Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Understanding Jane Street

thediff.co

Post reply on HN
Print view View original

201–210 of 392 posts

Re: Understanding Jane Street

#201

Post by SonOfLilit »

mgaunard wrote: post #134
Gene_Parmesan wrote: post #125

Earlier quoted context omitted.

Yes, it uses ML-style type inference, but that doesn't mean you are literally unable to annotate types, and the compiler is absolutely still doing type checking. I'm sure people like Jane Street annotate everything. In addition, perhaps a small point relative to the first, but in Ocaml, the arithmetic operators perform no type inference; there's a separate operator for float-addition versus int-addition, and so on. T…

That is not the idiomatic way to use OCaml, so I wouldn't assume they do that.

At least in similar languages like Haskell, what you do is let the compiler infer types and then press an IDE key combination to write out those inferred types as part of your code so you would get an error if you ever accidentally change the types.

Re: Understanding Jane Street

#202

Post by orbifold »

data_maan wrote: post #110
ChadNauseam wrote: post #103

Earlier quoted context omitted.

What good has cancer research done to a person living in Madagascar?

So you've established: People in Madagascar don't get cancer.

I think it is fair to say that most people discussing Madagascar here, have no conception of how it is like to live there.

Re: Understanding Jane Street

#203

Post by Ntrails »

> the winners get a job from which people routinely retire rich in their 30s

I'd love to know where this claim comes from! I'm not sure how much/what supporting evidence there is. It doesn't fit _my_ experience of people in Quant Finance. Of course, what does "routinely" mean here? 20%? More?

Re: Understanding Jane Street

#204

Post by em500 »

FabHK wrote: post #157
jaycroft wrote: post #127

Earlier quoted context omitted.

But it doesn't have to happen at the exact same time? Limit orders exist, no?

Further: Suppose you had an auction on shares once an hour, or even once a day, but no continuous trading. Would that make the world any worse off? (except for high frequency shops?)

I doubt it. I've never heard a coherent explanation how liquidity on sub-second scale is a great social good, while at the same time the largest equity markets in the world are closed 2/3rd of the day, plus all weekends and holidays.

Re: Understanding Jane Street

#205

Post by JonChesterfield »

bitcharmer wrote: post #199
fiprofessor wrote: post #188

Earlier quoted context omitted.

> I think people go to high-technology finance because they want to test themselves against a harder class of problem in a more adversarial setting against people who feel the same. Maybe this is so at other finance firms, but my experience with developers who go to Jane Street is quite different. Because Jane Street heavily advertises OCaml as part of its recruiting strategy, I know many people who ended up there ju…

Having worked in HFT for well over a decade now, I'd say OCaml is more of a deterrent than helps in getting access to larger pool of talent.

Every few months or so I run out of patience with C++ and wonder about the alternatives, of which there seem to be very few in HFT. Jane Street stands out for that reason. It's a smaller talent pool but I don't think it's a subset of those that the C++ shops can attract.

Re: Understanding Jane Street

#206

Post by SnowHill9902 »

spoonjim wrote: post #87
benreesman wrote: post #79

Earlier quoted context omitted.

Not to mention that tightening spreads, deepening books, and equalizing prices across regulatory/financial/geographical regimes is a pretty serious social good in its own right. I understand that (as the article mentions) these folks clean up when the wheels have already come off anyways, but day-in-day-out, the spread on AAPL is one tick ($0.01) nowadays, rather than the 1/8ths that you'd get quoted by some loud guy…

If you're going to hold AAPL longer than a quarter, then the tick vs. 1/8 doesn't matter, and if you're not, your trade doesn't need to happen to support the core goal of financial markets which is to finance companies.

That’s a moralistic positivistic take on a real friction problem which has been reduced. It doesn’t matter what people use the market for.

Re: Understanding Jane Street

#207

Post by fiprofessor »

bitcharmer wrote: post #199
fiprofessor wrote: post #188

Earlier quoted context omitted.

> I think people go to high-technology finance because they want to test themselves against a harder class of problem in a more adversarial setting against people who feel the same. Maybe this is so at other finance firms, but my experience with developers who go to Jane Street is quite different. Because Jane Street heavily advertises OCaml as part of its recruiting strategy, I know many people who ended up there ju…

Having worked in HFT for well over a decade now, I'd say OCaml is more of a deterrent than helps in getting access to larger pool of talent.

That makes sense to me. But, in light of that, the ones who do go to Jane Street are relatively more likely to be interested in their tech stack and OCaml, as opposed to wanting to "test themselves" in the "adversarial setting" that the post I quoted describes. In contrast, the couple of people I know who went to HRT or Jump Street are much more like that description. They deliberately targeted HFT work, whereas Jane Street has more people who "fell into it" because of this outside interest.

I mean, Yaron used to go around a lot and give guest lectures about OCaml programming "in industry" at all sorts of functional programming courses in universities. I have to imagine they thereby recruited people who would have never considered HFT shops otherwise.

Re: Understanding Jane Street

#208

Post by JonChesterfield »

red2awn wrote: post #196
nly wrote: post #129

Earlier quoted context omitted.

They'll have their cost centers in places like NY and London incorporated separately, making consistent losses, and then the arm doing actual trading will be in Bermuda or somewhere with 0% business tax.

Under this structure, how are trading profits shifted to the cost centres to pay for operating expenses?

Licensing IP would probably do it. Oh, the other direction - just make some returns in the place with operating costs.

Re: Understanding Jane Street

#209

Post by SnowHill9902 »

em500 wrote: post #204
FabHK wrote: post #157

Earlier quoted context omitted.

Further: Suppose you had an auction on shares once an hour, or even once a day, but no continuous trading. Would that make the world any worse off? (except for high frequency shops?)

I doubt it. I've never heard a coherent explanation how liquidity on sub-second scale is a great social good, while at the same time the largest equity markets in the world are closed 2/3rd of the day, plus all weekends and holidays.

Because real world information has sub-second resolution and a healthy market should reflect that. It’s a continuous auction. Some markets are open longer such as FX.

Re: Understanding Jane Street

#210

Post by bitcharmer »

JonChesterfield wrote: post #205
bitcharmer wrote: post #199

Earlier quoted context omitted.

Having worked in HFT for well over a decade now, I'd say OCaml is more of a deterrent than helps in getting access to larger pool of talent.

Every few months or so I run out of patience with C++ and wonder about the alternatives, of which there seem to be very few in HFT. Jane Street stands out for that reason. It's a smaller talent pool but I don't think it's a subset of those that the C++ shops can attract.

Rust is getting pretty popular among low latency firms. Many of my peers (myself included) would definitely show interest in a rust shop. OCaml - not so much.
Post reply on HN