Hmm there are so many disappointments that at times it is laughable. Here are a few but I could think of many more if pressed
- You take no risk. By constructing stochastic models, pricing models, risk models, etc, you are essentially constructing a toy version of how finance works. But you never trade on that information yourself and don't really profit from it. The way around this is to become a quant trader or just a trader.
- You have to deal with idiots, real stupid people. Sales, consultants' business analysts and project managers deserve a special place in hell for how they deal with quants. Traders are usually better. Often these people bug you and annoy you.
- It is impossible to create good models in bureaucratic environments. I remember looking at the models used to price CDOs and credit default swaps and knew they were garbage. So what. They were still used and banks lost a fortunate. A lot of quants felt useless, the ones who didn't got fired.
- You have to run a lot of boring computational business tasks to keep the pricing functions up to date and you have to always deal with shitty data.
- There is very little scope for new ideas. Quants have mutant strains and often choose a model by looking at what the competitor is using.
- Academic quants deserve a special place in hell - they know absolutely nothing. Same with regulators, consultants and the virus of all viruses, senior management. These people need quants but understand nothing. I mean absolutely nothing. I had meetings with senior guys explaining why CDOs were so dangerous - I ended up explaining why CDOs were so dangerous with the following analogy: imagine a house of cards, some cards have high ranks others have low ranks. If just one or a couple cards fall, they all fall, it doesn't matter what ranking have have. Senior risk managers were happy with that explanation and they felt it was rigorous. I felt humiliated working with these people.
Long story cut short: lots of idiots, boring tasks, computationally intensive and lots of bureaucracy. But you get to play with geometric Brownian motion all day - so that makes up for it.