I think when parsing that statement it's important to understand his (and your) definition of "programmer".

We (I) tend to use the term "programmer" in a generic way, encompassing a bunch of tasks waaay beyond "just typing in new code". Whereas I suspect he used it in the narrowest possible definition (literally, code-typer).

My day job (which I call programming) consists of needs analysis, data-modelling, workflow and UI design, coding, documenting, presenting, iterating, debugging, extending, and cycling through this loop multiple times. All while collaborating with customers, managers, co-workers, check-writers and so on.

AI can do -some- of that. And it can do small bits of it really well. It will improve in some of the other bits.

Plus, a new job description will appear- "prompt engineer".

As an aside I prefer the term "software developer " for what I do, I think it's a better description than "programmer".

Maybe one day there'll be an AI that can do software development. Developers that don't need to eat, sleep, or take a piss. But not today.

(P.S. to companies looking to make money with AI - make them able to replace me in Zoom meetings. I'd pay for that...)

There are almost no programmers today (you need to do malloc and low level sys calls in C to be considered a programmer).

I don't think you can be considered a programmer if you can't write your own syscall firmware code in assembly.

I don't think you can be considered a programmer if you can't perfurate your own punch cards.

That's right. We invented programming AI a very long time ago, and called it an "assembler". All you had to do was tell the assembler what kind of program you wanted, and it would do the programming work for you!

Then we invented another AI to tell the assembler what kind of program you wanted, and called it a "compiler". All you had to do was tell the compiler what kind of program you wanted it to tell the assembler you wanted, and it would do all the not-exactly-programming work for you!

And so on...

P.S. Visual Basic with its GUI designer was a quite effective way to rapidly build apps of questionable quality but great business value. Somebody should bring that paradigm back.

Excellent as long as you jump ship when the complexity is over 5kloc

The paradigm never left :)

My day job is programming in an environment which originated in the mid 90s. A contemporary of the Visual Basic era, but somewhat more powerful, and requiring substantially less code.

While I, and a few thousand others still use it (and it gets updated every couple years or so) it has never been fashionable. Ironically because it's perceived as 'not real programming'.

We routinely build systems with hundreds of thousands of lines of code, much of it founded in the 90s and having been added to for 25 years. Most of it was built, and worked on, by individuals, or very small teams. Much of it today is still active doing the boring business software that keep the lights on.

But its not "main stream" because programmers pick language based on popularity, and enterprises pick programmers based on language. A self-fueling cycle of risk aversion.

A lucky few though hot off the treadmill a long time ago and "followed a path less travelled by". And that has made all the difference.

The CEO of nvidia said basically the same thing so I don't know if that's the best metric.

You mean the CEO of the company that just rode the AI hype wave to become one of the top 3 most valuable companies in the world? It's his fiduciary duty to say things like that, whether or not he believes them, the same as every other CEO in the AI space.

There is no fiduciary duty to lie, or make up stories you do not believe.

People really oversell fiduciary duty. Yet the whole point of top-level corporate roles is to steer a company predicated upon opinion, which means that you have great latitude to act without malfeasance.

Unfortunately when Stability shuts down Wall Street might get spooked that the bubble is popping and downvote NVDA

The worst thing here is not that he doesn’t believe in it himself, but that he does. As George Costanza said, “it’s not a lie if you believe it”

No, my problem is that I don't believe he believes it. It is a lie.

Well it's been almost a year since he said that. You would think we'd have lost the first chunk of programmers to AI by now.

We have, there are a bunch of bottom of the barrel executives that have done hiring freezes and layoffs under the assumption that AI would replace everything. It will go the exact same way as the outsourcing craze that swept the industry in the mid aughts. The executives that initiate the savings will be praised and rewarded lavishly, and then when everything blows up and falls apart those responsible for the short sighted and ineffective cuts will be long gone.

Are there any good numbers on how many positions this has affected?

I'm pretty sure Elon Musk also didn't believe in "FSD in 6 months" every time he said it, but it was just marketing.