Open models recap: more on Kimi K3, Qwen 3.8, Xi's WAIC speech, distillation, the open-closed gap, and what's next
Interconnects Nathan Lambert ● Covered by 12 sources
Nathan Lambert and Florian Brand discuss the rapid advancement of Chinese AI models like Kimi K3 and Qwen, analyzing performance gaps between open and closed models and the role of post-training optimization. They cover benchmarking challenges, infrastructure requirements for deploying large models, and how the open-model ecosystem is maturing in infrastructure and tooling. The conversation explores how open-weight models can be fine-tuned for specific tasks and the geopolitical implications of China's commitment to open-source AI development.
Why it matters
A podcast with Florian Brand.
Also covered by
- The New Stack — Moonshot launched Kimi K3. Then demand shut down subscriptions in 48 hours.
- Latent Space — [AINews] not much happened today
- Interconnects — Kimi K3: The open-weights escalation
- Zvi (Don't Worry About the Vase) — On Kimi K3: Its Capabilities And Related Discontents
- Exponential View — 📈 Data to start your week
- Import AI — Import AI 465: Open vs closed gaps; Kimi K3; Demis' big policy plan
- The Neuron — Kimi K3 Pushed Open Models Toward the Frontier
- TheSequence — The Sequence Radar #897: Last Week in AI: China, Compression and the Open-Model Race
- Latent Space — [AINews] not much happened today
- The New Stack — Kimi K3 tops Arena’s coding leaderboard — and it’s open-weight
- Latent Space — [AINews] Kimi K3 2.8T-A50B: the largest open model ever released; Opus 4.8-class at Sonnet 5 pricing
