Ideas No Longer Need Permission

I think I am addicted to AI. Not in the usual terms as people talk about addiction. It does not feel like a distraction. It feels very familiar, something which I have felt before, but haven’t felt in a very long time.

When I got my first computer, I had to choose between buying a computer or buying a motorcycle. I chose computer. This was semi-urban Maharashtra. Having a PC was not a very common thing. It never felt like a big decision but in hindsight it was decisive.

This was a Pentium 4 with 128 KB RAM. Dial up internet could take an entire night to download a 2 MB file. I could start download and wait for hours, sometimes get up and check if it had downloaded. That machine was my life. It was not like I was doing something meaningful. It was always about trying something new, installing something, breaking something, rebuilding and figuring out ways to fix it.

As we progressed, this phase ended or was replaced by work. Work brought structure. It made you cautious as trying things had a cost attached to it. The cost was not just time but effort attached to it.

There was a gap between an idea and executing it. Even a small idea required a significant commitment. You had to write code, wire up things, debug things that made no sense, and then push through all these frictions to show your idea the light of the day.

Because of this, you become selective. Not necessarily about which idea is good, but about which ideas were worth the effort. And this meant a lot of ideas never got a chance to exist. Not because they were bad, but because they were too early.

Ideas did not compete on quality. They competed on how much effort they could survive.

This is the part which has changed.

Vibe coding does not just make execution faster. It reduces the gap between thinking and trying. When this gap is small, you are not afraid of trying them. You stop over planning and start building. This makes path to validation shorter.

Most of these attempts still don’t go anywhere. But they don’t die before they begin.

And something else changed when cost of trying is low. You stop working on one idea. In my case it was an accident. As I was waiting for agents to finish generating, I moved to the next. I then worked on three ideas from extremely different segments together. This made me realize, Earlier the bottleneck was execution. Now bottleneck is the Judgment.

The hard part is not building but choosing what deserves to be built.

When effort was high, it acted as a natural filter. Only ideas strong enough to justify effort survived. This effort filter forced you to think before you built.

Now this filter is not there. It means the responsibility to filter is yours.

This redefines rules of success. Earlier those who were good at pushing through could win. In the changing landscape, those who can decide when to stop will win. They try more things but also abandon more things. They don’t just keep building just because they have started it. This all requires a different skill. Judgment.

There is a reasonable criticism to this way of working. Lowering the cost and effort of building means not just good ideas but every idea gets built. Filtering out good ideas will become more difficult. It is also true that systems built this way can be harder to understand and maintain. If you rely too much on AI generated code without engaging with it, you risk building things you cannot debug. The risk is real. But it comes from how you handle the shift.

In my case, I don’t ignore the code. I just don’t let code delay the idea. Earlier, rigor came before validation. Now it can come faster. I find myself building things I could not have attempted earlier. Some of them are serious. None of them started as a project. They started as a thought. What if I try this?

I feel software development is more alive than ever. Not because more lines of code is written, but because more ideas are being tested.

There is a risk in this too. What if you kept exploring and never ship? You keep iterating without choosing.

Earlier ideas died because they were expensive. Now they die because we don’t choose them. This is a very hard problem to solve. No tool can solve this.

If you have been sitting on ideas, this change matters. The cost of trying has come down to few tokens. You can’t hide behind old reasons. You don’t need to commit fully. You only need to begin.

Choosing is the hard part now.