Eric Ries#368
1/ I spoke with @ericries on @analyseasia about how #AI changes the #LeanStartup paradigm.
The big insight: AI makes building faster, but not necessarily learning faster.
Watch the conversation here as Eric also talks about his new book #Incorruptible: https://youtu.be/WuHchkMO-Ts
Users agree AI makes building faster and cheaper, making Lean Startup principles more relevant by emphasizing rapid learning as the key scarce capability.
Success makes companies a target—not a shield.
@EricRies joins us on Incorruptible, governance for mission-controlled companies & what AI can’t automate in Lean Startup.
Read/listen at https://open.spotify.com/episode/0NLowrUXxvedcrSEjPJokT?si=8T6LjWAMR4O6j6aGbrrtBQ
2/ In the original Lean Startup, the challenge was building under high uncertainty.
The answer was not “ship more features.”
It was validated learning: test assumptions, learn from customers, and reduce uncertainty.
7/ My takeaway:
The Lean Startup is not obsolete.
It is more important.
Because when AI makes building cheap, the scarce capability is learning faster than everyone else.
Listen on Spotify: https://open.spotify.com/episode/0NLowrUXxvedcrSEjPJokT?si=8T6LjWAMR4O6j6aGbrrtBQ
5/ That is the founder trap in the #AI era.
More output is not the same as more learning.
A faster prototype is not the same as product-market fit.
A working demo is not the same as customer trust.
3/ AI changes the speed dimension.
Today, founders can generate code, landing pages, MVPs, pitch decks, and proposals in hours.
But when everyone can build faster, speed alone stops being the advantage.
4/ @ericries made a great point:
If one person uses AI to turn 3 bullet points into a long proposal, and another uses AI to summarize it back into 3 bullets, did productivity improve?
Or did we just create the illusion of progress?
6/ What has changed:
Cost of building has collapsed Iteration is faster MVPs are easier More people can start
What has not changed:
Customers decide Learning matters Judgment compounds
@bernardleong @ericries @analyseasia i agree speed wins, but the real bottleneck is auth & state. i cut iteration time 3× by swapping openai → groq free tier and adding a tiny redis cache layer. how do you version your system prompts?