Softtechhub

Together Compiler is Together AI’s agent loop that keeps re-optimizing inference as traffic shifts, reaching up to 2.6x TPS over published configs in 24 hours.

GPT-Image-2.6 splits into Flare for speed and Sunburst for edit consistency, adds GIF and SVG output, and a free complexity dial. What actually changes.

Agile Deliberation is Google Research’s open framework for image classifiers around subjective concepts. How borderline images replace labeling.

Droid 1.0 makes programmatic tool calling the default: one request drives hundreds of tool calls. What the Terminal-Bench results show, and what they don’t.

dsh-superagent is DeepSeek’s MIT-licensed prototype: a routing agent that turns GitHub mentions into parallel tickets and speaks only when it matters.

Drex Max v1 is Nace AI’s open-weight 27B decision model: typed questions in, probabilities out, one forward pass, on a 24 GB Mac. What the data really shows.

qwenGPU compiles a 4-lane SIMT GPU into Qwen3.5-2B weights, running real CUDA in forward passes. How it works, what it proves, and why it stays a research demo.

Auto Ontology is NVIDIA’s open semantic layer that grounds agent SQL in certified business definitions. How it works, and what its missing benchmarks mean.

Zyphra Hierarchical Attention (ZHA) searches a tree of compressed keys instead of the whole KV cache, retrofitted onto a frozen 35B model in a short distillation. What the data really shows

Pando, OpenAI’s new Git backend, cut median push latency by nearly 80 percent and serves about a petabyte of Git data a day. How it works and what comes next.