2025 Wrapped: AI Edition
At the start of this year, I shared a few (non-exhaustive) ideas about how I thought AI would change things during the course of 2025. Shall we see how much I got wrong?
To be clear, I’m not a journalist, I wasn’t then and am not now trying to write an exhaustive piece on the impacts of AI. I was just sharing some personal experiences and ideas, and I thought it might be fun to look back and see how wide off the mark I was.
Developer Supply Chain
In January this year I had an open role for a junior(ish) engineer. Right now I have an open rule for a junior(ish) engineer. This does still seem to be the exception though. Graduate unemployment levels are at the highest they have been in a long time with AI often being blamed. Tick. 1 point.
Who Trains Future AI
The next item was more of a question, where will the ever-larger training data sets come from if we no longer populate the internet with knowledge in the form of Stack Overflow, Wikipedia and Reddit? Well, all of those still exist. Even so, perhaps we won’t need newer languages or tech stacks if we can just vibe code our way to everything in C instead. :laugh I still don’t have an answer to this, but I suspect it will take longer than a year for its impacts to be felt. 0 points.
Code is a Liability
And AI still produces lots of it. With little qualm about its maintenance. That’s still your job. We haven’t seen better tooling to identify duplicate code or help refactor code. In fact, LLMs have gotten better at writing new code but not at any of the other things human developers have to do with code and given that in reality we spend far more time updating, reading and maintaining existing code than we do writing new code I suspect human developers aren’t going anywhere for a while. (Yes I know we have better results on the SWE benchmark… I don’t buy it, that data is part of the model’s training set).
I don’t have any good metrics on how much code humanity is producing now in a post AI world compared to pre AI. Given the existence of tools like Lovable and Replit (even with their recent slight decline in usage) I suspect we have much more total code in existence. How much of that code becomes production grade as opposed to a demo or mock up is questionable I suspect. I’ll be generous and give myself 1 point for this.
Measuring Software
I wrote this in relation to a study that showed how AI had resulted in much greater code churn and implied this is bad. I was hoping for AI not just giving us a way to write more code quickly but that we’d see tooling for a world where code is created much more quickly and cheaply too. That hasn’t really happened. We still use the same source control systems, CI pipelines and so on. I’m still hopeful but 0 points.
March to Homogeneity & Mediocrity
I think I deserve about 100 points to for this one. No one else was talking about this and now everyone is complaining about the same sounding text we read everywhere on the internet courtesy of the myriad of AI tools available and embedded into everything where we might scribble a few thoughts. 1 point
3 of 5 isn’t too bad right?
Possibly though, the most important point I made was in the closing paragraphs. I wasn’t right yet but I might still be soon enough.
“We have however seen this big tech playbook before. VC funded startups get us all hooked using the VC’s cash to foot the bills. Like a ketamine dealer giving you the first hit for free. What are the odds that it costs Microsoft more than $10 per month per user to operate GitHub Copilot? OpenAI CEO Sam Altman has already stated that even their most expensive tier at $200/month is losing money. At some point both big tech and VCs will want to see a return on their billions in investment. As we know all too well from our now subscription defined lives, when this happens the cost of your AI service will increase.
We aren’t about to the put the genie back in the bottle so doing your ostrich impression is probably not the best idea. At the same time though the real costs (both monetary and societal) of using gen-AI will not be apparent for some time. I’m not sure I’d bet the house on it just yet either. If SaaS companies that have adopted AI already aren’t happy with their reduced profitability just wait till the VCs and big tech companies start trying to cash in on their investments.”
