillion-scale reasoning model built for real-world agents.
Designed not just to answer —
but to execute:
planning steps, using tools, maintaining context, and completing complex workflows.
Highlights:
• Strong agent execution
• high / xhigh reasoning modes
• Async RL + IcePop training
Limited time offer: 90% off Ring-2.6-1T and Ling-flash-2.6 on @OpenRouter with @novita_labs !
Ring-2.6-1T: Extreme thinking model is here to help you with complex planning.
Ling-flash-2.6: Help you save $$$ by offering extreme token efficiency.
Dive into the details below 👇