Adapting to this Age of AI
It’s a bit of a divisive topic depending on who you ask, but one thing is for certain: AI is changing the development landscape. I’ve gotten pretty good at wielding it, and fortunately I work a job where I have the privilege to access some of the more advanced AI tooling for productivity. I’ve also seen the downsides of it (and believe me, there are a lot of downsides), but a lot of that I chalk up to a lack of experience from people who use it. In my opinion, generative AI and its related tooling are great resources that can help speed up processes that would traditionally take much longer to complete. I’m not afraid to admit I wouldn’t have been able to get this website up and running so fast without it.
Now, with speed comes the caveat of quality, and you can just look around here to see what I mean. Some things about this site might seem a bit off in terms of design or usability, most of which comes down to me generating lines of code that looked “good enough” and I couldn’t fully scrutinize. I don’t know how every facet of this place works, but I have a general idea based on years of web development experience, that’s kind of just “good enough” for a project like this. Nevertheless, the trade-off in quality for speed is quite evident.
I’ve had it thrown back at me that, oh, we just need to wait for the next model and the quality will be much better. Or that it’s a skill issue and that if I take one hour to write up an entire markdown file of instructions, the LLM will create its own implementation plan and get everything right in one go. Sometimes that does work, but it depends on the scope of the project and of course the scalability. I feel like most people don’t know or don’t care about any of that stuff though, and I mean why would they? It’s not like they’ve ever had to build anything like that themselves before. The computer is just a magical electric box on the inside. I mean, how hard can it be to center a div?
We can produce way more than ever before now, but the cost in cognitive surrender is pretty apparent. If enough folks in the chain of command hand-wave a product or process off as “good enough,” how much longer will it be before something catastrophic happens due to shirked duties and negligence? We’ve already seen it happen multiple times over this past year with industry giants, causing global outages across systems with indeterminate resolution times. It almost feels like walking on thin ice, towards the center of an ocean as the floor drops deeper and deeper beneath us.
That said, I do think AI is pretty cool. I just hate seeing all of the misuse and waste.