jakecrouch

I view modern AI as a combination of things that are too early and too late - data science hasn't changed much since the 1990s, but people have a misconception that there is still a lot more to do in data science because of the undeserved excitement around deep learning.

Raising taxes is not the only way to improve public infrastructure. One can also lower the trade deficits, improve the technology the government uses, and legislate away some of the problems (jails would be less full if drug offenses didn't carry such harsh penalties). Taxation should only be tried when nothing else has worked. With 50% marginal income tax rates in many states, the US leans much more towards the communist than Randian side of the spectrum as it is.

Interstellar travel will consist of information travelling at the speed of light. Pretty much nothing else is practical.

What's your reasoning for this? You would only have to accelerate at 1g for 3 years to hit 99.5% of the speed of light. You can use an electromagnetic shield for dust. It seems conceivable that this will be possible eventually, even if warp drives are not possible.

But having good social skills confers life-long benefits. So, don't write them off. Get good at making a good first impression, being funny (if possible... this author still working on it...), speaking publicly.

I would be careful with this point. Young people who don't have great social skills already tend to feel that they are somehow missing out on something important. But those who have good social skills in our culture will have a hard time forming new and unique ideas, and will often get talked into believing big, fundamental ideas that are wrong or crazy. I'd compare it to being a smoker in the 1960s. In exchange for being cooler, you suffer irreducible risk, that is made more dangerous by the fact that few people realize it's there. Until our culture changes, it's probably better to encourage people to learn about human nature than about "social skills".

While this is a cool result, I wonder if the focus on games rather than real-world tasks is a mistake. It was a sign of past AI hype cycles when researchers focused their attention on artificial worlds - SHRLDU in 1970, Deep Blue for chess in the late 1990s. We may look back in retrospect and say that the attention Deepmind got for winning Go signaled a similar peak. The problem is that it's too hard to measure progress when your results don't have economic importance. It's more clear that the progress in image processing was important because it resulted in self-driving cars.

One explanation of the inflation of the public and private markets in the US is that the Chinese are in the middle of a massive debt bubble, anyone with cash there has nothing good to do with it, so they've been willing to invest in the US at almost any price.

It's nowhere near as bad or pervasive as the competition to get into student debt. But I'm not really sure what psychotherapy would work. Maybe convince people to read about mimetic theory.

It's worth noting that this basically gives you competitive tips - how to be more effective at solving a problem after it's been given to you. I think it's much more valuable to be good at recognizing important problems that others don't believe are important.

I believe that their problem will be finding the small group of users that derive enormous value from it. When you really dig into all the applications they've suggested, they don't quite make sense.

The book "The Seven Signs of Ethical Collapse" tells the stories of companies that suddenly collapsed due to hidden issues, like Enron and Worldcom. Tesla fits every sign: larger-than-life CEO with massive executive turnover, board of directors very deferential to the CEO, conflicts of interest (Solarcity acquistion), "innovation like no other", attempts to silence criticism, pressure to maintain numbers (Model 3 production and resulting problems with factory workers, firing employees and contractors to reach profitability).

I find the fact that they have an executive departure rate on par with what Enron and Valeant had most concerning - when large numbers of senior people don't vest their stock grants, it's normally because they know there's very something wrong that's not yet public.

It's especially suggestive that the VP of finance and chief accounting officer left in March, at the same time as the Fremont factory was put up for collateral and a "special purpose entity" was created to hold $546m in car sales, presumably to make the company appear more creditworthy.

I would be skeptical of new entrants to scooter / bike sharing. In China, ofo and mobike together captured about 95% of the market. If it were possible to challenge the companies that have reached scale, one would think one of the many dozens of failing competitors in China would have figured out it by now.

Lightsail Energy tried and failed to bring this to market - not sure whether this can ultimately outperform Lithium Ion Batteries, which have all the pressure of the electronics industry behind them.

I am interested in knowing if there are advantages to using blockchain generally outside of Bitcoin - I don't understand it in this case from reading the post. A centralized exchange could be fully transparent, publishing every transaction, and handle every type of debt, if it wanted to.

What is the advantage of using the blockchain rather than creating a fully centralized debt exchange?

This is a legitimate concern. The Revolutionary War, Civil War, and World War II are all spaced roughly 80 years apart, and all coincided with deleveragings.

There are many other parallels that can be drawn between the 1930s and the 2010s. Ray Dalio's theory about long term debt cycles explains this in a non-ad-hoc way - every 50-75 years an economy's debt becomes too high and must be deleveraged. In a globalized economy this will happen to many countries at the same time and lead to tensions between debtor and creditor countries.

It's obvious that there are limits to how well machine translation can work unless the models have sensory grounding. I wonder if the problem is that people haven't figured out how to do sensory grounding or that the hardware is still too slow for it to work.

I'm somewhat uncomfortable with brain fixation because there is no way to actually verify that it works. Mind uploading is very far off if we still can't upload C. Elegans. It is more plausible that we would be able to test biological revival within the next couple decades. Going from atoms to bits is much harder than going from atoms to atoms.

Perhaps this is the solution to the problem segways had, social acceptability. If it becomes mainstream among college students, it could easily spread to everyone else.

And they call this a "product".

To be fair, this is an absurdly competitive space. If they achieved something of similar magnitude for a problem dozens of other teams weren't already solving, it would seem much more impressive.

My theory is that the only reason the SEC hasn't classified all cryptocurrencies as securities is that doing so would be too unpopular politically. Anything that fails to get wide backing before the regulators hear about it, will get cracked down on.

The best way to advance AI is probably to make the hardware faster, especially now that Moore's Law is in danger of going away. The people doing AI research generally seem to be fumbling around in the dark, but you can at least be certain that better hardware would make things easier.

There is kind of a Gresham'a Law with these buzzwords, where if you actually have new and interesting ideas around them, it's much harder to get people to take you seriously.

This is why the US should not worry about China as a long-term economic threat - authoritarianism has never worked. The Soviet Union, 1930s Germany and Italy, all collapsed. I think the Chinese realize this and are pessemistic about their own future. The Constitution was designed to make the US Government as weak as possible, because a strong government ultimately turns on its people.

Tough question: How will this get distribution? How would I find out about it without reading Hacker News?

People at the top of our economy are generally oblivious to how bad things are for the bottom half. The top 0.1% commands the same wealth as the bottom 90%. Historically wealth inequality finds its outlet in political unrest, and there is probably more of that to come in the next decade. In the last election a socialist and a populist both got suprising amounts of support, even though the broader economy was supposed to be doing well. The next recession will hit the bottom 50% hard, and we could see truly strange things if that occurs near an election.

Wealth inequality is the result of a debt bubble that's been growing for several decades, not technology. High household debt to GDP and wealth inequality probably coincide today for the same reason they coincided during the Great Depression. The ways these debts have to be restructured or dealt with distort the economy, in a way that tends to massively benefit some while hurting everyone else. If you own stock in or work for a health insurance company, the last decade has been fantastic. On the other hand, the government will soon find that it's having trouble keeping up with the debt from its healthcare programs. If you started a company in Silicon Valley, you benefitted from enormous pools of VC money that might have been in treasury bonds if the Federal Reserve hadn't already been buying trillions of them. If we look back historically, wealth inequality began to normalize by the 1950s only after the US's debt was erased. The story that runaway technology is creating massive inequality is unlikely be true in a country where GDP growth is only 2%.