Dimon to hire more "AI Experts" and less "Traditional Bankers"
Is this the AI purge coming to fruition? How do you guys think this sentiment will be reflected in upcoming SA recruiting or even incoming SA classes return rates?
They will be lower
sensational contribution
Congrats on NYU
Stupid answer to a stupid, unanswerable question
I mean I’d guess the better the bank ur at the more you’ll see an impact in the form of lower return rates since they’d likely more aggressively implement AI solutions. For boutiques I’d think stuff won’t change as much immediately as in 1-2 yrs
it just means the interns who can't leverage AI will look crap in comparison the one's who can and therefore don't get returns
As someone who's worked with CS and Stats PhDs who specialize in Ai, none of them want to work for JPMorgan. If you're smart enough to be really good, you go work in silicon valley if you have passion, and if you want to make money you work for trading firms/hedge funds.
This is something I'm rather shocked by, is that Dimon seems to want to make JPMorgan an Ai focused firm, but still pays and treats ai staff like cost centers instead of revenue generators. Until that changes, don't actually expect to see massive changes. Watch what they say, not what they do. Dimon still pays bankers like revenue generation and ai staff like cost centers. That means they're just saying stuff to get investors on board, not because they're planning some major internal revolution. If they were, Dimon's not dumb enough to think he can pull top talent paying below FAANG company rates.
Basically he means hire a few people that develop ai agents that have organizational context
I'm friends with a few devs internally where we've talked about this and it's really not that complex
You also really don't need that many of these people either
I'm guessing you're either talking about using some sort of open source model, maybe Gemma or Llama, and retooling it for your own needs. That's what a lot of firms are doing. It's that or an API to Anthropic or something.
The problem is that this isn't terribly complex in a window. But you still need a lot of smart people to deal with the edge cases. Scaling is a big one. What happens when most people internally start using this Ai? How do you ensure it scales to internal usage? JPMC has 300k employees, at least 500 petabytes of data. That's across many organizational levels. How do you ensure proper scaling, proper data intake and data engineering, proper security? Do you utilize mix of agents? How do you ensure that you don't have token costs that explode? How do you utilize thinking models properly, so that you aren't constantly billing to high or xHigh Opus and costing yourself billions per quarter?
If it wasn't so complex, you wouldn't need thousands of people to do these things. Tbh, most devs at banks aren't very good. I say this having worked in banks, prop shops, and hedge funds. The devs at banks were usually bottom of the barrel. I cannot tell you how bad the technical debt at my bank was. This seems more like a Dunning Kruger affect of some people who maybe had 1 ML class in college and never spent any time on MLOps or data pipelines like ETL. The infrastructure to serve ML is oftentimes much harder, especially at a place like JPMC where the legacy code is older than most bankers.
Iusto in culpa inventore consequatur est. Natus odit aliquam omnis excepturi quaerat. Magni odit voluptas provident libero labore ex. Ea asperiores sit delectus quis voluptas aut delectus id.
Quis facere consequuntur quasi ex eius. Consequuntur omnis eligendi et et aut. Ut nobis explicabo aut sed laborum praesentium. Perferendis nobis ex tempora unde.
Maxime voluptas sit qui. Dolorum ipsam voluptatem libero consequatur incidunt. Recusandae quae non vel blanditiis.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...