The following observations are based on an informal survey of developers I know, with some of my own bias sprinkled in. This is not scientific, it's sentiment based, but the patterns were pretty consistent in my discussions. Consider this "directional" information.

• AI-assisted coding offers roughly 20% overall uplift in programming productivity.
• The size of uplift varies greatly depending on the task, from say 10%-90%.
• The uplift is greatest when navigating a new language, unfamiliar APIs, or a new problem space.
• The uplift decreases as the problem space gets more specific or narrow.
• AI is much better at writing new code than editing existing code.

While none of this is earth-shattering information, I felt compelled to speak up following Mark Zuckerberg's recent assertion that his company will begin replacing mid-level developers with AI in the coming year. An extraordinary claim like that needs context.

Based on what I'm seeing, AI is not going to "replace" developers in the next year or anytime soon. In order for that to happen, you have to believe the following:

• A developer's sole job is to write code. (It's not.)
• Product owners always clearly articulate what they want the product to do. (They don't.)
• Customers always accept what is delivered. (They don't.)

What is going to happen in the coming year is that companies will shift their investments based on their needs and how their teams are performing, just as they do every year. I think Mark was really hypothesizing that AI provides enough of an uplift for the best junior devs that it puts the weakest performing mid-level devs at risk of being laid off.

A couple thoughts. Bear with me as I go off the deep end for a moment.

First, the way he phrased it conveniently makes his company more alluring in the eyes of the stock market, so it's important to take it with a grain of salt. The hype cycle is real.

Second, when was the last time you worked for a CEO who publicly stated he or she was looking forward to eliminating your job? I'm no expert but I'm pretty sure that's not good for morale. The way that leaders and investors talk about eliminating jobs is shockingly gleeful and ultimately short-sighted.

I have an investment pitch in my inbox projecting that by 2050 that 40% of the workforce will be replaced by humanoid robots. I don't believe that, but I could easily be wrong. What if it's true?

Let's use round numbers. If the US GDP is $30T, consumers represent roughly $20T of that. Eliminate 40% of income and you've removed $8T from the GDP. The systemic effects on financial markets would be devastating. And keep in mind this is just for so-called "blue collar" manual labor jobs. The systemic effects of a similar reduction it white collar jobs would be multiplicative in its effects.

Of course, this gloomy assessment doesn't account for new opportunities created as a result of technological progress. So without a doubt this is an extremely pessimistic view. Should it come to pass, the reality would likely not be as bad.

But what I find interesting about that pessimistic scenario is that it's possible and in that dire scenario everyone loses. Wouldn't it be ironic if, in a coordinated attempt to create massive economic gains through automation, we collectively did the opposite?

I personally don't believe the automation will advance as quickly nor as far as currently projected. This belief is based on the fact that historically the projections of automation progress over-estimate the speed at which improvements will occur.

That said, I could be wrong. I also could be the horse saying, "those ugly and slow metal horses will never be as good as me." So I continue to keep my eyes and my mind open, and stay optimistic that we'll figure this all out and land gracefully in the end. And I work to keep the panic to a minimum.

What to do with all of this? Same as always: stay focused on solving problems, stay on top of the current trends, be open to new potential, but be skeptical of claims that sound too good to be true.