Are Programmers Obsolete? Will AI Replace Them?

[karolmusia]

If entry-level jobs disappear, how will highly skilled engineers be trained to work on AI-assisted projects?

[shaunjackson]

I'm a software engineer with 30 yrs experience of designing telecoms systems all around the world. My primary job was the technical delivery of the solutions I designed, but my secondary job was the mentoring of new developers so they could progress into my role in the future. My worry with ai is that these junior roles will not be filled by people and nobody will get trained into the senior roles

[Anankin]

Before reading: no and yes.

Junior programmers are indeed obsolete, though, but this is just a big problem because you can't have senior programmers if you don't have junior ones first

[tekperson]

I've been programming professionally since 1982. Every year since there have been predictions of the end of programmers. The technologies used to predict the decline have varied over the years, but the predictions have always been overly generous about the capabilities of the technologies. Last summer, I mentored several software interns and asked them if they were afraid AI would take their jobs - they just laughed. I do agree that if your job is easily automated with upcoming technologies, you should be worried. In general, if you are employed as a "spec to code generator, " then you should be trying to upgrade your skill set.

[xgu]

If all entry level programmers are going to be replaced by AI, then 30 years after, there will be no human developers.

[ewasteredux]

Hope you feel better Dave... and yes, I think you are right on the money. The days of trivial programming are limited but the days of problem solving are here to stay. The only thing that changes are the complexity and sophistication of the problems.

[jej]

You didn't mention the most important difference, from a business point of view, between humans and AI: accountability. Nothing mission critical can be entrusted to an entity that might hallucinate. Therefore, anything an AI creates will have to be thoroughly analyzed and understood by a human engineer. You can't just do black-box testing and assume it works in all cases, if you don't know how it works.

[SirKenchalot]

I suppose it'll be a bit like when libraries and larger codebases came along; suddenly you didn't get your hands dirty flipping bits all day, you just imported a library or 2 and glued together a bunch of functions from within. Maybe the job of the dev for now will be breaking the task into sections that AI can build, then adding the glue, until the sections get large enough that no glue is required.

[knotfinley]

I once was the main programmer on a system and trained 3 others to do what i did. The only problem was when the system had a major issue none of them had the ability to solve it. I was always called in, that's the indispensable part. My boss at the time asked me to write down the process I went through to solve issues. I told him I have no process and in order to do what I was doing you would need my experience also. I've never figured out how I did it, best I could come up with was stay clam and be persistence.

[WilliamHaisch]

I’ll probably be the guy screaming *“I have people skills!”* to the AI-powered efficiency consultants. 😂

[roody_io]

glad to see your channel pop back into my yt feed. always quick to tune in and hear some wisdom from the sages.

[Micha-pfpl]

so much clarity here mate, kudos to you

[DevilsHandyman]

I've been programming professionally for 38 years now and one thing I know for sure is that you have to adapt and change. I did programming in BASIC and even some database programming languages before I wound up programming in C and eventually C++. Still learning new things as over the years since I have done C# and Python and a few processors assembly languages. You have to stay curious and try new things if you want to stay competitive.

[frubert]

My experience of working in the industry is that people still can't write good specs-at least for "applications" that they want. I'm not talking about "take this number and shift it left" or whatever, but for high level apps, people always miss out what to them is obvious, but which they don't realise is specialised domain specific knowledge. So the functionality isn't as expected but they are genuinely surprised.

So I can see the new market Dave describes for "the AI operator and input refiner".

[Jason-mknn]

This was a great video and in an interim of great change, it can serve to show the way forward and motivate those that were not quite sure which direction to pick at the mlm in the road. There have been great forks in the road in the past, and there will certainly be in the future, but it was a nice sanity check to help others avoid a potential fork bomb.

[douglasgoodall]

Dave, I bought your book. Thank you for taking the time to share your experiences and wisdom acquired. 🙂

[funknick]

I agree with Dave with his predictions. My career as a senior engineer who architects large changes for a system to handle complex user requirements isn't going anywhere anytime soon. How it gets written might change, but the up-front knowledge of knowing what can and cannot work hasn't changed. You still need someone who fundamentally understands the science and algorithms enough to apply them. Maybe AI just writes all the glue code, but it'll still be humans designing and verifying its input/output process every step of the way.

[dtran]

Thanks for the video. I mainly work on the firmware side designing safety critical systems for medical and aerospace. Coding is just a tool that I use to solve real world problems. It's so much more than coding so I wouldn't mind ai taking over that part but i doubt it would anytime soon.

[josephgarner]

Thanks for the insight Dave. I work in Infrastructure but I'll adjust my skillup efforts accordingly.

[muddyexport]

Good vid and info. Prior to retirement 80/20 debugging/development and I said that prior to you saying on the vid. Level II support. And keep up the good work and fight!