I’ve been thinking of a bunch of random stuff this week and I haven’t really had time to formally write them down. I wanted to write perhaps a few blogs but since I don’t really have time, I’ll just ramble randomly here.
So here goes:
I used to live under the assumption of “let’s hope that AI doesn’t progress that fast, because if it progresses fast, then it’s all over anyway”. I remember saying almost exactly this several times.
I remember two years ago, people were laughing at AI counting letters in strawberry or understanding “what happens when you put a cow in a blender?” as a joke. I remember chenmark remarking that “[openai] was working on reasoning”. I remember thinking “ok, we’ll see in a year or two”.
5 months later, the reasoning model was released.
Now, only a bit less than two years later, AI has probably reached my capability in competitive programming (I didn’t bother to check but I think this is certainly true). It’s certainly as good as me when programming robotics stuff, in fact the only thing I bother to do is make sure things are done my way, but it can certainly everything, and I do very little actual coding (as opposed to debugging hardware issues etc).
It’s clear to me that in a few years I won’t really be needed for this type of stuff anymore. Or useful. Even right now Claude Code beats me at most stuff.
In fact I’m not at all aware of anywhere in CS where AI is worse than me. Perhaps in the type of thing where you have to track a ton of details and invariants, but it seems like AI is now, at least as good as me.
It took me on the order of 10 years to progress to this point. It took AI only 2 years. And it costs $20 per month.
What to do now? assume that “it’s all over anyway”? ???
I’ve been offered a variety of advice on what to do. For example:
- get a senior-level position asap b/c entry-level jobs will be replaced
- join a startup in the summer
- take a gap year and work
- drop out now and go work
- do a year or two and drop out
- have fun since it’s all over anyway
At least, I think perhaps the traditional way of going to college and all that is, perhaps, just wasting time? What is there that I can learn that is useful, that AI can’t learn? The only value of college now is really to make friends and etc.
But what value can I offer, at the end of the day, that AI can’t offer? What use am I to the world? or to anything? Why should I even exist? Or will I exist, in 5 years?
“perhaps in 5 years the only jobs left over will be like [this hands-on mechanical work] *chuckles*” – my robotics mentor. He might just be right…
For the first time in perhaps a forever, I have no idea what I should do. Partially because it is senior year and I don’t feel like grinding anything, partially because I don’t really see the use of grinding anything, and partially because, what is there really to do?
For right now I’m just burning time at robotics. But what to do afterwards? or over the summer? or after that?
i dont know…
Also, I’m more-or-less retired from CP at this point. My last major competition will probably be the USACO US Open this weekend. Or camp if I make it. But after this USACO cycle I’m probably never going to seriously touch CP again.
Ever since I camped in sophomore year I feel like I’ve had less motivation to do CP. There are several, perhaps equally-important reasons. For one, I didn’t feel like I was getting better. And getting better really entailed solving more problems and getting better at thinking, rather than learning new DS or algos. For that reason getting better lost a lot of its novelty. You can’t really learn new algos and expect to get better always…
And more of the algos that were left to learn didn’t, have that sort of simplicity and beauty that I like, and the finer details of the algorithms were more important / annoying than the big picture, or at least it seemed that way to me. And so I didn’t really continue learning new algorithms which was the part of CP that I really enjoyed the most.
It’s also important to note that, at this point, I thought I should try other things since what I had to learn with CP had mostly saturated to some point, and I was really at that point just training to get better at CP and didn’t really, like, become better in general. Which led me to do all sorts of not-really-side-quests like math and physics, I guess.
And I guess, many of my friends who used to CP with me had graduated, and were no longer seriously competing. And to be honest, I’ve kinda burnt myself out grinding CP for too long.
Once upon a time I dreamt of being GM and IOI. Well, I guess I will remain master (peak international master) for now on…
