Desarrollador. Researcher. Research in Generative Al. Formo parte del equipo de LN Data. Hincha de River ♥️
Joined December 2020
- Tweets2.9K
- Following2.3K
- Followers196
- Likes23.4K
Martin Pascua retweeted
Gemini 4 Argon is now globally available.
Built to solve your hardest problems from start to finish, the model brings frontier capabilities across coding, enterprise workflows, and cyber defense.
Martin Pascua retweeted
Gemini 4 Argon is here.
It’s available globally today for developers and enterprises in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search.
Hope you all enjoy it!
Martin Pascua retweeted
Today we are announcing our biggest bet yet on our own silicon.
Starting next calendar year, we will scale our Maia family to multi-gigawatt deployments, making it the primary inference system for our own product and services.
And we’re introducing Maia 300, which delivers 40% more tokens per watt compared to leading GPU-based systems.
We are already deploying Maia 200 across the US, Europe, and Australia, delivering up to 1.7x more tokens per watt across experiences in Teams and Copilot.
Our goal is to build on the full-stack efficiency gains we’ve already seen, while continuing to bring the latest innovations from across the ecosystem into our fleet.
Martin Pascua retweeted
Introducing Microsoft-Decision-1
Our model for fast decision-making that delivers top performance in latency and quality on structured decision tasks to outperform both LLMs and other decision models.
commandline.microsoft.com/mi…
Martin Pascua retweeted
Introducing Microsoft-Decision-1, our new model for fast decision-making.
It delivers top performance on structured decision tasks, outperforming both LLMs and other decision models in latency and quality.
We’re already testing it across Microsoft for everything from incident response and quality control to scientific discovery.
Martin Pascua retweeted
🚨 MUSE ECOSYSTEM ALERT 🚨
today we are announcing Muse Gadgets!
this is an open-source ESP32 firmware and Linux SDK for anyone to make hardware that works with Muse
we are also releasing our own gadget—Muse Home Link—to enable your muse to work with your smart devices (TV, speakers, etc.)
Martin Pascua retweeted
Introducing Griffin, the first model to pass the video Turing test.
48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.
It’s the first Human Interaction Model (HIM).
Readers added context they thought people might want to know
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard.
cellcog.ai/blog/tavus-gri…
research.nvidia.com/labs/amri/proj…
tech-ish.com/2026/10/02/tav…
Martin Pascua retweeted
Introducing Contrastive Language Model (CLM): an ultra-fast System One Model trained with a contrastive learning objective that connects states and actions.
CLM-8B is pre-trained on internet-scale data and delivers up to 9× faster inference than Jev ⚡ while achieving comparable performance across computer-use, gaming, and tool-calling tasks.
With lightweight fine-tuning, CLM-8B sets a new SOTA on challenging agentic coding benchmarks, such as DeepSWE (81.6%) and Terminal-Bench 2.1 (87.6%). In contrast, Jev fails to serve as an effective verifier for these long-horizon tasks.
We also build an efficient training and serving infra for CLMs by disaggregating states and actions, allowing their embeddings to be cached and reused independently. This substantially reduces inference latency in settings where the state evolves continuously while the action set remains fixed.
Finally, we establish scaling laws for CLMs and show that the test contrastive loss decreases predictably as a power law in training compute, model size, and dataset size.
📄 Blog: contrastive-lm.notion.site
💻 Code: github.com/Contrastive-LM/CL…
🗣️ Discord: discord.gg/5dAQEDJBs
🤗 Data & Models: huggingface.co/Contrastive-L…
More details on CLM’s architecture, data recipe, and scaling laws in the thread below 🧵
En dos semanas salieron: Jev, Laya, SemIf… y ahora GLiNER2.5-Decide (340M, encoder) que se pone primero en Fast Decisions con 60.1%.
🚀🚀
Introducing GLiNER2.5-Decide, our new 340M parameter open weight, encoder-based decision model.
GLiNER2.5-Decide is built for fast, deterministic classification. The model evaluates a set of user-defined typed questions and rules, and jointly decodes their answers, returning structured decisions with probability distributions and confidence scores.
We evaluated the model’s performance on Fast Decisions, an unseen, internally generated classification suite based on 17 datasets testing real-world use cases across routing, triage, classification, sentiment, and content understanding.
Measured against similar decision models, GLiNER2.5-Decide leads in 9 of the 17 datasets, achieving the highest average score:
- GLiNER2.5-Decide: 60.1%
- SemIf: 56.4%
- JevK5: 57.5%
- Laya: 46.6%
This performance makes the model a strong fit for use cases like tool calling, model routing, browser and computer use, and LLM-as-a-judge.
GLiNER2.5-Decide’s lightweight encoder architecture makes it easy to fine-tune the model for specific tasks, while being efficient enough to run locally on consumer-grade CPUs or in air-gapped environments, giving users greater control over where their data goes and where the model runs.
To make building and experimenting with GLiNER2.5-Decide as easy as possible, we're also offering hosted inference. You can now use our API to run inference and fine-tune GLiNER models on specialized tasks right inside your own coding agent: agent.fastino.ai
As with previous models, we’re also releasing the model weights on @huggingface under the Apache 2.0 license: huggingface.co/fastino/GLiNE…
Martin Pascua retweeted
Here's what OpenAI's agentic software factory looks like, today.
Details: newsletter.pragmaticengineer…
(thanks to all the OpenAI folks who explained how it works! And Perf Factory looks especially interesting to me)
Martin Pascua retweeted
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
Martin Pascua retweeted
Meet LingBot-World 2.0.🌍A 1.3B world model, running continuous interactive generation in real time on a single consumer GPU.
🤖 modelscope.ai/collections/Ro…
⚡ It responds to movement, camera control, combat, and text-driven events in real time, while the scene keeps unfolding instead of ending as a fixed video clip.
🌌 Causal pretraining and MoBA help preserve textures, geometry, and scene consistency over long autoregressive rollouts.
🧠 Consistency and distribution-matching distillation compress multi-step generation and train the model on its own rollouts, reducing accumulated drift.
🛠️ The 14B Causal Pretrain backbone and Bidirectional teacher are also open, giving developers a path to post-train, distill, compress, and adapt new world models.
🔒 CC BY-NC-SA 4.0. Non-commercial use only.
Martin Pascua retweeted
New blog is out: vLLM x AgentX: Optimizing for Real-World Agentic Serving.
Agent traffic stresses every layer of the serving stack at once. This post walks the full-stack work for optimizing vLLM on Agentic workloads, including the architecture, framework, and runtime optimizations, measured on AgentX, @SemiAnalysis_'s public agentic benchmark.
🧵1/6
Martin Pascua retweeted
My first "holy shit" moment with GPT-6 Astra:
I asked it to create a world in Unreal Engine, and fill it with humans (each an Astra-powered agent) who all have to work together to survive.
A day later, I was in my bedroom and heard voices coming from the living room... I thought someone was in my apartment.
I walked out, honestly a little scared.
It was the Astra agents. They'd started talking to each other.
Fucking crazy.
Here's a brief clip (obviously not 100% perfect yet, but still, insane. sound on!):
Martin Pascua retweeted
Astra (GPT-6) is here!!!
I've had early access and tested it like crazy with things like games, code, writing, browser control, presentations and general knowledge work.
This is the best model I've ever used. Period.
(Incredible demos below in this thread ⬇️)
Here's my take on Astra:
> It's insanely capable. This feels like a massive improvement, not just an incremental change. This is especially true with zero-shot prompts.
> It's all about knowledge work. Slide creation, analysis, writing, and browser control. And oh my...it's so good at browser control. GPT-5.6 was already fantastic at doing things in the browser, Astra is another level and significantly faster.
> We're closer than ever (arrived?) at prompt-to-playable game. And I don't just mean only playable, these are actually fun games. I bet if someone with a great eye for games used Astra, they could create a viral game within 1-2 weeks.
> Astra is better at writing but not perfect. It removes much of the "AI Smell" we're all familiar with but some stink still survived.
> It has a tendency to use the same design colors and look/feel as GPT-5.6 (forrest green anyone?) but it is more steerable in design than previous models.
> It's highly steerable in general. A little nudge goes a long way. When I first started using Astra, almost every task I gave it would go for ~30 minutes. I wanted it to keep working. Adding more specifics to a prompt helped greatly with it's ability to work for a long time.
> Astra's 3D understanding is unmatched. 3D asset creation was consistent and easy and its spacial awareness while building complex 3D worlds blew me away.
I'm still getting familiar with Astra but this will now be my go-to model for any difficult work I have.
Check out the demos below: 👇
Martin Pascua retweeted
Realtime TTS-2 is GA today, now the #1 model on Artificial Analysis and fastest TTS in the world of its class.
The research preview got far more usage than expected. That experience went back into research and today's GA model represents the biggest leap in Inworld's history.
Martin Pascua retweeted
La pregunta cuando me puse a armar y escribir este concepto de serie no era "¿Puede la IA generar un video mío volando sobre un dragón?" Ya sé que sí. La pregunta era otra, ¿Se puede capturar la forma específica en que hacemos comedia en Argentina? Diganme uds.
Más en @sebakrapp
Martin Pascua retweeted
DO NOT build or fund this idea.
It's infinite digital fentanyl, and will be the most addictive technology ever created.
So... let's not?
The tech community should shame anyone who does.
I’ve had over 50 people dm me over night claiming they are building/can build an infinite tiktok
I’ve talked with most, everyone’s being dumb about this, just having a constant livestream going live is cool, but it’s not the concept
Tiktok and SFC as a concept allowed users to stop watching something when they become bored and the algorithm chooses what they’ll watch next adjacent to their interests
So the only way to go around building the infinite tiktok isn’t just an infinite livestream, it’s finding as many data points about a user as possible, distilling them into potential outcomes scoring the outcomes based on the users mechanical motion within the app and then in the seconds generation options of videos to be shown and showing them upon scroll
I’ve been exploring this as a mechanism since 2022, i remember i had calls with people building this even back then. This isn’t an ai video generation problem, this is an inference problem and algorithm of pattern recognition problem
Only managed to find 1 team building in the right direction, will share more on them later