The idea that you have to be the most dazzling job cantidate or the most amazing romantic partner to succeed in these areas probably holds people back a lot. In my experience if you don't do anything overtly stupid, the other party will subconsciously fill in the blanks so that you match "what they were looking for".

Regardless, it's still like a sport - if you don't make unforced errors, you will be more likely to succeed.

Consequently practice does help. You're better at dating if you go on more dates. You're better at interviewing if you get more interviews. Practice allows you to "not make silly mistakes" unconsciously, which frees up your more creative side to show the best you.

I'd say 90% of the stupid things I do are because I wanted to do something new or risky and as such I have a neophytes mind. Sure, when I start to hit an intermediate level I can then self-analyze and make that jump from average to 'sometimes good' but I can't do it all the time and I sure as hell can't do it in the beginning.

So lets look at interviews. I've been on a few. I've played the ultra-conservative role of being super-careful and treating them like I'm on trial. That approach doesn't seem to work. I've recently played the role of someone who is very socialable and even tells a joke or two and other "stupid" behaviors

Hey, you know what? It turns out most humans are far, far from these rational logical creatures. They aren't thinking "Lets fill this position" they're thinking "Lets hire someone I can work with who doesn't seem like an overserious ass or a nut and has at least the basic competency to understand the job and learn more." Geeks of the world need to understand the great importance in social skills, risk taking, socialization, understanding social culture, etc. The idea that we can just solve everything by being super careful and super critical of ourselves is not a smart strategy.

and btw, first dates are great!

Quite the contrary actually: leaning back, having fun and "being yourself" is sustainable because you don't put on an act.

I dunno, man. I graduated last month. 9 interviews in the last 2 months. Finally landed a job, so I guess I can talk about how the other 8 went. Honestly its not so simple as you describe. There is definitely a supply-demand dynamic at work. Too much supply. Literally too many sailors and too few ships. Every university of repute ( Courant, UCB, MIT, Princeton, Stanford, UChicago, CMU, Columbia, Cornell ) and the 2nd-tiers ( UMich, Rutgers, UToronto, Baruch, Boston ) together graduate some 1500+ solid quants each year. That's 1500 people with a Financial Mathematics Masters, in addition to C++ programming, whether via BSCS, MSCS, or PhD ( several PhDs in my pgm ), with typically another 1-2 graduate degrees thrown into the cocktail! I used to think with 3 Masters I was in a good spot, but I realized I was actually in the middle/bottom tranche. There are many people ( some 25% of class ) seeking a Masters in Financial Math who already have either a Physics PhDs or Statistics PhDs or ofcourse pure Math PhDs ( aka God ). So they get this science/math PhD, realize academia sucks, then head for a financial math degree, then get a C++ certificate ( there's a mini industry minting money off of "C++ certification for quant") and then show up at the interview. A first-timer ( ie. one with just Financial Math & C++ ) has zero chance. These are not the only ones you are competing with. You also have insiders ( people with MBAs and CFAs and 5 years at an IB/PE ) who may not do very well on the C++ or the math, but their experience at a trading desk gives them a huge edge. IBs surely cannot absorb 1500 bodies year after year. So the interviews are super-technical and nobody wants you to succeed. I wouldn't say I was stupid/obnoxious at 8 of the 9. 2 of them said I was "clearly overqualified and would quit in 3 months out of boredom" ! Some others were "not a right fit" type ( too much programming not enough math or vice-versa) , but there were a few hedge funds where I felt Boy this is my dream job but they were technically off-the-charts. The questions were goddamn hard. And the math was hard. Write C++ code for pricing a barrier option on an instrument with some specified parameters where the volatility is stochastic ( ie. use Heston's volatility model to price a barrier call. ) Now that's not the sort of stuff you look forward to without a stiff one. A classmate was asked to walk through a SABR and a GARCH and asked to explain in detail which one would be a better fit for a certain class of fx bonds and why. While the programs give you a good overview of everything, that is also their flaw - they give you detailed knowledge of nothing. So you have to have mandatory side projects to succeed. I had written a few thousand lines of java to price condors & written an optimizer over the holidays, so that was what got me my job. But even that wasn't sufficient. Walk me through your code. Did your model actually make you money ? How much money ? Did you make money because of your model or because of the view of the instrument ? I mean that stupid side project suddenly took a life of its own and became like a real production system which should jump through all sorts of hoops and make money from day 1! There was much more relief than happiness when it was all over and I got the job.

Bottomline, right now there are simply too many good candidates and too few positions. You have to be really really smart and lucky. Stupidity and obnoxiousness is a very trivial filter - most will get past that pretty soon. What happens after that is what counts.