pdovy

Ah apologies, I only took a cursory look at his dispatching logic. It looks to me like it supports both modes, but you have to hand roll the dispatching logic if you want to use it at runtime (but he provides an example). If you _really_ need the runtime dispatch then yeah I'd agree FMV is probably cleaner.

FMV is pretty neat but there are scenarios where it's not ideal, so I'm not surprised to see it not used in what is meant to be a lightweight high performance library.

Notably the fact that the dispatching is done at runtime means you are trading off the convenience factor for code size and running extraneous dispatching code in your critical path. Additionally I've anecdotally seen on modern Intel hardware the power heuristics can penalize you for even _speculatively_ running some of the wider instruction sets.

Yeah, it seems like this would be better served to be split into two (or more) courses if you're assuming almost no prior knowledge. "Software Engineering" is a pretty ill-defined term, but I'd generally think of it as the _process_ of developing good software, and that sort of presupposes you can develop _any_ software. Coming in with some ad hoc ability to program (and probably some bad habits!) would be a plus - then the "process" part is actually solving a problem, rather than just feeling onerous.

Regarding the direct listing, I wouldn't view that as a negative - sounds like they had no need to raise additional capital, but wanted to provide liquidity to their investors and employees.

They've raised a _lot_ privately, and some of those rounds IIRC were described as just "opportunistic" by the founder - i.e., raise money while it's cheap. Also given that they have a VC arm it sounds like they have more capital than they can actually deploy in their core business.

Presumably they would make gains in efficiency, but only to a point - if a gallon of gasoline contains 34 kWh of energy you'll always need at least that much to create one, and at 60 kWh/gal now they aren't that far off.

Thanks - that makes a lot of sense. I suppose the same issue exists with a large scale up of electric vehicles, so your interests would dovetail with other parties there and hopefully collectively spur grown in renewable generation capacity.

You've elsewhere quoted 60 kWh per gallon of renewable gasoline. The US consumed 391.40 million gallons per day on average in 2018 [1], and concurrently had generation capacity for 1.153 billion kWh per day from non-hydroelectric renewable sources [2]. That suggests that full replacement of current gasoline usage would require ~23.48 billion kWh per day, which far outpaces what is currently available.

Of course any replacement of fossil fuels is a great goal, but just curious where you see this going at scale, if it were wildly successful? It seems like if you or others were able to scale this up massively, you'd run into unit economics issues as demand for the required renewable electricity outpaced supply, raising the price.

[1] https://www.eia.gov/tools/faqs/faq.php?id=23&t=10 [2] https://www.eia.gov/outlooks/steo/report/electricity.php

Yeah, if they can't pull that rabbit out of the hat, it looks a bit bleak. With US MAU's stalled for the last four quarters, they are faced with extracting more revenue from the same set of users - i.e., show everybody more ads and presumably water down the user experience in the process.

If I'm reading their MAU charts correctly, they have way more MAU outside of the US than inside (82m vs 184m last quarter), not only that but international user growth is relatively strong compared to the US. That said, 94% of revenue last quarter was from the US, so they are not monetizing internationally really at all. If they can improve international monetization then there is a lot of room for revenue growth it would seem.

Will be interested to see what the standard range pricing ends up being, they would really be on to something if they could make a version of this that is closer to Model 3 pricing. It seemed weird to me that when other car makers are abandoning the sedan style entirely that Tesla would pick it for it's stab at a more affordable model. As a city dweller that still needs to tote around kids / pets / supplies a hatchback / crossover is a great combo of utility and not being ungodly large.

There's a great, if a bit cheesy video from FortNine [1] on why motorcyclists can be basically invisible to even cautious drivers due to the way our brain processes information, and I think the same thing would apply for bicyclists. When I'm on my bike (of either type) I assume nobody sees me and act accordingly.

For cyclists and pedestrians to really be safe we need to design our road infrastructure to counteract those issues from the start. I visited Copenhagen a few years ago and biked everywhere, and it was a real eye opener how safe and easy it was compared with my home in Chicago.

[1] https://www.youtube.com/watch?v=x94PGgYKHQ0

Seems to me like they are asking the wrong questions here. In my experience people work hardest when they are coming in every day really excited to work on a problem. If nobody is feeling that way towards the product then that in and of itself should be a big red flag. I'd be skeptical you can pull enthusiasm out of thin air with forced culture changes.

I think this to some extent depends on what kind of software you are interested in working on. From the perspective of someone working as a developer in finance, SF is nowheresville, you want to be in NYC or Chicago, typically. That's my personal anecdata but I'd imagine other smaller markets have their own niche industries.

If what excites you is web-focused startups, then yeah, you're probably not going to be happy in a lot of other places, but that is just one area of software development and I'd argue it doesn't have any exclusive license on interesting problems or top talent.

Running a race is a good analogy here. Privilege is the equivalent of starting the 100 meter dash 20 meters past the starting line, while someone else starts at zero. You can both work equally hard and achieve vastly different results. If you make the advantage large enough, no "extra effort" can overcome that advantage.

Hard work and personal achievement is not all that matters. I believe I worked plenty hard to get where I am today but I had the fortune to be born to stable parents who could always put food on the table, took an interest in my education and could afford to send me to college without massive loans. Take away any of those things and my same effort would likely not have yielded the same results - I was set up for success.

I don't feel guilty about that, but I do try to recognize that people who are struggling can't just "work harder" to fix their problems.

I'd agree with other commenters here that it seems pretty expensive for what you're getting, or at least the value proposition is poorly presented. For example, I don't see any reason why I'd want to pay for this at a US airport (at least from the perspective of a US citizen). On the other hand if you told me the immigration line at XYZ airport was typically 45-60 minutes long and you could guarantee me priority access, there is a value there that I might pay for.

The pricing also seem really opaque - the prices even within a single country vary (from $250 at JFK in the US to $375 at SFO, why?). Certainly the going rate for a fixer at the location must play into this, but from a customer perspective I'd expect the fee to mostly depend on the value to me - I'll pay more if the expected level of delay or hassle is high. That might actually be inverse of the rate you pay the fixer - there's less value in US/Europe for me than if I'm traveling in e.g., Southeast Asia.

Overall though I think this is an interesting idea and I'd be a customer at the right price point, at least for certain destinations. Good luck!

>> i dont disagree with arbitrage and liquidity, and a globally invariant price for goods. i just think HFT isnt really improving that in a meaningful way. i ask you- at what timescale does it stop mattering? thats really what im saying. not that liquidity and invariant prices are suspect, but that things are happening on such a minute time scale that it doesn't matter. hey, i could be wrong, but no one has argued why going from milliseconds to picoseconds is improving our lives. thats the thing i would like explained.

This is a little more complicated than it would initially appear, I'll try to explain from a market makers perspective why the speed race exists and why being faster (as a liquidity provider) is better for the market, at least the way it's currently structured.

Market making 101 is basically that you want to come up with a fair value for the product you're trading, and then put out orders to buy for a little less than FV, and sell for a little more than FV. If you buy and sell at those prices you're providing liquidity to the market and capturing a small spread for your effort, great. Do that repeatedly and you have a business. But how much should your "a little less" and "a little more" than FV actually be? The smaller the better for the market, and this ideally should be the primary vector on which market makers compete with one another.

Okay, so in our optimal scenario you'd always quote as tight as possible (limited by the granularity of pricing on the exchange) around the fair value. (I'm glossing over a lot here, calculating the fair value is non-trivial and your level of uncertainty about it will also determine the spread you can quote, but ignoring that for the moment). The problem with this is that if the fair value moves then some of your orders become mispriced, as they represent an opportunity to buy below or sell above fair value and if they execute will be a loser. Smart participants will recognize this and race to pick those off before you can reprice them. If this happens too often, you're not making money anymore, crap. Really only two options here, #1 is to widen your quote so that you are less sensitive to such movements and you are capturing a fatter average spread which compensates you for the losing trades. #2 is to get faster than those other guys.

#2 yields a better outcome for the market, but necessitates a speed race as an additional vector of competition. Exchanges recognize this as well and have long played around with various schemes to give liquidity providers a systematic advantage, e.g., via rebates, or otherwise. It's a tough problem.

I think if you're considering AirBnB and a hotel room as essentially fungible products, you're missing some of the value proposition of AirBnB. For me personally, I really hate most hotels because I don't like being lumped together with a bunch of other travelers in cookie-cutter rooms. A hotel is convenient and reliable, and sometimes that's most important (e.g., if I'm on a business trip) - but if I'm actually on vacation AirBnB provides a more private and less generic space to unwind and relax.

I'll be very interested to see what happens if Uber does decide to go public.

Prior to recent events I'd have thought if they did go public they'd almost certainly follow the recent trend of setting up a class structure that gives little to no voting control to most common shareholders. Will they be able to get away with that if it means Travis retains control?

Here's an analogous situation - let's suppose you are a buyer in an auction for a car. You're competing with all other interested buyers to get the car, and after a bunch of back and forth, you win the auction.

Later, you find out that one of the most aggressive bidders in the auction was actually just a buddy of the seller, trying to increase the price in his/her favor but avoid at all costs actually winning the auction.

You'd probably rightfully think this was unfair, and this is exactly what spoofers are doing in an electronic market. They are generating the illusion of interest to buy or sell, without the intention to actually do so, in order to move the market in their favor.

I work in HFT, though we don't deal with Bitcoin. I don't think volatility creates any value, but it can be a sign of market inefficiency if it's caused by a lack of liquidity (demand > supply). A typical market making strategy would be to provide liquidity while betting on mean reversion, which will have the effect of dampening volatility. The price won't necessarily trend in any particular direction, it should just dislocate less from some "true" theoretical value.

Yeah as a transplant Chicagoan I'd have to disagree, on balance our infrastructure is pretty good. CTA works well enough for me to commute daily for the last 8 years without owning a car, and they've been systematically replacing water and gas lines around my neighborhood for the past few years. Not to mention new green / public space investments like the 606, Maggie Daley Park, etc.

Road resurfacing is pretty good where I am, but oddly is largely a function of how effective the alderman for your ward is, since a lot of the cost typically comes out of their budget.

The real problems with the city IMHO are massive unfunded pension liabilities and a huge segregation problem between the north/south sides.

Thanks, definitely going to give those a read. Best of luck!

Congrats, that sounds awesome! What are the economics of that business like? I love woodworking (as a hobbyist) but it seems like most people that manage to make a living at it are doing so through volume (e.g., cabinetmakers) vs. something like custom fine furniture commission work.

> Trading companies, including market makers such as UBS , KCG Holdings and hedge fund Citadel LLC, are willing to compete for retail investors' orders because they are considered "dumb money" that shows the professionals where markets are headed.

This is contradictory, if retail is "dumb money" it's not indicative of the future direction of the market.

The basic deal with internalizers is that they just have to fill you subject to Reg NMS rules, i.e., at or inside the NBBO. The catch is that the NBBO is set based on the lit markets, where as a market maker you have to take flow from any counterparty. As a result the spread that gets offered reflects that built-in risk that some flow you get will be adverse. The internalizer on the other hand has the great deal that they can filter down their flow to retail only, which is on average much less adverse. This gives you some options - you can offer a tighter spread to entice business based on better execution quality, or you can use the extra headroom to pay brokers to drive volume to your business, or both.

Argument for internalizers: knowing the risk profile of their customers should let them improve execution quality (offer tighter spreads) over the lit market.

Argument against internalizers: transparency is king, if retail flow was driven to lit markets there would be less adverse flow there in general, and spreads would be tighten as market makers competed for the new flow.

As I work for a non-internalizing market making operation, I selfishly like the second option, but it's not really clear who is right. My guess is nothing is going to come of this investigation unless there some real shady stuff going on behind the scenes we don't know about - the basic premise of the business model is legal, even if maybe not optimal for the market.

Typically C++, Java or C#, although the technology in the currencies trading world has generally lagged behind say, US cash equities where exchanges are catering more to electronic customers (a substantial portion of the currency trading user base is still trading manually).

Reuters changed over to a more modern platform around 2013, but prior to that for electronic trading they would basically designate a Windows host for you that ran their GUI application plus a server, which you could connect to and would just forward your requests via COM to the UI.

Simple latency arb strategies are totally an arms race now, with very high barriers to entry. That does not mean however that latency becomes unimportant - if you are a market maker you're essentially playing the same game in reverse - you want to move your quotes out of the way to avoid getting your lunch eaten by those aggressors.

The quality of the questions on Bernie Sanders' AMA were pretty good, but it was informative only because he didn't shy away from answering the interesting questions.

One of the Chicago Mayoral candidates did an AMA in the recent election and it was pretty abysmal. The good questions were skipped and the other answers were evasive. In that respect, Reddit lacks something an in-person interview can deliver (with a good interviewer) - the ability to push back on a non-answer and not be ignored. You also can't catch someone off guard on Reddit, really - the answers are going to be very carefully curated.

Flexibility on work hours and the ability to work from home occasionally. Being able to flex around doctors appointments, errands that can only be done during normal business hours, etc makes life a lot less complicated.

It's worth pointing out that this kind of behavior gets people in trouble on a fairly regular basis, what is new is that the DOJ is starting to bring criminal complaints.

It's long been the case that the SEC, or in this case the CFTC, will fine and temporarily/permanently ban you from trading for this kind of behavior, look at http://www.cftc.gov/LawRegulation/Enforcement/EnforcementAct...

This is a great rundown. It's worth mentioning though that if you are interested in HFT in particular but not necessarily being a "quant", there are lots and lots of really interesting technical problems to be solved as a pure software developer.

The upside is not the same as being a very successful trader, but it can still be head and shoulders above working at the Amazon/Google/Microsoft's of the world. Writing high performance, high availability trading infrastructure is hard, and scaling it globally is even harder. A successful HFT firm will be iterating on it's infrastructure fast enough that there is always interesting work to do, especially if you like distributed systems.