← all notes

AI Is Compressing Timelines

June 2026

Something has disappeared from the process of building things. Not effort. Not difficulty. The queue.

For most of modern history, “I know what I want to build” was followed by a long period of acquiring the ability to build it: learn the skill, hire the person, raise the money, wait for the roadmap.

That delay shaped how we thought. AI is collapsing the gap for a large class of work.

Given enough compute, execution is rarely the bottleneck for software.

But this compression is uneven. AI compresses implementation, not every source of delay. It does not compress regulation, capital, trust, adoption, or human coordination.

This is why expectations have shifted. Investors, users, and employees are less willing to accept a promise that something useful will exist in five years: what exactly is taking five years?

If the problem is clear, the output is digital, and the main constraint is implementation, the answer cannot simply be “it takes time.” Increasingly, it takes compute, judgement, validation, and access to the right context.

Some work has earned the right to ask for patience. Physics has it. Biology has it. Safety testing, physical infrastructure, distribution often do too. Most product roadmaps do not.

Ask yourself: can you justify waiting many years for a single improvement in the digital world? AGI, maybe. AI made supply responsive. It did nothing to make demand coherent.