I think neural nets and deep learning systems are misunderstood so profoundly because humans are bad at many kinds of reasoning and the explanation of the neural net being "smart" is so psychologically appealing.

I think this is the aspect of it that applies to Bitcoin. All of the steps along the road to hyper-inflation in a fiat currency may be quite reasonable and may constitute the smartest move using the available tools.

There is also the question of whether currencies should be tied to governments. The risks that (frequent) government failures pose to one of the core purposes of money (storing value) are not costless. Society bears the costs of those risks even though they are hard to measure.

Of course many of the same problems could apply equally (or more) to Bitcoin depending on how the governance process proceeds.

Goldbugs / conspiracy theorists have a line of argument where they describe all the fiat currencies that have ultimately experienced corruption and hyper-inflation. So I think it's fair to say that running a fiat currency successfully for more than 500 years is something that is unprecedented in the history of the world.

So over a long time horizon this makes Bitcoin extremely interesting. It's not a question of whether 1 BTC will one day be worth millions of USD, it's a question of when. There is a good chance it will be hundreds of years from now if at all.

Much financial risk is in fact sovereign risk in one form or another. This doesn't mean it's rational to hoard gold or BTC, but for entities that have a long-term view of their own future, it makes sense to care a little bit and to diversify.

Even if 10% of the long-term entities diversify into Bitcoin, that alone will drive the price up substantially. It's far too soon for that to have happened.

We are also entering an era where politics are once again a bit part of international exchange, which adds additional sovereign risk for many areas of international business and financial planning.

But Bitcoin faces the same kinds of risks as governments for corruption, mismanagement, etc. The genius of it is that the governance model makes it a lot harder for one party to really control what happens with it.

Now, a few years later, BTC is blessed by regulators. The next phase of world domination is for the governance model of Bitcoin to start to seem far better than fiat governance models. This is just a matter of time.

More specifically, price is a function of supply and demand. Bitcoin was created with a very specific bootstrapping plan baked into the design.

What people are missing is that the bootstrapping plan is well known and obvious to investors, and is meant to incentivize a speculative motivation for mining, which it has done successfully.

But think about it this way, the price of a currency is only loosely linked to supply and demand. Nobody really knows how many dollars exist, yet the currency has characteristics that make it trustworthy.

Bitcoin is the same phenomenon. The price is based on the success of the governance model and the appealing characteristics of the ecosystem.

Based on these appealing characteristics, there is the widespread expectation that Bitcoin will win market share from other currencies over the long term.

We know there will be a finite number of Bitcoin mined, what we don't know is how much market share Bitcoin will have in comparison to other currencies.

Market share is not a function of money supply as much as it is a function of the holders of the currency that rely on the currency because of its governance mechanism, fungibility, etc. Many countries hold USD in reserve because they find the governance characteristics of the USD appealing. Bitcoin is just a novel way of doing currency governance.

For all uses of currency other than holding inventory, the governance mechanism matters very little, since there is little risk exposure to price fluctuations or the risks associated with bad governance.

Critiques of Bitcoin get mired in an imprecise understanding of all of the above, but the most notable blind spot is that Bitcoin is a governance mechanism first and a currency second, and investors are pleased because the governance mechanism has been tested a few times and has (thus far) performed admirably.

The answer is that regulators did an abysmal job of understanding the perverse incentives that plagued the industry. Among the biggest causes of bad judgment was the incorrect price signaling created by GSEs that were not following proper accounting and disclosure procedures.

On one hand the financial firms had significant regulatory capture and had been enjoying lots of profits due to their success influencing regulators.

But on the other hand, the entire system of "markets" that were most relevant to the crisis were the most heavily regulated and tied to specific policy goals.

It seems odd that we'd ever expect a system that is so politicized to ever be regulated in a rational and appropriate way.

Broadly viewed, we can see that the combination of regulations and areas loosely regulated and left up to the discretion of firms constituted a significant degree of centralized control, which turned out to be "corrupt" enough to result in a lot of bad decisions and the crash that ensued.

If we assume that all systems are prone to this sort of "centralization risk" we can better appreciate the benefits of decentralized governance that exist with some block chain systems.

Forget about jail for what happened in 2008, why should we ever trust financial regulators of financial firms again to regulate our financial system responsibly? Are we supposed to believe that the perverse incentives for regulatory capture, socializing risk, etc., suddenly ceased to exist?

I'd argue that we should not. We should realize that human institutions typically require participants to have some trust in other participants, but that the more trust is required the more vulnerable the institution is to the kind of problems that plagued the finance industry.

We trusted the GSEs to be acting responsibly even though no financials were released. Regulators trusted ratings agencies to apply disciplined processes to rating generation in spite of the profit motive not to do so, the public trusted regulators to ensure adequate underwriting of risk capital, etc., etc. All these things, many of them not even measurable due to the significant accounting slop involved, were vulnerabilities waiting to be exploited.

When a building has marble pillars outside and everyone inside is wearing expensive suits, what we are seeing is signaling of trustworthiness. When the banker is wearing $1500 Italian leather shoes and a $50K watch we can assume he's earned those things by being trustworthy over time. When we enter the high ceilinged room and see the marble we are meant to trust the institution itself. After all, how could this structure, meant to last thousands of years in the elements, not indicate the highest level of accountability and honesty?

We must realize that even the most well-intentioned institutions are vulnerable to centralization risk, aka the corruption of the inner workings and mechanisms in a way that is not at first noticeable but benefits insiders.

The finance industry used to be simply about risk, money and time. But in today's world it typically follows the pattern of taking money as an input, and producing as an output financial products that foist off some of the risk to society so that the owners can make a profit, with the downside risk being borne by society as a whole.

We see this process in action time and again, and it will continue to happen as long as our regulatory approach rewards massive firm size, prevents competition, and socializes losses.

Let's hope that we see an emergence of an alternative system that relies on a lot less trust and is much less vulnerable to centralization and corruption.

I'd argue that it's not really about ideas as much as it is about the public being much more accepting of authoritarianism.

The broader values of free speech and free expression come not from an aesthetic judgment of particular acts of speech, but from a desire to limit the power of authoritarian governments to suppress speech/expression.

We're seeing in both parties an amnesia about the downside of authoritarianism, and this filters down to things like office culture, university culture, and the perception of the trade-offs associated with various freedoms.

Using the pan repeatedly as a griddle will result in a superbly seasoned pan that does not need much attention or cleaning.

Once in a while if the pan is used for something too gooey or saucy, or if it gets put through the dishwasher, I simply do a quick seasoning as follows, before resuming its typical use as a griddle:

1) Rinse or wash the pan, use a scouring pad or steel wool if there are any burnt bits of food stuck to the surface.

2) Wipe cooking surface of the pan with a thin layer of olive or vegetable oil. Wipe it in with a paper towel and do not leave any excess in the pan.

3) Heat the pan over a burner at medium heat, so that it gradually warms up, then becomes hot enough to produce a small bit of smoke.

4) Pour in a bit more (room temperature) oil, and once again wipe/rub this around the cooking surface of the pan.

5) Let the pan heat for another minute or two until it is once again at the smoke point, then turn off the burner.

This quick process takes about 5 minutes. It thoroughly dries the pan after washing, which prevents rust, and it gives it enough initial seasoning to bootstrap the seasoning process for subsequent uses of the pan as a griddle.

But now when the price is high Coinbase can't manage to scale the website so that people can sell?

It would not be hard for coinbase to engineer a meta-strategy using targeted service failure to extract additional profits from the ecosystem's obvious trends. Not saying this is going on, but there are now to indications that this could be happening.

Someone is intending to buy at $12K so you fail the transaction for 3 hours if your simple model expects a higher price tomorrow (meanwhile buying some coin at market to use later when you succeed the transaction).

Suppose at every turn there are n possible future states of the game based on the rules. To avoid "brute force" the AI must be able to ignore many of those states as irrelevant. In effect, the AI is learning what to pay attention to, not just considering what might happen, thereby conserving computational resources.

Chess and Go are interesting for two nearly opposite reasons: 1) because they are too large for humans to consider the reasoning obvious, and 2) because the input to the reasoning is simply a small (and easily perceived by humans) grid of rule-constrained pieces.

But when you think of AI in an information theoretic way, so that given representative training data the system (if large enough) will always "learn" perfectly, it's not really all that remarkable. It's just a different computational way of doing the same transformation from input states to moves. Given a problem (chess, go, etc.) the researchers must simply learn what network structure and training regimen will do the job with the least computational cost.

To see why this is relevant, consider a deep learning model that could continually generate successive digits of pi (or primes) without having the concept baked in already. Would the result be computationally cheaper than highly optimized brute force algorithm? No, because what it would "learn" would be something already known by humans. Perfect chess is simply a function from input states to moves that humans do not already know the definition of. Most humans do know the definition of this function for the game of tic tac toe by the time they reach middle school.

I'd argue that while this is useful it's ultimately not hard. Comparing it with Stockfish mainly demonstrates how chess is hard for humans to reason about and hence hard for humans to write non-brute-force algorithms to solve.

Thus, I think this is an example of "weak AI" even though humans associate chess with high degrees of exceptional human cognition. Chess data contains no noise, so the algorithm is dealing only with signals of varying degrees of utility.

I'm looking forward to AI that can be useful in the midst of lots of noise, such as AI that analyzes peoples' body language to predict interesting things about them, analyzes speech in real time for deception, roulette wheels for biases, and office environments for emotional toxicity.

Chess is interesting because we can't introspect to understand what makes humans good at chess (other than practice). So many human insights and intuitions are similarly opaque yet the data is noisy enough that it will take significantly better AI to be able to do anything that truly seems super-human.

Thus far, Google's major platforms have had roughly the same longevity as a typical series C startup. I will keep this in mind the next time Google announces a promising platform technology.

My personal view as a user of both companies' hardware and software is that Amazon is taking the lead. I bought a Fire TV a few weeks ago and it comes with a remote and turns any monitor into a useful, standalone entertainment device.

I also have a chromecast, which requires an app to stream (and the streaming tends to freeze or hang and frequently gets confused if you try to stream something from a different app while another stream has hung).

I also bought a $29 Kindle Fire that strongly outperforms my $200+ Nexus 7. When Google started getting into the hardware market I was hoping we'd see affordable, subsidized hardware, but instead we're getting high priced "luxury" phones which happen to be a bit cheaper than those sold by the market leader.

Of the two companies, I think Amazon has a strategy that is poised to take the lead in hardware. I say this after also buying a BLU Amazon ad-supported phone for $59 that has a super large screen and performs somewhere between an iPhone 5 and iPhone 6+ (I bought it a year ago).

One nice thing about the Alphabet reorganization is that we can see how hard this sort of competitive landscape is really hitting Google. We've already seen lots of free tier services getting taken away and many other signs of lower margins, and so the threat to block Youtube feels like an admission of weakness.

Also, FWIW, the minute Google blocks the Fire Youtube app, Amazon can simply make the icon load Youtube in the web browser.