Hywel Carver at LeadDev West Coast 2023
summary
Claims that LLMs already handle implementation and boilerplate extremely well and will make ticket-to-code work near-instant, while humans must keep responsibility for architecture, functional review, shipping and monitoring. Cites hallucination (Billy Shears) and malformed-output examples, warns juniors lose learning if coding disappears, and proposes moving from passive to structured interactive learning using Bloom’s taxonomy so engineers become reviewer not author and teams retain deep system-level skills.
AI-written summary · about
chapters (7)
- 0:00AI and software engineering future
- 4:51AI risks and opportunities
- 9:19Just let me code mindset
- 13:31AI-assisted programming paradigms
- 17:03Team skills for AI development
- 21:14Learning for AI-era teams
- 24:02Closing remarks
related talks
03mSteve Yegge, Sourcegraph, ETLS Vegas (09)
54mTech Truth: Agile Evolution & the Future of SW Engineering
10mRethinking collaboration: LLMs, teams, and cognitive load
16mVibes won't cut it
38mNobody Pages the LLM: Engineering Rigour for Vibe Coding
44mJoint Keynote Address: Accelerating Software Development: The LLM (R)evolution
01h 02mAround Learning
26mMore. Better.