50%+ of programmers still don't use AI tools to generate code in their normal daily workflow

When AWS appeared, engineers were hesitant to use it. Many of the arguments were the same as with AI tools like Copilot.

It's incredible that at least 50% of programmers I know still do not use AI/GPT and tools like Github CoPilot (or similar) in their normal daily workflow.

It reminds me a little bit of the rise of AWS. When it appeared, many engineers refused to use the cloud as it was unknown and people wanted to have more control over their tools and deployments. They felt safe in their comfort zone.

Deploying code to Amazon felt like giving up too much control over your intellectual property - some people were afraid that Amazon would access and steal their code or get hacked (implying their own deployment was more secure than Amazon's - which is funny from today's perspective).

We hear similar arguments about CoPilot etc.

Because of resistance, I think that the AI-assisted programming revolution will take longer than we think.

Yes, the tools are still imperfect and there are potential IP and other issues with them. There is also a learning curve and it requires a different mindset.

I believe all of this stuff will be fixed and forgotten 6-12 months from now and the tools will be seamless.

Still, there will be many people who will find a zillion reasons not to use these tools -- as always, the biggest roadblock in innovation is humans blocking progress because of the old ways of doing things.

Give it a year or two -- AI-assisted programming will be the norm, as it unlocks the "10x programmer" productivity. It drives cost down and iteration speeds up. That has real economic effects.

A normal programmer will literally be 10 or 100 times slower than an AI-assisted one, especially because of debugging speed, and you can bet that every single new young programmer will be using these tools from day one.

Until now, the 10x programmer was a myth. Now it could become a reality.