'Capability gaslighting' is real—but not in agentic engineering, where code is consistent
PawelHuryn · x · 2026-09-25
Maggie Appleton (GitHub Next) coined "capability gaslighting": models impress you on one task then fail on another, making you feel deceived. Pawel Huryn adds a distinction: in agentic engineering it barely applies—mistakes are checkable and answers often binary, so code performance is very consistent. Taste and design are where gaslighting lives, since nothing flags wrong answers and mistakes are hard to articulate.
More from Models
- Could the Jev model add reasoning offloading and N-gram augmentation? Fan speculation — ProposalOrganic1043 · 2026-09-25
- Redditor: models are now good enough that I stopped caring about prompt engineering — Large-Excitement6573 · 2026-09-25
- User says OpenAI's Sol 6 wastes their time, switches to $100 Claude plan and prefers it — AirportEither2456 · 2026-09-25
- Nagi-ENORMOUS beats Jev, Semif, Laya on Game Arena Benchmark — No_Skill_8393 · 2026-09-25
- Principia: video models know less high-school physics than you'd think — mariyaivasileva · 2026-09-25
- Post-scarcity intelligence defined: Opus 5.5-level smarts at $0.10 input / $0.50 output per million tokens — Sect-Sister-Reads-42 · 2026-09-25