PhD thesis explores intra-turn interaction: AI interfaces beyond turn-based dialogue
OfirPress · x · 2026-11-13
Researcher liuxingyu argues that today's AI interfaces remain largely turn-based. His PhD thesis explores what interaction could look like beyond taking turns, demonstrating "intra-turn interaction" — letting users step in and steer while the model is still generating or acting. OfirPress amplified it as a cool idea; the work is relevant to designing more fluid agent and assistant interactions.
More from Research
- Qwen3-32B Has a Body Schema: Wheels Map to Heels, Grippers to Palms — Sauers_ · 2026-11-13
- scaling01 replicates looping transformer: 4B model reasoning climbs 42.9→52.4 over 4 loops — scaling01 · 2026-11-13
- Continual learning series: keeping training data clean when feedback, noise and bias creep in — goyalshaliniuk · 2026-11-13
- 7 Engineering Challenges in Building AI Systems That Learn Continuously — goyalshaliniuk · 2026-11-13
- LLM edits 65 embedding rows to fix robot failures, lifting held-out success 74.3% to 87.5% — 机器之心 · 2026-11-13
- Tsinghua's CoACT cuts Coding Agent tokens by 33% while slightly improving pass rate — 机器之心 · 2026-11-13