Unique Research · Insight / Unique Friend Q&A · January 18, 2024 · Guest: Qin Xuye, CEO, Future Speed (未来速度)
Inventory type remains insight (Unique Friend series — not partner welcome).
Guest
Qin Xuye — CEO, Future Speed; SJTU graduate; formerly Alibaba; years in distributed computing, AI, and big-data systems.
Hangzhou Future Speed Tech builds enterprise AI infrastructure. Its Xorbits model platform helps enterprises deploy and connect foundation models — private or cloud — and enhance models with owned data. Team from internet majors and Fortune 500 firms.
Q&A
1. Product and core scenes?
Xorbits model platform deploys open-source foundation models in the customer’s environment (private or cloud) with enterprise HA, performance, and observability, plus integrations to Dify, LangChain, LlamaIndex and other AI developer tools. Live in finance, government, healthcare and more.
2. R&D / GTM — 2023?
Active R&D; open-source edition liked by the community; GTM progressing. Biggest challenge: AI moves so fast — how to keep up and still ship something useful.
3. Opportunities vs. past waves?
Nearly everywhere — every industry can be reshaped by LLMs. Scope of impact dwarfs prior waves.
4. Challenges landing AI in China?
Delivering the promised effect — model uncontrollability and hallucination make landing hard.
5. Keys for product and GTM?
Collect user needs fast. Build only around them. AI evolves so quickly that founder “hallucinations” (clever ideas) abound — better go to users and ship only what they need. Execution: MVP and speed; lowest cost, fastest delivery.
6. Competing with giants?
Do what giants will not; focus on useful product.
7. Are LLMs still too weak?
Models have limits — apps must match today’s capability.
8. Human–machine collaboration?
Copilot-style human–machine collaboration is inevitable; enterprises should embrace it.
9. 2024 tech to watch?
Open-source models reaching GPT-4 level — that unlocks a lot.
1 CN image; attach on ingest. Signup CTA in CN omitted.
— END —