​

Roadmap

BLAST is actively being developed by a team at Stanford University, with a focus on building a lightning-fast and cost-effective serving engine for browser-augmented LLMs. The moonshot goal is to complete tasks that previously took months of research and human collaboration in seconds. It’s hard to imagine how much society would be transformed by such a drastic reduction in latency. As an analog, I would think about how much the world changed when we went from library trips to Google searches. BLAST is a similar step towards a future where we humans can be far more creative and inventive.

​

Focus

• Stability gotta catch dem bugs
• Cost optimize under constraints
• Performance need to get way faster
• Human-Teaming should allow maximally natural human-AI interaction
• Reliability benchmarking

What we’re not focusing on is building a better Browser-Use, Notte, Steel, or other vision LLM. Our focus is serving these systems in a way that is optimized under constraints

​

Contributing

Please do! See our Contributing Guide.