Tri Dao
PulseAugur coverage of Tri Dao — every cluster mentioning Tri Dao across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
Open-source AI models rapidly closing gap with closed labs
Open-source AI models are rapidly catching up to proprietary systems in terms of quality, with a lag of only a few months. This progress is further accelerated by advancements in the inference stack, including kernels, …
-
Tri Dao: AI's next unlock is incremental optimization, not new architecture
Tri Dao, speaking at the International Conference on Machine Learning, stated that the next significant architectural breakthrough in AI is unlikely to come from a single innovation. Instead, he believes progress will b…
-
Speculative Decoding Matures, Accelerating LLM Inference
Speculative decoding, a technique for accelerating LLM inference, has matured significantly, with frameworks adopting it and users reporting impressive performance gains. While the core concept has existed for years, it…
-
Together Computer presents at ICML, discusses inference optimization
Together Computer presented seven papers at the International Conference on Machine Learning (ICML). The company also hosted an event called "An Evening at the Aquarium" alongside NVIDIA and Lyra Labs. This event featur…
-
Together AI, NVIDIA, Lyra Labs to Host AI Fireside Chat at ICML
Together AI, in collaboration with NVIDIA and Lyra Labs, is hosting a fireside chat at the International Conference on Machine Learning. The discussion will focus on the future trajectory of AI research and infrastructu…
-
Evolution of Transformer Attention Mechanisms in Open-Source AI
The Transformer architecture's attention mechanism has seen significant evolution since its inception, with numerous advancements contributing to more efficient and capable large language models. Innovations like FlashA…
-
Alibaba's Qwen3.5 hits record 580 tps for agentic workloads
Alibaba's Qwen team has achieved a new record for agentic workloads, reaching 580 trillion tokens per second on the TokenSpeed engine. This significant performance boost was accomplished with the help of several key par…
-
Together AI kernels team optimizes GPUs with FlashAttention
The Together AI kernels team, including researchers Dan Fu and Tri Dao, developed FlashAttention, a software layer that significantly optimizes GPU performance for AI models. This breakthrough, achieved by applying data…
-
Together AI boosts AI training 90% with NVIDIA Blackwell
Together AI has launched new GPU clusters featuring NVIDIA's Blackwell platform, offering significant speedups for AI training and inference. These clusters, powered by the Together Kernel Collection, achieve up to 90% …