uhh guys can we talk about ai…

This is a long one but I think about ai non stop and I need a space to share my thoughts. I rarely talk about ai seriously with coworkers because I don't want to be an autistic vibe killer, so you guys are going to have to hear it instead.

I've been thinking about how ai is going to change or replace my job for the past year now, partially because there's no definitive answer yet and partially because my career matters a lot to me. For a while it made me depressed. At some point that shifted and now I mostly think about it in terms of what skills I should be building right now and what a given role or industry is actually going to look like in 5 years.

Here's what I've actually seen. I've been using ai extensively for deal work all year through a tool we have called rogo (not saying rogo is anything special its just what we have and it gets me the most recent models with a big token budget). As recently as January/February I was double checking every little number that came out of it and finding errors pretty regularly, nothing crazy but enough that I couldn't send anything without going through it line by line. Then sometime in the last month or two that just stopped. I give it a task with some initial direction and maybe a follow up and what comes out is largely copy and paste usable. The nuanced stuff still needs hand holding but the boiler plate is just done, and done better than I would do it. I still spot check and on anything standard I basically don't find errors anymore. That six month window has dramatically changed how I do my job.

Based on my experience and reading in the topic I'm convinced execution work is out, and it's going out fast. I'd estimate 1-2 years before the frontier models are better than any human at the modeling, the slides, the materials, the analysis. I mean better than the best analyst you've ever worked with. Then maybe another year before the industry actually gets there. Maybe models plateau, but I've been hearing that for four years…

And I don't think it's just an analyst problem. If your job is to run the process on a deal, keep the model, turn the drafts, chase down diligence items and all that, the role is going to change whether you're a first year or a VP. The roles that stay similar (for now) are the ones where you own the relationship or your name is on the materials, because someone still has to be on the hook and it isn't going to be the model.

Part of why I think people on WSO aren't taking this seriously is I saw a poll where 80% of americans think ai will eliminate jobs but only 6% of workers think it will replace theirs. That's the exact dynamic on this forum, everyone agrees it's coming and everyone thinks it's someone else’s seat.

So if you read any of that, what are you actually doing about it? For me the answer so far has been to move from an execution role where I spent 14 hours a day head down working at my desk to a downmarket role where I can start building relationships with clients now. I’ve begun focusing less on technical skills and more on being a well rounded person that’s easy to talk to. Not sure that's the right or only move, which is why I'm asking. What skills are you prioritizing now and what do you think an associate or VP seat is actually worth in 5 years?