Which software jobs will AI take first?

[InventedPrime]

I had to rephrase the question “Will AI take ALL programmer jobs” into this one because it makes sense. If it’s possible, what types of software developer jobs will be automated by AI or is at risk? Some would say that making decisions on what tools/frameworks a company will use is like the backbone as to why AI isn’t as close to causing a problem. However you can still ask it for pointers if you’re a good prompter and go from there… But anyways what thoughts do you guys have?

[readyplayernull]

What would happen if we had, not a general AI yet, but easy access to the knowledge required to perform any job?

I like this prediction by Alvin Toffler:

> People of post-industrial society change their profession and their workplace often. People have to change professions because professions quickly become outdated. People of post-industrial society thus have many careers in a lifetime. The knowledge of an engineer becomes outdated in ten years. People look more and more for temporary jobs.

https://en.wikipedia.org/wiki/Future_Shock

Instead of "losing" jobs (ok, some tasks of those will be replaced), there will be re-mixed jobs: dev-designers? doctor-programmers? more and more specialized branches of every permutation.

The only way to "lose" jobs, instead of transformed, would be by full AI replacement, which didn't happen in this AI summer.

[muzani]

The benchmarks are mostly Python. Python is mostly used by AI engineers. So most likely the AI engineers' job, lol.

Another angle is that it's based on training data. Something like Android and web changes drastically every damn year. We've seen people applying for jobs and using LLMs; the key indicator is that they're using architecture from 2018 or so.

It hasn't been well trained on game dev, either. Most of these are not benchmarked, so they tend to fly below the radar.

[OnionBlender → muzani]

Android is a good example. For Android questions, I find that stackoverflow and LLMs give outdated answers since Google keeps changing things in Android.

[jaykoronivo]

It probably enhanced productivity, which means one guy may do two or three people's jobs. But still, AI is not perfect for taking over entire web dev, UI design, or DBM jobs.

[deisteve → jaykoronivo]

backend devs now do all of those now

[mnk47]

I wasted too much time and energy trying to look for an answer to this, and the only one I ended up with was:

Frontend at higher risk, but other than that it doesn't really matter. If LLMs continue to succeed, no dev field is safe. We might all have to become product managers or systems analysts.

There is nowhere within dev that you can safely pivot to, because the most in-demand fields also have the biggest datasets (from online/PDF content) and if LLMs replace frontend dev roles, it won't be long before they effectively replace or transform backend roles, DevOps, SRE, gamedev, etc. It all largely depends on whether or not we hit a ceiling in the next 2-3 years as the next frontier models are released. If we're still rapidly progressing by then, it won't matter because by the time you've learned a whole new domain, Claude Opus N will be around the corner to catch up to you. If we do hit a ceiling, then it won't matter because, well, we hit a ceiling. So you can stay in your favorite dev field for the time being.

[gtirloni → mnk47]

Have you tried to develop a decent sized project with LLMs? What was the experience?

[JSDevOps]

Non Technical Project Managers. Don't need them just hook AI to the Jira Board and assign tasks to programmers.

[greatpostman]

Front end will go first. Designers will become front end engineers

[fabianholzer → greatpostman]

Web designers used to be the combination of what you'd today call UX designers and frontend developers. The technical knowledge and understanding in the UX-only tribe that seems to have become the majority is abysimal. They are nearer to product management than to engineering. The capability of their current tools enables to cheaply spit out high-fidelity prototype which are good enough to give the technically clueless business completely false impressions about the cost for the promised capabilities, ever increasing the demand and sophistication of the actual engineering part of the frontend.

I don't see that being automated by language models in the sense of that it would enable the technically unsophisticated designers of today to come up with even an actually deployable UI, let alone one that is maintainable, so that a change in requierements would not lead back to square one.

I'm willing to be a spectator in attempts of business to chase that folly, though. Once systems start to degrade engineering demand will go up.

[deisteve → greatpostman]

I'm already seeing it. Our BE team is now expected to churn out FE applications.

BE will generate FE templates and to wire it up to the API they built

FE, designers are screwed. BE is safe.

[coding123]

CRUD backends and their types can be easily boiled down into various statements about permissions. I don't think CRUD needs to be hand-written anymore. I think that will go first.

That said, I think the data engineering aspects will be around for a while. For example integrating outside data, analytical processing.

One domain that won't be heavily affected for a while is gaming. Gaming will only hire more and more AI developers to make the content more interesting/interactive.

[VirusNewbie → coding123]

> CRUD backends and their types can be easily boiled down into various statements about permissions.

RoR did this like 15 years ago. I don't see how CRUD is eliminated by LLMs if they weren't eliminated by the hundreds of frameworks and nocode platforms that people have been building forever.

[deisteve → coding123]

but crud is only 20% of BE

rest is biz logic and workflows and those are tough

[petabyt]

Web devs? I don't know what that would look like, maybe it would be a AI enhanced Weebly/Wix/CMS/no code platforms? Then again those platforms are fairly capable and didn't make much of any web devs obsolete in the past few years so I don't know what to think.

[aristofun]

The only job llm would be able to successfully take over is journalism.

Especially that anxiety-ignorance type ;)

[infamouscow → aristofun]

I am very confident in this prediction too.

Reminding these soulless demons on X of this inescapable fact is extremely entertaining. In addition, it's good to remind them:

Corporate journalists are the enemy of the people, and the enemy deserves everything that happens to it.

[aiagentsdir]

I think developers are first in the line. I'm collecting AI agents into a directory and I see a lot of real good options. You can check and filter here by category aiagentsdirectory.com

[matt_s]

I think you need to take a step back into the business realm first and ask this: what bullshit jobs go away with AI solutions? Those areas of software development supporting them are likely the first groups of developers that will become redundant.

I think there will be an inflection point with things like insurance claim processing, non-sales call centers, help desk/IT (any non-physical elements), time tracking, payroll and HR. Those areas of business are the ones that are cost centers, aka SG&A (sales, general and admin costs) in the corporation's P&L statements. Look at software jobs that support these areas of businesses.

If all the mundane tasks in those areas are mostly handled by AI with consistent, repeatable outputs then its cheaper and changing systems will be faster. The largest cost of changing systems in those areas is re-training staff on a new system. I think this means rapid consolidation of solutions in those areas (i.e. VC/BigTech money scooping up companies).