re: AI is killing software developer jobs

quote:

---

Coding bootcamps were dying long before AI took off.

---

As they should. Basically scams. Coding is a pursuit that certain people have a natural gravitation towards. These boot camps give the illusion to some people that they can do this, but they can’t. Coding takes a certain mindset to do it day in and day out at a high level. I don’t know why we started putting value on these things. Maybe it was HR or the fricking media.

You don’t need to go to college to code, but you better have a mind for it or you’ll fall out in 2-3 years of doing it professionally.

quote:

---

You don’t need to go to college to code, but you better have a mind for it or you’ll fall out in 2-3 years of doing it professionally.

---

We gave up hiring anyone who wasn't a college grad many years ago.

Every now and then you got someone, but that was rare.

People with the degree just understood the fundamentals so much better, they could generally actually problem solve.

Paying someone just to code doesn't actually exist. That's not a real job

quote:

---

But it's just like an Engineer needs to understand the math that the software does for them in the real world. Without a solid fundamental base in database design and SQL programming skills, AI won't get you across the finish line with any degree of data integrity or accuracy.

---

I think this touches on the misunderstanding people have about how white-collar jobs get replaced by “automation.”

It isn’t like being a worker on a factory floor who gets outright replaced by a machine. It’s all about productivity. I think the engineering analogy is a good one. If you look at the design tools available today vs. 30 years ago, they have certainly made engineers more efficient. I don’t think most engineers would say those tools have killed engineering demand, though.

Meanwhile drafters are an entirely different story. CAD hasn’t necessarily made “drafters” obsolete (you still need CAD operators after all) but it’s made them so efficient that there just isn’t as much demand for them as there was before CAD became the norm.

So what’s the difference? In my opinion, it comes down to the question of where the bottleneck in a given industry lies before and after automation, and the question of how much of the job is replaced by automation. When workers become more productive, the cost of -insert work product- goes down. But as long as the rest of the industry (e.g. the demand for that work product) increases accordingly, times are still good for everyone.

In the case of engineering, I think we’ve seen that the demand for actual design work has kept up with increased productivity while the demand for drawings, specifically, has not. As a result, I think we’ve seen a shift in the industry from having folks who strictly work on creating drawings to having designers who do design work and CAD work.

I think AI will likely be very similar. Some people will simply become more productive and probably have an expanded job description as a result, while other less skilled/capable people will fall out of demand.

quote:

---

Paying someone just to code doesn't actually exist. That's not a real job

---

True statement

Learn to coal mine..

quote:

---

So what’s the difference? In my opinion, it comes down to the question of where the bottleneck in a given industry lies before and after automation, and the question of how much of the job is replaced by automation. When workers become more productive, the cost of -insert work product- goes down. But as long as the rest of the industry (e.g. the demand for that work product) increases accordingly, times are still good for everyone.

---

Bingo. When I started in this field in 2001 I supported 1-2 systems at a time. I now support 5 in all facets, from administration, dba maintenance/tuning, to reporting/kpi/dashboards from said systems.

quote:

---

It's going to be the biggest job killer in history,

---

Yep, the tech is almost there with things like the LLM.

AI is still terrible at coding big picture things. Good luck using it to set up a complicated website with hundreds of routes, etc.

But junior level developers are in trouble.

It’s not just SWE’s though. AI will eventually hit everyone. People used to hire graphic designers for $500 to do an album cover (music), now all the indie artists (who far outnumber the Taylor Swift’s of the world) are saving money having AI do the album covers for them.

The coding boot camps are dying also because companies are shifting more to prioritizing 4 year degrees. Which is stupid, because the students coming from bootcamps tend to be way better coders than people with 4 year degrees, simply because coding is literally all they do in bootcamp.

Bootcamp is like a trade school for coding. I think they just got over saturated.

quote:

---

It’s my understanding the AI we are seeing were not ready to be released.

One company released a product that was maybe 70% and got a butt load of press and attention in the medias. The 70% ready product got a lot of oooohs and aaahs so the other companies figured why not get their share of the spotlight and released their own 70% ready product.

---

I wouldn’t really characterize it that way. AI/ML work has been surprisingly collaborative. There are certainly some big-time players who try to keep their secret sauce hidden, but there’s also a ton of open-source work and published research.

Many of the LLM’s that are available have been made public for research purposes. OpenAI probably led the way with monetization of GPT, but we are really just now reaching the point where LLM’s are being integrated into other tools in a way that actually has a real commercial impact.

I think it's an opportunity for devs to pivot. AI still needs SME to translate technical value into business value for executives and non-technical owners. Any software developer who learns bed side manner will 100% be better off than they were before.

Junior developers should definitely consider technical project manger or sales roles.

AI is still very much a shitshow of a technology. People don't realize that the real money, opportunity, and challenges are more in Data quality and transformation. Everyone wants AI solutions, but a very little subset has the data readiness for it.