Machine learning and NLP techniques have started to show you can get some very intelligent software without strong AI. Watson was able to beat the best human players in essentially a Q&A game, similar to what a digital teacher would need to be capable of. If you look at the evolution of chess AIs, affordable personal computers capable of doing what Watson did will be released before the end of this decade. Cell phones will be able to do it a decade from now. Two decades from now Watson will appear downright stupid compared to the Q&A software we will have access to in our watches. And it won't require strong AI.

Just like all sufficiently advanced technologies, the AI of 20 years from now will appear like magic compared to what we have today. With or without strong AI.

He is talking about 20 years from now. Technology has a habit of changing at an accelerated rate, so to envision what technology will provide in 20 years it is probably better to compare today's technology with 1975 tech. So take a look at Siri and Cortana compared to voice recognition and natural speech processing in 1975. Take a look at the amount of information is retrievable in Wikipedia with what existed in an Altair 8800.

Imagine going back to 1975 and describing the world wide web, ordering on Amazon, the iPhone 5, etc. If you aren't thinking of a world as different in 2035 as we are from the 70's, you aren't thinking in the right frame of reference. Teaching may not have changed much in the past few hundred years, but locomotion didn't change much until the railroads and automobiles either. All it takes is a tipping point of technological advance to make the future unrecognizable to what we have today.

In 20 years, a student's computer will likely be able to teach then a subject in 10,000 subtly different ways. Voice recognition will be so good it will seem like magic by today's standards. It will be like having a 1000 to 1 teacher to student ratio. We probably won't even rely on prerecorded video lectures that far into the future. The software could generate the lecture in real time, adjusting to the slightest input from the student (direct questions, facial expressions, heart rate, etc).

Humans will almost certainly still be involved, but the role could shift to supporting the technology instead of the other way around. Just as hybrid human-computer chess teams are still better than chess AI programs alone, hybrid teaching styles will likely still be the best. But technology is likely to open up possibilities in teaching that are unheard of today.

Then again, we still don't have flying cars, so who knows for sure.

I'm just waiting for someone with money to bring the situation to its obvious parodic conclusion and be carried up to the top on a gilded litter dressed like Xerces.

Thanks to Google's project Calico, I would like to be the first to wish Dan a speedy recovery.

Philips runs 59 R&D facilities across 26 countries. It takes advantage of talent in all of these countries, including the US. The fact that two of its many subsidiaries are headquartered in the US is no indication that Philips is a US company at heart (like you insinuate in the last statement of your post).

Sony also has various headquarters in many different countries. It is no surprise that its movie and music subsidiaries are headquartered in the US, but that is no indication that Sony is a US company that just happens to be headquartered in Japan.

I am not ignoring anything. My post was not a detailed analysis of every pro and con of outsourcing labor and I didn't claim it was. I merely stated that outsourcing exists, and that industries can and do move overseas. Neither of these claims are false.

There are plenty of complications that still allow massive discrepancies in pay between the developed and developing world, but no complications are impossible to overcome. My father in law travels to China a half dozen times per year to fix these kinds of problems in his company's Chinese based manufacturing plants. These problems are very expensive, but overall it is still far cheaper to manufacture overseas. Many of the problems you mention make it very difficult to offshore IT jobs as well, but there is always still a cost point where it is better to deal with those problems and offshore anyway.

Plenty of companies offshore services that are not completely commoditized, although yes they rarely offshore the core and most innovative aspects of their company. But just as there are plenty of engineering related jobs that shifted overseas when a large amount of the US manufacturing industry moved offshore, there are plenty of other STEM related jobs that are not that innovative as well. I wouldn't doubt that the US could lose half or more of its STEM related jobs without moving much of the innovative sectors of the industry offshore.

Also, it is not a given that the US will continue to be the center of most innovative aspects of the economy. The Large Hadron Collider is one high profile example of the US dropping the ball and letting some of the most innovative physics research in the world leave the US. Many if not most of the greatest large scale engineering achievements in the last couple decades have been accomplished in Asia, not the west. China's total R&D spending has already eclipsed the EU and probably will beat out the US within a decade.

While the US still has a lot going for it, simply assuming it will always be the world's leader in innovation is naive.

I am a native born US citizen (with 3 native born grandparents if it matters) who works in the IT industry. I just don't have any naive ideas about American exceptionalism that make me believe my country is untouchable by the rest of the world. I want our country's economy to stay strong for my children and other future descendants and simply feel that protectionism is not a good path for the US.

People who complain about H-1B visas usually have a misguided view of what the real options are in this debate. They see an option where companies don't use H-1Bs and simply hire more US citizens instead. The reality, however, is that the real options for companies are:

1. Bring in H-1B visas so corporate IT teams stay in the US
2. Build corporate IT teams in other countries

Option #2 is essentially outsourcing, and it is not just some boogeyman intended to scare US workers. It really happens. Entire industries have already moved overseas in the past century, and the software developer and other engineering industries are not immune to it.

If US citizens cannot compete with foreign labor that live in the US, with a similar cost of living as US citizens, we have no hope of competing with foreign labor abroad with a much lower cost of living. There has been a push back against outsourcing software development jobs in the past decade, but if we start practicing protectionism the trend can easily start moving in the other direction again.

Yep. It's wierd because the symptoms can correspond with many different causes. For example, the climate change thing makes sense because bees can be tricked into thinking it's spring and start foraging or even swarming in the middle of winter when they really should stay in the winter cluster. The occasional warm day is good for them to be able to get out and void themselves, but longer periods of significantly fluctuating weather can be bad.

But it also matches other problems. Diseased or dying hives often lead to "desperate" swarming where bees start abandoning the hive to try to establish a new, safe place. Most of these swarms, however, will die. The behavior could be seen as a general "exteme stress" behavior. It could also be seen as a neurological disorder from pesticide exposure.

In short, it could match almost any possible cause. And probably is a result of many of them.

I dunno, they might be a bit boring.

I think it's important to ask questions because there's been literally "dozens" of different things "definitively linked" with CCD. The public likes to seize on neonicotinoids, but they're probably one of the least supported of these many different "definitively linked" reasons. Whole countries have gone so far as to outright ban neonicotinoids, with no effect on CCD. France, for example, banned them. The next year they largely switched to blaming the condition on Asian Hornets when the decline rates didn't decrease.

The problem is that when you ban a certain pesticide, people start using others. And going from neonicotinoids to organophosphates is a massive step backwards in terms of general safety, not just to pollinators, but especially to more complex animals as well. But the biggest problem with the neonicotinoid theory is that neonicotinoids are only used in a small fraction of the areas where CCD exists. Bees can only fly several kilometers from the hive, they're not going cross-country and picking up every pesticide in every farmer's arsenal. It even exists among people who are in places where no pesticides at all are used.

It's easy for the general public to latch onto a particular cause. But once you learn more about beekeeping you realize how incredibly much out there is that can utterly f* up a hive. And which have in history regularly collapsed bee populations, far worse than the collapses we have today. Trachael mites once nearly obliterated beekeeping in Europe, saved mainly by the development of the Buckfast bee. Check out this very inexhaustive list of bee pests and diseases. There's even some really counterintuitive effects in that small levels of some pesticides can actually increase hive survival rates, in that they're deadlier to bee pests like mites than to the bees themselves.

The public also tends to totally understand colony collapse disorder in the first place. Normal winter colony death levels are about 15% in most locations (though where I am it's higher). CCD raised the US average to about 30% at its peak. This is painful and expensive to beekeepers, but it has literally no impact on the ability to sustain bee populations. A new beehive can be started with just a queen and a handful of workers. Hives can be made to produce queens en masse through proper management. Hence people can mail order starter hives, and there's never going to be a threat to the ability to produce these starter hives - a single hive can make many dozens per year. Even normal hives not managed for breeding starter hives will naturally produce several swarms every year; beekeepers try to discourage and/or catch these swarms.

In all likelihood, neonicotinoids are one among many different stressors to bees in the modern era that causes CCD. Modern bees are much more "stressed" than bees in the past. We've created an environment where new bee pests and diseases have spread far and wide to bees that never would have encountered them in the wild. We raise them on corn syrup and sugar water in the winter (good for reducing dysintery and increasing honey yields, but robbing them of certain vitamins and minerals). We transport them on flatbed trucks hundreds or thousands of kilometers (these are animals that get confused if you move their hive a couple meters; their ability to navigate by sight is poor, they're best navigating by the sun and dead reckoning). And countless varieties of poisons, even unintentional ones, affect them every day of their lives. There's so many factors now that weaken hives that any "new" factor to an area can push them over the edge.

I'd really like to read the paper but unfortunately it's down. But for example, do the neonicotinoids add a UV signature to the liquid not present in the sugar water? That would have little to no influence in the case of flowers in nature (where they're not looking at the nectar, and there's all sorts of other chemicals in the nectar). What other chemicals are in the neonicotinoid solution (they're rarely pure, they usually have all sorts of other chemicals to increase their effect)? What's their cleaning and handling procedure for preparing and filling the sample containers? I want to know how they controlled these experiments against factors that humans can't detect but bees absolutely can.

Just the very act of hooking electrodes up to bee neurons I'd have concerns about. Is there any induced electric field involved, or even rubbing against the bee hairs? Bees transfer information to one another via dances, such as the waggle dance. Bees build up an electrostatic charge on their body, and a waggling bee imposes an electrostatic force on the antennae and hairs of all adjacent bees, causing them to feel dance over a short distance. Their stereoscopic sense of the dance lets them know the direction, and that combined with the time allows them to work out a direction to a food source relative to the (moving) direction of the sun. It functions like transferring a memory from one be to another. There's also "negation" behaviors, by other bees who don't like the information giving out; they have a different frequency buzz to say "don't go there", and sometimes different bees may even fight with each other over what's good and what's bad information.

Also note that the linked articles refer to a second study published simultaneously which showed no effect on honeybees next to rapeseed fields sprayed with neonicotinoids versus an altogether unsprayed field. Which is pretty remarkable, because you expect almost *any* pesticide next to your hive to have a profoundly negative effect on it.

I'm sorry, but calling flagging a "troll" because they misread an article is beyond the pale. None of their behavior was "trollish". Saying that a study is flawed is in no way shape or form engaging in "fraudulent research", aka, deliberately falsifying data to push an agenda.

The Nature article appears to be down. But I have to caution, studying bee behavior is very difficult. Many of our senses, bees lack or have only at low resolution. But they have a number of senses that we don't. They see UV. They see polarized light. They sense electric fields. They're sensitive to a lot of chemicals that we cannot detect. And so forth. It's very, very easy to accidentally give bees signals, which will alter their behavior, that you didn't realize you were giving. I'd like how they attempted to control for all of this, but unfortunately that's not possible now.

Cause and Prevent are two different words. One requires an all or nothing, the other does not.

You can say that someone caused a car accident even though they were not in control of every aspect of the accident.
You cannot say you prevented someone from getting beat up if you stop one assailant but his buddy still beats the guy up.

Of course vaccines don't prevent measles. That is one reason why the herd mentality is important since it is not 100% effective (the other is for people who cannot be vaccinated for medical / age reasons). Vaccines are over 99% effective at preventing measles, but you need to put those stipulations in there if you are going to talk about their effectiveness.

If you had originally said but almost certainly eating right can prevent 80% of cancer cases then there would be no problem with your statement, except for possibly the validity of whatever percentage you gave. You could have also said but almost certainly eating right can reduce your chance of getting cancer. But instead you said it could prevent cancer, which is not true.

Small differences between words are important, especially when you are making scientific claims (which all health claims are).