Organizations that don't force AI

Work
1

Hey folks,

If you happened upon my intro thread reply, you may have seen this coming. Simply put, it’s a fair statement to say that most of us aren’t interested in using AI to code for various reasons. It’s also a fair statement to say that a large amount of software companies and tech companies are pushing encouraging, or flatout mandating AI usage for either coding, or everything everywhere as much as possible. This is not a thread to get into the reason and analyze what they’re doing. This is a thread focused on solutions for us. That is, what does a professional software engineer or (in my case) software security engineer, etc do now? How do we put food on the table and pay the bills going forward in this current landscape? Now, the reality is, there’s a ton of software developers and firms out there, and just because mine and a few friends’ is forcing AI doesn’t mean they all are, I’m aware… But also seeing some of the replies here makes me think it really is a more widespread issue right now.

So yeah, I’d like people to list here any organizations or companies that they know of (for a fact) does not force AI usage, or perhaps even downright bans it. This is not a category I’ve been able to find on job sites :wink:

But that’s not all, eventually I have a feeling that some of us may have to create one and prove that it can still be done, be profitable, be enjoyable, and we may be able to outperform companies who are so heavily pushing AI. So, if anyone is interested in that, do tell as well; it’ll be a marathon, not a sprint… The point is that, while I’ve had my qualms with AI, I’ve also sat around and listened to everyone complain about it, with no real solution, and taking no real action. This is not going to help us… Eventually we’ll all be jobless grumps spinning in circles complaining and I don’t know about you folks, but that’s not the way I want to live the rest of my life or career. So please do feel free to discuss this topic, give us any leads, reach out to me about future non-AI Projects, collaborations, etc…

6 Likes
2

I have a feeling that some of us may have to create one and prove that it can still be done, be profitable, be enjoyable, and we may be able to outperform companies who are so heavily pushing AI. So, if anyone is interested in that, do tell as well; it’ll be a marathon, not a sprint… The point is that, while I’ve had my qualms with AI, I’ve also sat around and listened to everyone complain about it, with no real solution, and taking no real action.

My perspective on this is that I have never worked in the industry, but that I finished a bootcamp - fully funded by a “Women in Tech” scholarship - just as the mass layoffs in software began.

Back then, it wasn’t about AI.

The lay-offs were a perfect storm of the covid lockdowns ending, C-suite execs introducing return-to-office mandates and the end of the ZIRP era (zero interest rate policy) for central banks.

In the early days of Chat-GPT4, many engineers wanted to use the new tools at work but were actually forbidden from doing so over concerns about sensitive company data, possible IP lawsuits and the dubious reliability of the tools themselves.

But now, the new tools are often mandated and rarely forbidden at work.

Its actually funny to me that engineers on Twitter in the Before Times were always banging on about tech debt, how you should keep your PRs really small, how organisations should have rigid self-imposed standards, how LOC is a stupid metric, how bugs increase with the size of a codebase etc.

But today, they are saying “don’t look at the code” and bragging about the vast quantities of non-deterministically generated slop that they are merging into production.

It seems to me that the only thing that caused the shift was the attitude of the business executives and venture capitalists towards software.

In terms of a fight-back strategy, I have heard that many freelancers are offering Slop Clearance As A Service now (SCAAS?). If tech debt is the problem that isn’t going away, then perhaps tackling it is the way forward?

Of course, AI is really bad at novelty. It can remix what was in the training data, but it can’t learn anything new.

Many of the most famous people in software aren’t there because of technical brilliance, but because they were business savvy, well-connected, just technical enough to produce MVPs, and they worked on the right thing at the right time.

4 Likes
3

I recommend listening to Jake Wharton on the JetBrains podcast. He left Cash App before they fired 70 % of their employeed. His new employer had to align with values that he defined before - one was not to be foreced to use AI (timestamp 15:53). After 3 months, he joined Skylight since they “echoed all the values” (timestamp 37:40). I also recommend the rest of the interview, which is half about the job market, reasons not to use AI. The other half is about Kotlin and his career.

I looked at their job offerings. Their Senior or Staff Product Security Engineer lists “comfort working with LLM-based systems” as requirement. But it makes sense for this particular role since Skylight has added purely optional LLM features, which were “tastefully added” (according to Jake) and need to be checked for security issues.

1 Like
4

The only ones I know are FOSS, so not really a way to put money on the table sadly.