My sister is doing a PhD in machine learning, and she keeps telling me that I will regret not studying CS and aligning my career goals without considering the advancements in AI. Should I be concerned?

Why not both? Get a major in finance and take CS classes. If coding is the buzzword just acquire that skill on top of the finance background.

What's the alternative? Study CS and end up at a FAANG coding gmail buttons for a living. No thanks

I hope you sister knows that unless she's in the top 1% of her field she's going to be replaced much faster by automated coding AIs than those of us in relationship/deal-driven finance. She's just an academic. She has a maybe slightly better understanding than an undergrad student, but still woefully lacking perception of how the real world operates. Her entire existence is a bubble and the sum total of her experiences is like bowling the the guard rails up the entire time. The most probable thing that will happen at the analyst/associate-level is the seats will become even more competitive as various tools lower the technical and knowledge barriers, meaning even more people can compete for a shot. In other words, the sooner you can get yourself into our profession in the first place, the better IMO.

All the systems and things that will be automated across finance are only support material for taking decisions. Do you think that pitchbooks, models, and whatever else dictates the final decision? All of that is meant to guide the discussions/plan/negotiate certain aspects of different deals. To put in CorpFin 101: Finance is interested in how markets perform, how to fund and where to allocate resources, etc. So the things that makes finance be finance are the people that work in it and the decisions they take. Do you think that GS/MS/JPM/etc? etc. are there because of their tangible assets? Their entire balance sheet is built by decisions taken by previous seniors, by the "value" they provide in terms of expertise, etc. We live in an era where you can basically learn and do whatever the fuck you want because of the internet. Do you want to learn how to value some companies? No problem, open some books/online courses and learn about that. That's available since the late 2000s btw. Yet, I didn't see clients distancing themselves from requesting banking services. Financial services/investments will continue to be requested because an important thing that isn't learned in school is that you can feel you know a lot but you lack the confidence to do it because you never did it. That's when you call a finance professional.

Indeed, a part of the truth is that the useless work will be cut = useless analysts thrown off; so it depends solely on you what type of finance guy you want to be: a monkey or someone that really can bring some value to a company.

so I concur with the 1st comment, tell your sister to shut the fuck up because she lives in some type of utopia in her head (among others, expect her to be extremely biased because of her PhD i.e. all the world is guided by machine learning).

Fucking nonsense.

Also, to complete a little more, those systems are only as good as their inputs. There tend to appear some trends in the market, strategies that are more suited for certain economic environments, etc. etc. Who will update those systems for day-to-day trends? I mean, some economic trends are profited and created strategies on how to profit from them in boardrooms. So good luck to the tech guys updating the systems with up-to-date info.

If not, regardless of how much info you put into a system, the reality is that finance is common sense. You don't care as much about reading a paper with 20 pages with all the possible technical terms of finance book explaining why it's a good investment or a bad one. Instead, your thought process may be something like: Is this product good for this market? What is the expected demand? What makes me think that? And you may take your decision solely because you had a walk around in Manhattan and you saw that there is some type of trend that will lead inevitably to more distant product that at the moment people don't want it.

With corporations the same. You may build financial models and whatever the fuck you want. In the end, to finish the transaction, it's not even a finance question, but instead a conflict of interest in the corporate governance side and you're trying to sell the exting and thinking maybe how to plan an appropiate roll-over for a director to not impact a company's value. The system may offer you some guidance, sure, but the system can't get the character of that director and it can't negotiate with him to agree on some common ground.  

so you're sister sits in the bench of the IT/tech guys that think that human interactions/psychology/social behavior aren't important if systems are efficient and good, everything else is irrelevant but unfortunately, everything is guided primarily by human aspects and only secondary by technology (this bench your sisters sits in is the same where the guys that think that "making a good job" is enough to advance in a career sit, meanwhile politics and managing/reputation/finessing/soft skills are the main drivers)

Do not take career advice from a PhD student. Let alone one that likely knows nothing about your industry. AI has not penetrated services involving relationships, advisory, and sales, and likely never will. The other commenter talking about apprenticeship is correct. Recieved too much braindead advice in my life to let this slide.

This is funny because AI has the same probability of replacing regular coders (not specialized AI/ML coders) as finance folks. I mean, you know AI can now write code?

If you want to do IB or PE for the long term, a CS degree is definitely not required. However, you can bring something unique to the table if you understand math and how to build software. For example, M&A Research Institute based out of Tokyo resembles a fully automated investment bank.

I myself studied CS and have several friends studying ML for their Ph.D. The machine learning hype train is a bit over-emphasized. All AI really is is stats and optimization.

Also, CS may not be for you. There are many other technical majors like Data Science, Stats, Mechanical/Aero/Structural Engineering, etc. Yes, they are significantly harder than other majors, but they are worth it in the end. Don't be afraid to switch majors. By the middle/end of your sophomore year, you should know what you're going to be graduating in. 

Either way, you're going to have to grind for IB interviews.

I work in a quant fund with manny ML phds, and I partially agree with your sister. Machine learning is practically re-inventing human brains ( neurons, perceptrons, and neural networks ) using machines. OpenAI has been hiring brain scientists and they're doing things at full speed. This means AI intelligence will only get closer to human intelligence from this point onwards.

When you see one cockroach, there are already ten thousand cockroaches in the dark corners, and ChatGPT is that cockroach. Shortly after openAI released ChatGPT to public, Google and Meta also released their own chatbot. They can't possibly make chatbot out of thin air in that short amount of time, so this means that Google and Meta had the chatbot technology long before ChatGPT became famous, and they were HIDING it. We don't know how many AIs are still hidden in the dark and will suddenly come out and bite us in the future. Your sister is very right that ChatGPT is a tick of iceberg.

That being said, I don't recommend studying CS. CS is easy to enter but extremely difficult to excel, and 80/20 percent rules will be more apparent in future years, where 20% of elites ( like your sister ) are rainmakers and bottom 80% can't find jobs. My advice is to get into a career that rely heavily on intuitions & instincts that are unique to human. Sales, product management, film-making, fashion design, literature, marketing, strategy consulting, politics etc. Ibanking is too process-driven to be safe from AI.

I agree and disagree, but I find it too frustrating arguing with a bunch of non-quantitatve fundamental types. I feel like you guys don't understand how the ML side works and are quick to shit on its potential to solve the same problems you're looking at. I may not have worked in a fundamental job, but I did study both finance and engineering along with a very quantitative grad degree, so I dont think my insight should be discounted like you guys say about OPs sister here.

I have no doubt in my mind the implications of machine learning will be profound - as was the internet, the airplane, the car, the railroad, electricity, etc.

But I would argue that in an AI transformation, investment banking is just about the best job to get into. I can see how AI can substantially improve our processes and make the grunt work more efficient, which will make us more productive. But AI can never replace the core functions that create real value. For example, on a sellside M&A deal, a computer cannot create the conditions that force a bidder to negotiate with themselves and increase price, even when they are the only game in town. I have the skill and experience to do that, and that’s what clients pay me for, not the pretty CIM my team creates (not that it’s useless, it’s just undifferentiated). Similarly, on capital markets deals, pricing and distribution are a matter of psychology, market read, relationships and leverage with the buyside that are impossible to recreate without having done it for a long time. This is the ultimate high touch, human job. 

People forget 50% of the analysts job is being the fall guy.

For liability purposes, having a 23 year old to fire when a mistake is made is a go to for an MD.

Goldman had 600 cash equities traders and now they have 2. Finance has a special kind of hubris that "oh yea, machines are eating the world... but it won't happen to us, we're special..."

AI will eliminate half of all IB jobs in the next decade.

I graduated with a finance degree and currently work at a BB. I do understand where your sister is coming from. You have to be living under a rock to see that AI/machine learning isn't taking the lead in regards to the US economy. Deal flow right now at investment banks is garbage and breaking in is as hard as it's ever been. That said you have to take everything someone says with a grain of salt. Lots of companies right now don't want to IPO or continue on existing deals. Lots of people also don't want to fund these startup tech corps either. A lot of people pointed out that you should not be taking advice from someone that doesn't know the industry, I partially agree with this. Although, she is in the tech space but her options would probably look a lot different from yours depending on a lot of things. At the end of the day they are both good fields. Do what you like. I personally realized that tech is on an upward trend and I'm currently learning several different programming languages on my own. I think finance and tech are a really powerful combination. You can learn code on your own, finance is a little tricker. Just my two cents. 

People keep saying all these things about Wall Street jobs being replaced, but literally all its going to do is refine the work process. It is so ironic to me that people keep saying to study computer science when in reality, the skillset you get from studying that is much more replaceable by AI. AI is going to refine the work process and cut a lot of the senseless hours analysts spend on trivial things. After that is done, MD's will just find some other random task that is maybe a little less pointless for analysts to work crazy hours on. 

People always use trading as an example of innovation taking over finance (which is a very good valid example) but you also have to look at it from the perspective of cultivating talent for later. A lot of analysts in IB/ PE are not a necessity. They are not needed for the firm to be successful and their work could be done in a much more efficient way. However, these firms are training them so that the few that stay will take over the upper level positions that couldnt be replaced. 

The role of a salesperson will never die....