Which field would you enter if you were to start your career now?

[The Doc]

So, the private credit/debt industry?

[sigterm]

@The Doc



1. This is the perspective of an IT worker, mind you, not a quant. Data Science is a buzzword. It can probably be defined honestly as traditional statistics plus the new computational stuff popularised by e.g. The Elements of Statistical Learning. As you wrote, lots of people advertise "data science skills" but few truly are good at computer science and mathematics at the same time. It brings value yet can still be done "in a garage", with little capital, compared to most engineering, which ties you to moneyed people you'd rather not have to report to. So choosing Data Science would be optimising my career for autonomy, not money. I don't like working, and I especially don't like taking orders; other career goals will direct you in other ways.



2. I meant, look into big cloud providers' papers/talks/job offers, to see what kind of work is behind XaaS and see if that tempts you. Not everything cloud makes sense, but cloud services make economies of scale, and that's going to stay true indefinitely.



Everything is slowing down. Technology that actually brings value to the world will need smart people to prop it up for a long time, don't worry. If you want a very high-level perspective, think about what's essential: finding and distributing energy, food & water, controlling people, etc.



BTW, rftx713's advice is genuine. Do not be dazzled by the lights, they're there to enslave weak people with no internal compass. If you're really that good at making beats, then give it a shot. Just try to avoid Avicii's fate.

[jslade]

There is still money to be made in tech. I can't stand the culture though. Zippity doodah new framework angularjs, twee nincompoops making simple shit complicated because muh engineering, "product managers" who don't know anything, patches on patches masquerading as engineering improvements, insane political witch hunts, greasey VC antics, 80 hour a week death marches, "work for me for free" hackathons, trans-people on roller skates next to the ping pong table and catered lunches, plus virtually all of it visibly makes the world worse. I know people who had fun in early Google days and stuff like that, but I don't know anyone in the Valley having much fun now.



Always thought observational astronomy was pretty cool. Surprisingly everyone I know who went into it continues to do it, so it must be OK; better than physics anyhow.



I've thought about going into oil and gas. Tons of interesting mathy problems to work on, you can don the safari hat to visit remote locations, good compensation with potential for wall street sized upsides, I won't be persecuted for being who I am, and evil idealistic ding dongs are off trying to make solar panels that work at night instead of doing honest work drilling giant holes in the earth and drinking dat mofo's milkshake.



Criminally underrated: operations research. Practitioners are usually vastly more hard headed and well trained in fundamentals than "data scientists." It's also the type of engineering job where you can go home after an 8 hour day.

[Jurassic]

@jslade "but I don't know anyone in the Valley having much fun now." because of all the previous?



"good compensation with potential for wall street sized upsides," whats the job though

[Maggette]

Agree with jslade on the OR stuff.



IMHO there are a lot of fascinating problems in manufacturing, supply chain management and logistics. A combination of forecasting, risk management, planing and optimization or even game theory...sometimes coupled with challenging computational restrictions (lattency, availability and even limited cpu or memory).



"Data Science" and more important machine learning are very very relevant and powerfull tools here and OR engineer should know about. But so are simulation, classical statistical and mathematical modelling and numerics.



I do this stuff for a living. I love finance, but I am more interested in it, when it is part of a more "physical" systems, like in commodities.

[EspressoLover]

FaceApp is a pretty good reminder that it's still quite possible to make it huge in tech without touching the Silicon Valley nexus. All of those downsides are mostly avoidable by being lean enough to do without VC lucre, and the loss of autonomy that comes with it.



It's just tough to forego because the funding flows so freely. But most startups with $3 million seed rounds could pretty much do the same thing by rotating through $20k credit card limits. That probably means paying early early stage employees with actually decent stock options to keep salaries low, relocating outside the insanely priced tech center metros, writing C instead of javascript, and not paying through the nose for hyped up cloud-based serverless crap.



The way it looks, the FANG/VC/SV ecosystem is today's version of IBM circa 1975. Nobody could imagine a future of computing that wasn't dominated by Big Blue. Until it happened, then it seemed obvious in hindsight. Today's tech giants have built Robber Baron like monopoly juggernauts. But at it's core it's all built on top of the most decentralized platform in the history of the world. That's a house of sand if I've heard of one. There's no way that business model survives in the long-run. The walled gardens are a lot more pregnable than they look.



Right now, this generation's Apple Computer is in a garage somewhere. That garage almost certainly isn't in Palo Alto, and Andreesen Horowitz definitely won't be in at the ground floor. If you're young, insanely ambitious, and want to light the world on fire, imagine some version of 2030 where Google, Facebook or Amazon have been rendered irrelevant. Now figure out, how do we get from here to there?

[The Doc]

Some points from all of you:



1. You all mentioned commodities.

Yeah. Nobody gives a f anymore. They take those things for granted. They are busy with their insta stories. They all think Zuck is spying them too. lmao. You can look at the prices though. Metals are stable (dropping in very slow rates). Everyone is waiting for China to boom again lmao. Even battery metals are not climbing. Ags can be interesting but only for specialised crops where the margins are huge but volumes tiny. Energy is oversupplied. Oil consumption is peaking in like 10 years (my guess, i could be wrong but you get my point). As for gas, we produce so much that in SOME cases its price is negative (imagine being a commodity with negative price lol). People are paying others to take it in USA (cause of fracking, now thats a fiasco too, but w/e). Trump is begging EU to take his LNG. I remember when i was 10yo we had power and gas crises. Now its all stable. It takes the excitment out of the game. If you are talking about drilling or production engineering, then yeah, that could be interesting but it might get boring after a while. Still money to be made in commodity trading but i thing the era of commodities is over. Change my mind



2. @EspressoLover: FaceApp? That shit is already dead. They probably made big bucks during the week that thing was relevant. But thats not an argument.

"But at it's core it's all built on top of the most decentralized platform in the history of the world." Having decentralised platforms got them where they are. Its a much better experience for the user imo. Maybe you are old school, or im just stupid...

"how do we get from here to there?" Hmm, we dont?

People keep saying "markets are cyclical". Stop it. Use this instead "markets WERE cyclical". I got your point with IBM in 70s but its not relevant anymore. Its like using historical data to predict prices. Useless.

Very interesting poionts about VC. Can you expain this more?



3. @Maggette: Isnt the current state of ML (or at least 80% of it), image recognition, and NLP? The only big think i see coming out from this, is the self driving car. Oh, and very good smartphone camera software/firmware. As for operations and manufacturing, those processes are being optimised for decades. Nothing new or groundbreaking here.



4. @jslade: Agree with the culture in tech. I cringe even with myself sometimes. Space is cool but since its nothing out there i would say its boring. Oh, and you are entirely government funded (i wouldnt like that). Can you explain ops research? Sounds interesting, but my mind goes to boring stuff. Are you talking about manufacturing or tech?



5. @sigterm: So cloud is the future or the present?



- What do you guys think about a career in sales? Maybe sales engineering/tech sales? Maybe people are tired of the typical salesman and want more customised and faster solutions. Money can be good as far as i know but quite boring.



- There are some interesting uses of ML and DS in the finance industry that im interested in. Im talking about private markets though (public markets are kinda dead - dying).

There are some uses in the gaming industry too but its not paying very well xd.

Any opinions about those?



You can make good money by writing code that replaces humans or shittier code. - Me



Thank you all for sharing info and spending your time to respond!!! Sry for bad english..

[Rashomon]

image or video. stuff that can run on an ARM. (pete warden has written about this)



best CS classes are compilers (stanford--aiken), operating system (mit xv6). stanford fourier and mit linear algebra worthwhile.





there are loads of weird data formats to specialize in. just like FIX in trading, netflix needs people who know the details of yuv/hmac/mp4 (i am out of my depth here but you get the point).





get good at C, but not the first thing. start with python/node/ruby/friendly. don't waste time on scientific computing (unless it's macaulay2, which is a totally different story).















A friend who ran a jazz combo said he would always try to get a drummer who used to play with a metal band. Why? “They hear ALL the 64th notes. So playing a weird rhythm just meant leaving out certain 64th notes.”



Something similar is true with coding, the best people for clock issues are video game programmers. They really really really wanted that missile hitting their sprite to line up with their boom sound----more than a generic young ambitious wants to learn 2 play stock market good.











Put another way: skill is unmistakable in other fields.






---



added. Koopdeta, I would also invest 1-2-… months each in learning puppet (there is a learning VM) and docker (their standard docs). Old sysadmin skills are, for me, still like a bread-and-butter, but zillion other things may be replacing them. That said, if you buy the GNU Parallel book you'll see the argument made that PostgreSQL and GNU parallel (pi.dk/3) and cron can do most distributed systems work. Whatever. I'm picking 5 technologies for you to start with so you don't have to pick. Move on from there in another year or 2 when you know more. My 5 are: parallel, redis, docker, puppet, postgres.for salary data at FAANG, look up "team blind". As a bonus, you get a preview of how nasty some people are on the inside! Just look at their "marriage" section.

[deeds]

Super interested in macaulay 2 reference. would you elaborate? groebner applications?

[Sisqo]

I am too young to have first hand knowledge of this but it seems to me that old fashioned 20th century Wall Street partnerships survived for so long, when they did survive, because the structure reduced agency risk, and forced everyone in the boat to mostly pull in the same direction, or else be ruined.



What led to the general extinction of this model for joint risk-taking? Is it gone for good?



Mutual corporations at one time made up a large segment of the banking and insurance industries, and then demutualization wiped them out. There was some wealth unlocked temporarily for the engineers of the restructuring, but I am not sure that the corporate descendants of the old firms have been more successful than their forebears.



Is the mutual company gone forever?



Back office operations at one time were such a crushing problem, that it was a limit to growth and operational risk was significant, and I think this even significantly contributed to systemic counterparts risk.



Data used to be more scarce, yet less valuable.



It is a truism, or at least it was when I was still current on these matters, that there will always be a march of small denovo banks, and the vast majority of them will not go out of business, but rather will be acquired. What are the stylized facts about the stream of de novo buy-side shops that have been popping up since I was old enough to be paying attention?



Just ten years ago Jim Rogers was still being invited to tell everyone about the infinite boom in commodities that was going to inevitably follow the rise of China and the emerging economies.



As another commented here mentioned, the monopolies of Google and Facebook are largely taken for granted now. I personally am prone to perhaps unthinkingly accept that automation and technology will lead to ever greater consolidation.