Huawei researcher's Cambridge talk: agent-era code bottlenecks shift to specs and verification
satnam6502 · x · 2026-11-18
Dan Ghica, head of the Programming Language Lab at Huawei Research UK, presented 'Code in the Age of Agents' at Cambridge's Logic and Semantics Seminar, arguing that autonomous coding agents have dismantled the assumption that writing code is software's primary constraint. As marginal code-generation cost approaches zero, the bottleneck shifts to comprehension, review and verification—or higher still to specification and validation. Agent-generated code differs fundamentally in style and failure modes (slop, context drift, abstraction leaks), forcing a rethink of what programming languages suit the agentic era.
More from coding & agent
- Oracle agent memory cuts input tokens ~10x at turn 80 vs full chat replay — bibryam · 2026-11-18
- 50-company study: median enterprise knowledge base doc is 505 days stale, 1 in 65 AI citations hits expired verification — femke_plantinga · 2026-11-18
- Plane AI now writes PQL queries for you, keeping them editable — JosephJacks_ · 2026-11-18
- SciML lead: Google Antigravity orchestrates 60 parallel issues, outclasses Claude and Codex — ChrisRackauckas · 2026-11-18
- Qwen3.8-27B OpenCode adapter lifts SWE-Bench Pro 35.3%→51% with 6-8x fewer tokens — yb2698 · 2026-11-18
- Hermes agent can now fine-tune models in-chat via Adaption's AutoScientist plugin — sarahookr · 2026-11-18