Ideas are ___, Execution is ___

It’s incredible to see how good AI has gotten at programming and building stuff. I use Cursor almost exclusively now, and have noticed myself prompting more and coding less. Tools already exist to create full-stack apps from scratch, using only a prompt.

I started my undergraduate degree in Aug 2021, around a year before ChatGPT was released. And if you had told me then that this was how the state of programming would be right now, I would’ve laughed and told you to stop reading so many sci-fi novels.

Yet here we are.

How will software engineering look like 1 year from now? 2 years? 5 years? I’ve no clue.

The rate of change is too high to even predict how things will like by the end of 2025. We’re living through the world’s fastest technological revolution!

Nevertheless, I think at least in the near-future, technical skills are going to remain important. Even if you use AI tools, prompting them correctly and effectively is necessary for them to do tasks right now. You still need to specify some intermediate steps or break the problem down for the AI to solve. To do this, you need to know (at least on a high level) how to do the task yourself. This will eventually change as models get better, but until then, technical knowledge is useful at least insofar as enabling you to write better prompts to generate accurate code.

But even though I can’t predict how tech jobs will look like in the future, I think the real (and more important) question is: What skills are going to be valuable in a world where AI can do most of the intellectual work we do today?

Until now, the advice given to young founders and startups was mostly: “ideas are cheap, execution is expensive”. This is clearly changing. With AI, the time and effort to build things on the internet has been going down dramatically.

Does this mean the maxim should be reversed to “execution is cheap, ideas are expensive”?

Honestly, I don’t think so. In my opinion, the “execution” in the maxim was never really about technical execution, i.e., the technical skills needed to build something. It was more about the discipline and the mentality to execute. It was about the consistency and persistence to push through difficult times and brick walls. It was about being relentllessy resourceful.

So, in that sense, execution — having a strong conviction and high agency — will become even more important. The difficulty is just going to move from solving technical challenges to solving business problems and the increased competition.

And the real question is: Even though you have the power of AI to build whatever you want, will you? Will you take the initiative and make use of the resources you have?

I still think ideas alone are not going to be enough. Even with AI and all the possible resources one might need, there will be a lot of people with ideas who will not act on them (or see them to fruition).

At the same time, as a natural consequence of more products entering the market at a rate never seen before, branding and advertisement are going to become more important.

The reason is simple: the barrier to entry has gone down significantly (think: anyone who knows basic programming can build an app) so while there were once 1-2 apps in a niche, now there are going to be a dozen. This means you’re going to have to work harder to differentiate yourself from the crowd.

So, when everyone can build, the game isn’t about building anymore. It’s about finding distribution and making people care about what you’ve built. ¹

Footnotes