tdees40

So it may not be illegal in all cases (although it is in many cases), but how does failing to deliver close out a short position?

I work in markets. No, you cannot just fail to deliver. I'm also not sure how the ETF thing would work. If it's 1% of the ETF you're going to hedge out the position using a giant notional. It doesn't work.

It amazes me that a fourteen year old company with a mature business model requires equity financing. I'll never really understand how tech companies work, I guess.

Why do they need financing? How is a company that sells $2000 exercise bikes not profitable?

Wendy and Lucy! Also, TV is better: Atlanta and the Wire are great about poverty.

I'm genuinely fascinated by this comment. You're saying that mutuals offer lower prices (not typically true), but they were out-competed by private companies? I'm not sure that makes sense. In my sector (life insurance), there's no real pattern to mutual vs public company pricing, but maybe it's different in auto insurance, for instance.

Yes, but the dollar also has a very long history of being extremely stable. Said another way, the dollar has never had a 30% shock in a year (BTC did today...)

Sure, but there's like umpteen million terrible accounts...seems like a lot to ask of your users.

No, how about users (me, at least) are leaving Twitter because it's a rotten cesspool of misogyny, racism and stupidity. Maybe users should get a choice to avoid some of this?

If you were looking to exercise, wouldn't you just ... ride a bike? And if you aren't looking to exercise, why do you want to walk on a treadmill while riding an e-bike?

Yes, there are lots of complaints about it handling curves poorly, misreading overpasses as solid objects, etc. etc.

I just ordered a Model S with Autopilot, and as I've been reading the comments on the various Tesla forums, I'm not sure I'm ever going to use it. Some of the stories are honestly terrifying (sudden deceleration on the highway, swerving into other lanes, etc).

It's a food company, and its product is fully baked. Why aren't they bootstrapping? Are they not profitable? If so, how do they have a business? If I sold ketchup at a loss, would I be able to close a $50M financing round?

My weird and unrelated question is: if I donate software to an open source group like the Linux Foundation, can I write it off my taxes? And if so, how do I assess the value of it? RethinkDB probably has some legitimate market value...can the founders reflect that on their taxes?

Yes, but Keras works just fine using Theano as a backend as well...

At this point I've used PyTorch, Tensorflow and Theano. Which one do people prefer? I haven't done a ton of benchmarking, but I'm not seeing huge differences in speed (mostly executing on the GPU).

I mean sure, there are. In fact, there are lots of pretty standard ideas of what constitutes profitability. But if you just say, we're profitable, most people will assume you mean on a GAAP net income basis. If you don't mean that, you can say, we're cash flow positive or something like that. Or you could say, "we're profitable on an EBIT basis," or we have a positive gross margin. You can't just say, hey! if you exclude a bunch of our costs and count all of our revenue, the revenue is bigger!

I want to put this comment in a time capsule so I can tell my kids what start-up life was like in 2016.

I've always wondered why more companies don't do this. Offer a 30 or 35 hour work week and watch the top-notch candidates roll in, even though you can't offer a top-notch salary.

I often played FIFA on my phone, and the trick was just to hire to best possible scout and then constantly send him out, and your team would be unstoppable pretty quickly. Scouts found outrageously good players with unrealistic frequency.

Banks and Bitcoin serve radically different purposes. (Try getting a mortgage from Bitcoin!) A more appropriate comparison would be the percentage of bank energy that goes to payment processing, which is probably pretty low.

I read Koine/Classical/Homeric Greek reasonably well. Message me (e-mail in profile) if you have any questions.

There's a growing body of evidence that Alzheimer's is really just "Type 3 Diabetes":

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769828/

I dunno the model but my dad had one.

This is a pretty silly piece.

First of all, the vast majority of LTCM's trades had nothing at all to do with the Black-Scholes-Merton formula (BSM). They were just highly levered mean reversion trades. The fall of LTCM had nothing to do with the BSM formula, it had everything to do with leverage.

Secondly, vanilla BSM with constant volatility isn't really used to do anything important anymore. It was going away much earlier than 1997 anyway; the Heston model came around in 1993, and Derman was using a primitive version of local vol around the same time. And of course people knew that financial asset returns were not normally distributed with constant vol (Mandelbrot wrote a paper describing that in 1963!).

I won't even get into why his (and Salmon's) description of how the Li formula was used is completely wrong.

N=1

I love Idris as much as the next guy, but I doubt virtually anyone is using it for production code. By contrast, there are millions upon millions of lines of Haskell code in production all over the world.

I love the idea that Haskell users are all wringing their hands about .... Idris.

The language in this is a bit silly:

"Nowhere is that more evident than in the U.S., where lending to the government should be far safer than speculating on the direction of interest rates with Wall Street banks."

You aren't "speculating on the direction of interest rates with Wall Street banks", you're buying a synthetic rates position that is centrally cleared with daily variation margin. That's not quite US Treasury safe, but it's pretty damn safe.

It's pretty simple really. Swaps are now centrally cleared, so you have to post collateral through a clearinghouse. If on any day, one party cannot settle up on their daily variation margin, the trade is terminated.

Compare this to before, when swaps were often un-collateralized, so you might have a huge paper profit on a trade that you'll never actually realize because the counterparty lacks the cash to settle up.

So before, you could lose the total amount of your trade, but now you can only lose a day's worth of gain or loss.

This is a big improvement, and the risk to the taxpayer is pretty de minimus. Of course the central clearinghouse could default, but that's very unlikely for a variety of reasons (mostly that the central clearinghouse's whole reason for existence is not to default).