Can't wait to see the bugs! GAN's are famous for some... interesting reconstructions. And better still, nvidia will have no way to debug it since the model is essentially a black box.

It's somewhat pointless without a female voice though. I don't think most players will assume you are anything but male unless you speak...

That doesn’t make it pointless. Specially since the results did not show equal treatment (which would be one thing that the assumption of being male until proven would explain)

Not pointless, but calling out misbehavior and giving education on the matter would be more helpful long term.

I think it's incredibly disingenuous to call yourself a self-taught mathematician when you already have a masters in physics.

Undergraduate physics gets by with surprisingly little, if any, advanced math, so it is only natural when one realizes how inadequate one's knowledge of math actually is.

Have to disagree with this. I'm a math undergrad but a good chunk of my algebra class this semester was physics students. Additionally, the author said they have a graduate degree. As far as I know (which is not very far) modern physics makes considerable use of algebraic constructions such as tensors and lie algebras, so I'd imagine the author would be familiar with these concepts, which puts them lightyears ahead of any self-taught mathematician.

Let's just say that there was very little emphasis on Math in my Physics Masters, which in itself was very poorly taught, and leave it at that.

Math is being able to do logic with numbers.

I've tried creating my own math as an engineer, it's possible, but I also know every attempt I've made failed at finding something new.

My point is, we learned a lot about numbers in engineering school.

Yes but physics != engineering. You really need very minimal higher math to do engineering work. As far as I know, at my school the furthest engineers go is a course in vector calculus, ODES, baby linear algebra and simple probability. None of which would be really be considered higher math. In contrast, physics needs complex variables, Pdes and most students take at least a group theory class.

Is there something about those topics that don't follow logic?

I think the result would be the same no? Assuming the characters are stored in ASCII the result of a single bit flip can only possibly modify one character. Of course this assumes the program is resilient against modification in general and not just deletion (which seems to be the case).

I wouldn’t assume so. A character has a lot of bits. So it kind of depends on how the stream is processed I think.

In my completely amateur opinion, that is the one major reason why value investing is so much harder today. When Graham and later Buffett were executing this strategy to enormous success, most of a companies valuation could be traced back to it's assets -- excluding intellectual property. Now that IP is such a large part of valuations, it's much harder to execute this strategy because IP is inherently harder to valuate than tangible assets. And consequently, it's much more difficult to accurately access the difference between price and value.

But surely having 80% of some revenue is better than having 100% of nothing.

Can you explain to me how to solve the decentralization problem then? It seems that with the current generation of blockchain, the problem is that eventually it will become no longer feasible to store the entire ledger on one device, and therefore some level of centralization will be required in order to keep the system functional. Throw in the enormous amount of (ever increasing) energy required for proof-of-work systems like bitcoin, and it seems unlikely that true decentralization is possible with current technology.

Storing the entire blockchain isn’t as important as having check point blocks that are agreed upon via consensus. At that point, you can prune the historical blockchain state and new nodes can sync from multiple nodes to get the latest state, starting with the most recent checkpoint.

With things like state receipts, light clients can rely on other network nodes to provide the state of transactions that were recently confirmed, without giving up security.

This does require that some nodes are processing and storing the most recent N blocks and their transactions. Today, non-mining full nodes do this for free out of kindness, but in the future, this could be offered as a paid service instead.

Proof of work is only one option for securing a chain. Research into proof of stake and delegated proof of stake will hopefully make PoW a thing of the past and drastically reduce the energy requirements to secure a blockchain.

The Ethereum team has two main focuses right now:

1. Proof of Stkae transition (Casper) 2. Implementing sharding

At the same time, teams are working on sidechain technology such as Plasma, Plasma Cash, and Loom which will essentially allow transactions to happen on another blockchain whose state is persisted in the root chain.

This not only benefits blockchain users (higher scalability, little to no fees), but is also good for the blockchain network since the data persisted to the main chain is only a set of merkle roots from the sidechain, no matter how many transactions the sidechain has.

They even use one in an episode of the wire (circa 2003).

> The interesting property of cryptocurrencies is that the supply of money is constricted and can't be manipulated by any nation's central bank.

I don't understand how people believe this is a good thing (not implying you do specifically). One of the leading factors in the economic recovery following the 2008 crisis was Quantitative Easing and expansionary policies by the fed. Without these tools at the government's disposal, if there was another credit crisis, the effects would almost certainly be much worse and prolonged.

Like any tool, sure, it can be abused. But getting rid of it entirely by switching to a decentralized ledger (ignoring the reality that total decentralization seems infeasible), eliminates the very tangible benefits of having a central bank.

> I don't understand how people believe this is a good thing (not implying you do specifically).

It's probably the same people who think that the Gold Standard was a fantastic thing and abandoning it was the worst thing ever.

I think the effects would have been worse in the short term, but the long term effect is much worse -- signaling to an industry that wanton risk taking will essentially be rewarded and if you're large and connected enough the government will debase the currency to keep you in business, so you're free to fuck up again.

I think you're arguing on the side of "countries need monarchies because the people can't be trusted" and "printing presses allow seditious ideas to spread and must be outlawed or tightly controlled", but, again, I can't be certain.

How is it that we can have two totally different interpretations of the truth? Isn't economics a science? Why is it that there are such varying opinions? I think it's because economies are too complex for any one person to accurately model. What you need is markets to decide what works and what's valuable and what the prices should be. What you need is competition, and that's what cryptocurrencies provide.

Bitcoin is already competing with weak currencies of corrupt governments and winning. A lot of improvements need to be made before Bitcoin is anywhere close to making a run for the #1 currency in the world, but if you're watching the technology closely, there's still a lot happening in crypto and it might be sooner than you think.

> I think you're arguing on the side of "countries need monarchies because the people can't be trusted" and "printing presses allow seditious ideas to spread and must be outlawed or tightly controlled", but, again, I can't be certain.

Hard to imagine how you reached that conclusion from what I wrote.

I do somewhat agree that having the precedent of a prior bailout might encourage some potentially stupid behaviour that might not otherwise be considered. The alternative, however, is much worse (in my opinion). The point isn't to keep Morgan Stanley or Merrill Lynch or whoever in business, the point is to make sure that the rest of the country doesn't lose their entire life savings because of one recession. If the cost of economic stability is propping up a few rich people, it's probably worth it considering the alternative. Of course, I'm sure the anarcho-captilist blockchain experts might feel differently.