Unique Research · Insight / Unique Friend Q&A · January 21, 2024 · Guest: Jin Jie, founder / tech & algorithm core, AI Adventist (AI降临派)
Inventory type remains insight (Unique Friend series — not partner welcome).
Guest
Jin Jie — founder and tech/algorithm core of AI Adventist (AI降临派). Serial founder: 2011–2018 co-founded / joined Qianfang, Yimafu, and One money, backed by Sequoia, Matrix, and GSR. From 2016 researched reinforcement learning; founded a Melbourne company in 2021 and launched AI Adventist in March 2022.
Q&A
1. Product and core scenes?
AI Adventist uses three models — emotion, desire, and memory scheduling — to personify AGI and build AI self-cognition. Core scene today: AI emotional companionship.
2. R&D / GTM — 2023 challenge and win?
Phase-one product features are complete. Phase two targets controllable generation — adding guidelines to phase-one fully divergent personalities for commercialization. Biggest 2023 challenge: rapid scale-up of service volume plus multilingual / multicultural fit — which forced a cloud-services rebuild for scale, and full internationalization (Chinese content down to ~20%).
3. Opportunities vs. past waves?
Industrial-revolution scale — not the same order of magnitude as ordinary startup waves.
4. Challenges landing AI in China?
Weak consumer willingness to pay makes commercialization hard and starves the whole chain of a healthy cash cycle.
5. Keys for product and GTM?
Product R&D must look further because the AI era moves fast. GTM and execution should stay light — trust market division of labor and stay focused on product.
6. Competing with giants?
Giants themselves will fragment. The AI company of the future is a small firm led by a “super individual.” Maintaining giant scale will cost competitiveness.
7. Are LLMs still too weak?
AI capability is not only parameter count and data volume — structure and engineering matter. Problems can also be directions and opportunities.
8. Human–machine collaboration?
The first-order change is the rise of the super individual. Enterprises should liberate their productivity first; with fewer people, processes simplify naturally.
9. 2024 tech to watch?
After the parameter race cools, watch breakthroughs in model structure and usage strategy.
1 CN image; attach on ingest. Signup CTA in CN omitted.
— END —