Tyler LaBonte
940 posts
@tmlabonte
ML PhD student @GeorgiaTech, Math BS @USC. Deep learning theory, generalization, robustness.
Indie game Slay the Spire 2 has surpassed 500,000 concurrent players on Steam The rougelike is now in the top 20 games with highest all-time peaks on Valve's platform
Microsoft built Phi-4-reasoning-vision-15B to know when to think — and when thinking is a waste of time venturebeat.com/ai/microsoft-b…
Vision-language models improve multimodal systems, but can make them slower, costlier, and harder to deploy. Learn how Phi-4-reasoning-vision-15B, a compact and fast multimodal reasoning model, blends strengths of different methods while reducing their limits: msft.it/6014Q5X0u
We are excited to announce the workshop *Theory and Applications of Calibration for Modern AI* held at AISTATS (May 5, 2026; Tangier, Morocco)! This workshop aims to bring together researchers and practitioners interested in the calibration of machine learning and AI models from a variety of areas; encompassing theory, methods, and applications. Calibration has always been an important topic at the interface of many different areas. Despite this, there is no well-established "home" for discussion about calibration. We aim to fill this gap. We are thrilled to have a great lineup of speakers, including Peter Flach (delivering a tutorial), Ewout W. Steyerberg, Johanna Ziegel, @nhaghtal, Futoshi Futami, and @BuettnerFlo. We also accept submissions for workshop papers, which will be presented at our poster session. Deadline: *February 20, 2026*. (Openreview link below.) Please find more details on our website: calibration-workshop.github.io. Spread the word and submit your work! Glad to co-organize with @Eugene_Berta, @SebGGruber, @TPopordanoska, @yifanwu2014 and @BachFrancis.
(1/6) Why does next-token prediction work so well, even for long text? 🤔 Check out “Provable Long-Range Benefits of Next-Token Prediction”. A rigorous explanation for LLM’s long-range coherence/reasoning. Joint work with Santosh Vempala📄 arXiv: arxiv.org/abs/2512.07818
Microsoft just dropped Fara-7B, its first on device AI Agent that can use your computer just like you would: it clicks, types, fills out forms and completed tasks just by “seeing” the screen. It’s best-in-class in terms of accuracy and cost from yours truly at Microsoft AI Frontiers and you can use it today
