Software Engineers Are Facing an Existential Crisis As They Drown In Horrendous AI Code (Futurism, 6/25)

[highplainsdem]

https://futurism.com/artificial-intelligence/software-engineers-crisis-drown-ai-code

A lot of ink has been spilled on the programmers who got replaced by AI agents. But it doesn’t sound like it’s a whole better for the ones who avoided getting the axe, either.

According to Deedy Das, a partner at the VC firm Menlo Ventures, the rapid embrace of AI tools is tearing companies apart by creating a “class divide” between brainless vibe coders and experienced engineers. The veteran “craftsmen” engineers — the ones who actually care about their profession — are forced to wade through and fix the swamp of terrible AI code that comes their way, leaving them despondently questioning their livelihoods.

“Most software engineers are facing an identity crisis bordering on depression,” Das wrote in a lengthy X post, spotted by Business Insider.

-snip-

Some studies have explored this phenomenon. One harbingered the bureaucracy of “workslop” — shoddy AI-generated outputs that lazy workers pass onto their colleagues. This creates the illusion of increased productivity, but in reality needs to be corrected by a fastidious coworker. Brewing resentment ensues.

-snip-

I've seen lots of posts about this sort of thing on both Hacker News and tech forums on Reddit.

[BlueWaveNeverEnd]

1. how depressing

[highplainsdem]

3. I agree.

[Disaffected]

2. I'm not a SW Eng but have written a lot of code

from machine language to assembler to high level languages in the day to AI and, have found my recent experiences with AI written HTML and VBA to be very positive. It is not foolproof by any means - the functionality has to be carefully detailed and the code has to be tested and debugged but overall is IMO a great time saver and, importantly, facilitates coding (HTML in my case) in a language I have very little familiarity with.

So, I guess I'm a bit perplexed by the negativity - I dunno, maybe the detractors are referring to large integrated code bases with a lot of added complexity and security requirements etc.

[highplainsdem]

4. I've seen thousands of social media complaints about AI coding, many for the reasons described

in the OP. I've also seen articles and studies on people using AI to code being proven NOT to be saving as much time as they think they are, and in some cases spending more time on it than they would doing the coding without AI.

And I wince every time I read about someone using AI to code in a language they know little about, because I'm not sure they can catch any problems with the code.

Which is why vibe coding is particularly a mess.

I've read that people using AI to help them with coding in a language they should be more familiar with don't learn/remember that language in the way they would if they didn't use AI for it.

And I've seen lots of posts from security firms on AI code creating security risks.

I don't doubt that some people might find it useful for some smaller coding projects. But I don't think that begins to outweigh all the harm genAI does.

And AI coding tools, like other genAI tools, are built on a lot of stolen intellectual property. Even when code was posted online with a Creative Commons license, those licenses have certain requirements the genAI industry ignores. They're fundamentally unethical tools. Those AI coding tools' training data is almost certain to contain a lot of stolen books, since those are the basis of the chatbot's ability to communicate.

[Melon]

5. I spoke with an exec over major projects including AI

A few days ago. They converted all coders to basically supervising of AI roles responsible for checking and implementing code instead of righting it. His comments were as their in house AI has learned, the coding is just as accurate but the time savings is astronomical. The coders now review and implement the AI code. Those released were those that didn’t want to change. Everyone else kept roles.

[Disaffected]

8. That doesn't surprise me.

Producing a computer program/sw is much more than just coding - coding is often the easiest part. The more demanding parts are:

1. defining the functionality of the program i.e. what it is supposed to accomplish. We used to call that the BSD (Business System Design) and was always completed & approved before any "programming" was performed.

2. the System Analysis portion i.e. defining the program logic, computer resources required etc.

3. after the coding, the testing and debugging then follow (the testing and debugging being as important as any other stage).

4. documentation, both user and programming - often neglected but equally important.

As an example, I developed a browser app (HTML coding) that turned out to be about 5,000 lines of code. The AI agent Claude Opus produced that code in far less time that it would have taken me to even type it in. And, after a bit of enhancing & debugging, it worked fine. I was literally amazed at how well it worked - it even made a number of useful suggestions for improving the app. And it was dirt cheap - I cost about $20.

[Prairie Gates]

6. Let them choke on it

This is a problem that they created and celebrated over 40+ years.

The greatest event of karma in recent history is AI coming for the software engineers.

[fujiyamasan]

7. More alarmist nonsense frankly

These tools aren’t going away. Instead they’re getting better everyday. There’s a reason companies are paying for Claude code subscriptions and laying off a lot of overpaid software engineers. Coding is the one area where engineers will use them more than anywhere else.

There are a lot of concerns over AI and some are very valid, especially in some fields like education. But AI is exposing something far worse about the economy built over the last quarter century — that so many white collar jobs that were sold as the ticket to success provide little to no value aside from a paycheck.

What we instead have seen is endless degree creep, forced college debt, and endless paper pushing jobs. That supposedly dumb software could replace some of these jobs is more problematic because so many lack basic marketable skills anyways becuase college certainly didn’t give them that.

Time for UBI. The corporations will always try to save a buck. Humans can’t compete with a machine for routine tasks and just like factory automation, coding automation basically makes much of coding a routine task. Will there be mistakes? Sure, but test and script automation tools have been around for years. That didn’t stop them from being adopted. It just saved time with the testing process. The gen AI coding tools are the next step, and it just removes a lot of the grunt work. The engineers that embrace this will continue to find work and grow their skills.