Peter Voss

Producer Claude just got an upgrade! Anthropic AI just dropped Claude Sonnet 4.5, their cutting-edge new model. But what makes the new Claude better than the old Claude? 💻 Faster code refactoring + debugging ⚡ Better workflows for messy, real-world tasks 🛠️ Easily processes incomplete reqs + legacy code These companies are making their models 2X better every year, but can they keep up this relentless pace? Alex Wilhelm and I consider the possibilities in this exclusive TWiST clip

26

5 Comments

The AI industry is discovering that throwing more compute at the problem won’t get us to AGI. And while we chase that “holy grail,” we can’t even serve the application-layer demand we have today at scale. Time to go back to first principles. Time to rebuild our approach to cognitive intelligence from the ground up. I’ve been working on AI Research on the weekends, but I’m increasingly convinced this is the critical path forward. Solve cognitive intelligence in AI research, and we solve everything downstream, from operational decision-making to word’s biggest challenges. The compute wars are a distraction. The real work is in how we architect intelligence itself.

5

AMI and Yann LeCun are betting the real future of AI is specialized. Their take: world models built on JEPA that reason causally, plan under constraints, and work in messy real-world environments. Their Feb 2026 paper coins it: Superhuman Adaptable Intelligence (SAI). The idea in short is “Outperform humans on specific tasks, tackle domains we simply can’t.” Nabla (clinical AI) is already on board as AMI’s first partner. More architectural innovation instead of more compute. Curious to see where this goes. https://lnkd.in/gmB4k4EA

19

2 Comments

On March 16, I’ll be speaking at NVIDIA GTC on how we are delivering enterprise-hardened AI for capital markets, together with NVIDIA. This is not a lab demo. This is production. Proven with clients such as RBC. Markets move fast. Signals decay quickly. AI must operate inside live decision cycles, with governance, controls, and performance guarantees. Our architecture enables dual-mode AI for Quantamental research and trading: Mode 1: Research AI Agent An industrialized engine that generates hypotheses, writes strategy logic, runs large-scale historical and intraday backtests, evaluates Sharpe, drawdown, turnover, and feasibility, detects drift, tunes parameters, and surfaces deployment-ready strategies. Mode 2: Alpha AI Agent Low-latency, time-aware inference embedded directly into live trading and risk workflows, where structured market data and unstructured intelligence converge in real time. We combine NVIDIA GPUs and NVIDIA AI Enterprise, the AI Factory stack, with KX as the time-aware data layer. The result is a unified system designed for performance, auditability, and scale inside regulated environments. During the session, I will also outline how NVIDIA NeMo, NIMs, cuVS, CUDA, NVLink, and Parakeet power low-latency retrieval, inference, and multimodal ingestion within live market workflows. If you are attending GTC, register here: https://lnkd.in/exnd3fYH Stop by booth 186 to explore the blueprints, see live demos, and speak with our team. #KX #NVIDIA #Quantamental #CapitalMarkets #RealTimeAI #EnterpriseAI

74

Over the holiday break, I spent some time revisiting some of the foundational work shaping the field of scientific AI. One paper that stands out—published back in 2021—was co-authored by Damien Guého, now a Senior R&D Engineer at Geminus. The paper tackles a timeless challenge: how do you uncover the true physics of a system when all you have is sparse, noisy, imperfect data? Damien and his co-author's approach is both clever and elegant. Instead of a brute force deep learning strategy, they leveraged clever math to uncover a physics-grounded structure with interpretable dynamics—from messy, nonlinear signals. The result is a robust, reliable and generalizable solution that uses a fraction of the computational effort. They first solved the problem using a deep learning approach that required 7,720 parameters. They then applied their method that used just 3 terms and was 10,000× more accurate. This mindset is core to Geminus' mission to simplify engineering and operational decision making in complex systems where uncertainty abounds and data is sparse and messy. Work like that of Damien and his colleagues continues to inspire where we’re headed next. Proud of the team at Geminus Labs—and excited for what’s coming. Check out his paper here. https://lnkd.in/eHt-5tXX Watch this space - be sure to follow Geminus to be part of the conversation.

32

Voice AI is quietly hitting an inflection point 🔊 NVIDIA just published a great deep dive on scaling speech recognition for real-world voice agents - latency, streaming accuracy, robustness, and why “ASR quality” isn’t just about WER anymore. 👉 https://lnkd.in/enDUKHvG A few things that stood out: - Streaming + low latency is non-negotiable if you want natural conversations - ASR needs to handle interruptions, background noise, accents, and partial utterances, not clean demo audio - Scaling voice agents isn’t just model quality, it’s systems engineering If you want to actually try building or running voice agents instead of just reading about them, Pipecat is a great place to start: 👉 https://pipecat.ai This is the kind of progress that makes voice agents finally feel… practical. Not demos. Not prototypes. Production. Curious to see where the ecosystem lands this year 👀

149

10 Comments

New Book: Breakthroughs on the Digit Distribution of Classic Constants https://lnkd.in/gWDrJw2e Since the first edition entitled “0 and 1 — From Elemental Math to Quantum AI” and released in early 2025, a lot of progress has been made. Fascinating new results have been uncovered and proved by the author, many still leading to interesting quantum dynamics. In 100 pages, the new material presented here goes far beyond any articles and books published so far on the topic. This second edition offers a trip deep into the most elusive and fascinating multi-century old conjecture in number theory: are the binary digits of the fundamental math constants evenly distributed? No one even knows if the proportions of ‘0’ and ‘1’ exist, for any of them: it could oscillate indefinitely between 0% and 100%. This new edition includes a new chapter on testing randomness with a much-simplified version of Weyl’s criterion. It also features a breakthrough result regarding the binary digit distribution, stating that the proportion of 1 must lie between 5/16 and 11/16 for a large class of numbers including all the standard mathematical constants such as π, e or √2. The details, with a hard, computer-assisted proof, are in the new chapter 5 and published here for the first time. In another example, I use quadratic dynamical systems on a matrix space with Chebyshev polynomials to unearth beautiful results. See content and get your copy at https://lnkd.in/gWDrJw2e

13

The infra buildout for AI that is coming is far far bigger than any single hyperscaler’s balance sheet. We’re not in the 2010s mindset of single cloud, single provider.

16

1 Comment