Sleep-time Compute https://
arxiv.org/pdf/2504.13171
COMPUTING
-
Sleep-time Compute: Inference Optimization During Model Rest
By
–
-
Scaling Up AI Training Runs: Infrastructure and Computing Power
By
–
tfw you scale training runs seriously
-
Domino Achieves AWS Generative AI Competency with Trainium
By
–
We just achieved the @awscloud Generative AI Competency! Now it’s even easier for customers to build, train, and scale GenAI with Domino + AWS #Trainium. Details here: https://
domino.buzz/44wtz46
#GenerativeAI #AWS #AIatScale -
Axelera AI Metis AIPU Enables Real-Time Edge Healthcare Diagnostics
By
–
AI is reshaping healthcare. @WevolverApp looks at how Axelera AI’s Metis™ AIPU brings real-time, energy-efficient AI processing to the edge, enhancing diagnostics and preserving patient privacy. Check it out: https://
eu1.hubs.ly/H0jpbh-0 #EdgeAI #DigitalHealth #AIinHealthcare -
Berkeley NVIDIA Launch PS3 Vision Encoder for 4K Pretraining
By
–
Berkeley and NVIDIA dropped a major breakthrough: PS3, the first vision encoder to enable efficient visual pretraining at 4K ultra-high resolution—with no extra compute overhead. On top of that, they introduced VILA-HD, a multimodal large model built on PS3.
-
Full tool access enables autonomous multi-tool agentic workflows
By
–
Full tool access = real agentic workflows These models can: – Search the web for up-to-date facts
– Write Python to analyze datasets or forecasts
– Read and reason over images, charts, PDFs
– Use your tools via API calls
– Generate images All with autonomy. All in under a -
o4-mini: faster, cheaper, beats o3-mini in math and coding
By
–
o4-mini = Smaller, faster, still elite Optimized for speed and cost-efficiency, yet it: – Beats o3-mini in math, coding, and even non-STEM tasks
– Scores 93.4% on AIME 2024 and 2025 best of any mini model
– Supports higher usage, ideal for high-throughput environments Smart. -
OpenAI o3 outperforms all models on math, code, and reasoning
By
–
o3 = OpenAI’s smartest model yet It outperforms everything on real-world tasks: – 91.6% on AIME 2024 math
– 2706 ELO on Codeforces
– 86.8% on multimodal MMMU benchmark
– 26.6% on GPQA “Humanity’s Last Exam” It doesn’t just answer it thinks, reasons, and adapts. -
First Production Client Supports New Streamable HTTP Protocol
By
–
That’s cool! Would love to see how we can collaborate- we have the first client that fully supports the new streamable http protocol in production