a16z's Josh Elman: AI cut build costs to near zero, but product judgment costs the same
xiaohu · x · 2026-09-16
a16z partner Josh Elman argues that while AI coding tools have collapsed the cost of building software, the cost of making the right product calls hasn't dropped at all.
- A PM's real artifact is a story, not a spec: Elman recalls answering "spec" in his LinkedIn interview and writing a 120-page spec for a recruiting product, but years later concluded the true deliverable is a shared, retellable story of who uses the product and how it changes their lives.
- The dev loop flipped: expensive engineering once forced 6–8 cycles a year of upfront spec/costing/scoping/design; with AI it becomes idea → build with AI → play with it → visual/UX & engineering design → ship → learn.
- New risk is speed: teams can now churn out large volumes of mediocre, unwanted features faster than ever.
- He illustrates with his RealPlayer ad-revenue clash and lessons on helping teams ship the right product.
This is a Chinese write-up (xiaohu.ai) of Elman's piece; the latter part is members-only.
Related event: a16z: AI Cuts Build Costs, Not Product Judgment Costs(3 posts)→
More from AGI Musings
- vikhyat predicts alternating compute oversupply and undersupply on the road to AGI — TheZachMueller · 2026-11-02
- Sam Altman: AI has surpassed top scientists, and 2026 will be a turning point year — kimmonismus · 2026-11-02
- DeepSeek design doc: a person can only follow 3-5 agent sessions at once, attention is the real bottleneck — heyshrutimishra · 2026-11-02
- AI researcher: we have no reliable way to tell whether an AI system is conscious — dioscuri · 2026-11-02
- Toby Ord: track both what models can and can't do to see real capability gaps — tobyordoxford · 2026-11-02
- Forward Deployed Verifier will soon be as valuable as FDE, argues poster — vaibhavbetter · 2026-11-02