The State of Tech

Everybody that is in tech already knows that AI is the big thing and the bubble continues to grow bigger. Interestingly, this also means that there is a lack of practitioning that is happening within the tech world. A lot of folks are now schilling AI software in hopes that it nets them huge profit in association with some company that they start, and the tech isn’t even meaningful. Think about the releases of LLM’s that are constantly in flux and continue to have no real bearing on reality. I would ask the question if these models are making life any better for people that are in industry.

I personally will say, that my experience has been that through trialling these things, in particular different models for programming/research, there has been only one quality of life change that I’ve found. Aggregation of research has become easier, but, ultimately, it has made the actual development process significantly worse. Furthermore, it literally makes you a worse developer and takes away all of the thinking that actually went into development in the past.

Here’s a non-comprehensive list of things that have generally gotten worse:

  • Web apps
  • Enterprise Software
  • Code Review
  • Security Research
  • Certainly more things…

Imagine you write a web application using AI, submit a pull request, some garbage AI tool reviews your pull request and says LGTM!! Then your PR gets merged, the application security team (or security researchers) check out your web app and use some scanner with AI built in to it which says that there aren’t any issues, just for somebody that is doing something niche or novel to come by and say that you have a race condition or business logic error that loses the company money.

Now, of course the above is worst case scenario without a human in the loop and oftentimes that isn’t the case, but I’m only begging the question if reliance is becoming an issue and killing our ability to be creative and solving problems?

The State of Conferences

This is where the real downfall is… community. Recently, I returned home from a conference disappointed. Although the above passage reads like I despise AI, the truth is that it’s a cool technology that used in the right way can accomplish a lot of things. I believe that the right way is not to get it to do the actual work, but to get it to do the surrounding research and the returns should be research that you might need to complete the work. As mentioned, what’s actually happening is that vendors are coming to conferences and paying sponsorship fees, taking a speaking spot as though they’re practitioners and talking about how AI stopped X attack or AI is great at code review, while they’re not doing anything truly interesting.

So what needs to happen?

Well… There needs to be a re-invigoration of practitioners that are doing things that are truly novel and unique and speaking slots need to be given to those people rather than companies that continue to push the agenda that AI is solving problems that more often than not, can’t be solved by a probablistic system.

Honestly, though, the hacker culture of old needs to come back.

Needless to say, I likely won’t be going back to the conference and if I do, I hope to see some changes there for next year.

With That Said

I think that AI and LLM’s are doing some cool things and the things that I’m saying aren’t necessarily all that interesting or new. My other belief is that there are other people that are working on more interesting things and that’s where attention (and resources) should be put.

Either way, my only hope is that community comes back and vendors recognize that they’re able to make strides in the community without schilling an AI product, but rather marketing themselves in a way that makes them look like a domain expert. Give a talk about the challenges around your architecture or how you’ve changed your teams methodology when doing VR and how it’s made your team better and give concrete examples. It would certainly make me more interested in your company and products than saying we built a system on top of insert SOTA here and this is why we think it’s changing the game!

If you are going to schill your AI product, at the very least make sure that it’s doing something novel that would save me time at the end of the day rather than regurgitating what X company over there already did.