It's tricky because if you take the view that rightsholders licenses are necessary for training, it's not clear that the labels hold those rights. If you need the music producer rights, for instance, it's a lot more decentralized and harder to get a catalog.
I think my comment come of as too literal. It wasn't meant to imply that there is nothing of value to the club experience, but rather that a busy and loud environment like a club may over-highlight the stimulant aspects of the drug and tone down its more exceptional empathogenic traits, reportedly.
In theory you can claim it back, but it's a major pain in the ass. Some ETF issued outside the US and holding US stocks will do it.
Other strategies to avoid this annoyance include:
1. selling on the ex-dividend date and buying the next day.
2. buying a long-date call and selling a put instead of holding the stock
3. simpler version of 2: buy a deep in the money long-dated call, you won't be paying all that much for the convexity, and you don't have to think about dividends.
The reason to hold stock is that they return value to shareholders. Stock buy backs are functionally dividends but they are taxed more fairly. They also work better for foreign investors, and are a much cleaner and simpler mechanism than directly sending cash to shareholders.
Dividends are silly in a world where you can achieve the same result and save everyone 20% tax by doing stock buybacks instead. It's also silly to strongly differentiate between dividend paying stocks and non dividend paying stocks as opposed to focusing on the business' cash flow, unless you're purposely avoiding dividend paying stocks because they suck from a tax perspective.
Yet, dividends are entrenched in many investor's psyche. Even official regulatory training in the US teaches registered representative that it's a good idea to recommend dividend bearing stocks to people who want "income" from their investments, which is a fallacy. Large mutual funds also have special categories for stocks providing high dividend yields.
Sad that this is the strategy as opposed to leveraging the incredible opportunity they have from their fabs.
Whoever is backing him is taking on the risk of SBF jumping bail (and I'd assume the house is under surveillance and he's wearing ankle monitors), in exchange for either silence or valuable goodwill on the tiny off-chance he gets out of this.
But it takes an entire network of human neurons to simulate the behavior of an artificial one. There's overhead in translation, but that doesn't by itself mean one is superior to the other as a basic unit.
Spam on most platform is a complicated engineering challenge. However, for some inexplicable reason, Twitter features a kind of spam that you don't see on any other platforms because it could be caught with a regex.
It's almost never the case that spam issues on popular social networks can be alleviated with some easy fixes, because obviously if they could they would already have done that. Twitter is a very weird exception.
I wonder if there's any value in picking not one but a couple random directions, and then finding the update in their span. Increases memory consumption but reduces the noise. It could be that the trade-off is best when picking a single direction, but who knows?
If you pick 2 directions at each pass, one of them could be the direction of the last update and the other a random one, allowing for some kind of momentum.
I skimmed the paper, so I'm not sure I got this right, but my understanding is the following:
1. pick a random direction
2. compute the derivative along that direction using forward-mode differentiation
3. update the parameters along that direction based on the derivative
The idea being that this gives an unbiased (albeit noisy) approximation of the actual gradient. You thus need a smaller learning rate, but you also need less memory and computation and, net net, they argue it's a win. Is this correct?
Microwave frequencies used in ovens would be dangerous because they can penetrate deep in the body and heat internal tissues, but in the EHF range microwave heating could be safe. See for instance https://en.wikipedia.org/wiki/Active_Denial_System
With all that said, what I had in mind is in the far less scary and far more prosaic near-infrared range.
