Four months is the new eighteen

I used to tell people it took 18 months to release a product if you released it quickly. That sounds like a contradiction, but it wasn’t. Eighteen months was the fast version because that’s how long it took to discover your initial direction was wrong. If you spent all 18 months building and then released, the product would fail on day one. The timeline was never about construction. It was about learning.

If you released early instead, you’d keep improving, and the product kept evolving and the old 18-month wall never quite caught up with it. The product was already past it.

Timelines are compressed now. Some say hours, some say months. A discussion on Hacker News landed recently where someone thought about a product for eight years and built it in three months, and the comments found the timeline refreshing. So are the people who say they built something in a weekend. There is truth to both paradigms. Complete means different things for different people.

The time to get started is hours, not months. That’s what collapsed. What used to take weeks of wiring things together now takes an afternoon. You describe what you want and it exists. Starting collapsed. The bottleneck moved to learning what users actually need.

Once you’re running, the feedback loops can be that tight. Something real ships in the first week. It gets in front of people. You find out what you got wrong and fix it the same day. Each round of learning that used to take a quarter now fits inside a week. What used to take 18 months of discovering what users actually needed now fits inside four months, because you find out faster. The learning compressed, and that was always the long part.

With the right harness, customers literally talk to the agents building your product. The gap between what someone asks for and what gets built shrinks to the length of a conversation. When that loop closes, the timeline question stops mattering.