But wait, this is invasive and potentially harmful, is there some way we can be a bit more sure about things before we confirm?

Why yes! This kid developed a blood and urine test which is 90% accurate!

The point is to potentially reduce the number of large, expensive needles stuck into someone's pancreas, not to serve as a standalone test.

It also matters WHY the test is inaccurate. If it's consistent with each individual "if I get a false positive, it will ALWAYS be a false positive" because of a lack of a certain protein or whatever, then it's less useful (unless you determine the conditions that make it work). If it's actually just a random 10% due to lack of precision for a particular measurement, then it can be refined, OR you could just run it five times and do some math to get a result with >90% accuracy.

ummm.... why? Current driveless car technology may use GPS for some data, but it's hardly necessary.

Google's driveless car can keep track of the road and oncoming obstacles just through a laser system within a tens-of-meters range which gathers data on close objects and then uses that data to extrapolate the position of further objects from a video feed.

I see no reason why a driveless car would ever really NEED GPS. The computer is perfectly capable of keeping track of the available turns which it has taken/passed up, which should be enough to determine its position (assuming you knew where the car started from).

In fact, cheap, ubiquitous, and wireless internet is far more useful for automated driving vehicles, as you would need a lot of bandwidth for cars to effectively keep track of each other.

If I'm navigating through Nebraska, all I really need to know to get a good-enough position is what freeway I entered on and how many exits I've passed. A video-feed could provide this information. Or even easier, a small signal emitter at each freeway exit could allow the car to record this info.

Say I'm writing drivers for these GPU's and I want to give my customers the ability to block this. In theory, the calculations required to mine a bitcoin follow a particular and repetitive pattern, which means I could probably identify the pattern of instructions being sent to my GPU. Could I code a driver to render the functions that draw the screen for the game normally but just return '0' every time I'm asked to mine a bitcoin?

I feel like I would invest in a graphics card that offered such a feature, as it should indicate to any gaming company trying to profit on my extra GPU power that I do not view this hijacking of my hardware as 'OK'.

Is this to say they don't keep the full text? Even anonymized or just identified as 'student 1..2...3'? I find that to be a bit of a shame, as a conversational dataset this large would be quite helpful for many types of research.

At least that would strike a decent balance between privacy and usability, at least for the active research project.

Well they're on sprint... So about $70/month for unlimited text and data = 840 per year.

Over the last four years that's $3,360 per teen.

For 175 Teens that's $588,000.

Then you have the monitoring software, the backend database. Half a million messages per month? Over four years that's 24,000,000 messages in an uknown number of tables. You might want to pay a person to make sure that thing stays running and do daily backups to make sure there are no gaps in your data if stuff breaks.

Then you have the army of grad students who are probably funded through that grant who are either sifting through the data themselves, or coding up machine learning applications to draw conclusions from it.

And this research project also existed before they started using blackberries (since 2003). So this $3.4 million seems to have gone a long way.

Except this has nothing to do with the network engineering, it's simply about whether or not the particular streaming video service counts towards a user's cap.

If Comcast chooses to cap, then neutrality would mean that their own XFinity content counts toward that cap, regardless of the actual load placed on the network.

Of course nobody would want to, but since every country has a different set of laws and regulatory bodies with different rules, it's pretty much the default for a multinational corporation to have different code running in every country. Even in the US, it's not unusual to have different code running in every state. It somewhat depends on how regulated your industry is. Advertising definitely has a lot of rules in certain locales.

FTFY

But seriously, conflicts of interest don't apply to advocates who are interested in change, there's absolutely no conflict since everyone knows the agenda you have before you even enter the room. Conflicts of interest apply to decision-makers, not some people who speak on a panel that have no actual influence.

If you have the time to perform such research for every patient you have, you probably ARE NOT a doctor.

There's no magic ability an M.D. gives you when it comes to understanding drug trial reports. A few statistics classes will do more for your understanding of a drug trial than any amount of biology.

That's kinda the point, if you prevent the ship from building them up then you can no longer collect them, silly.

That's how it works in criminal court. Copyright Infringement is a civil matter, which is why people are getting sued for money rather than placed in jail. In Civil Cases you (usually) need something called "preponderance of the evidence", which is basically the same as just piling legal briefs onto each side of a sea saw until one side clearly tips.

I would put this in the "notebook wins" column, but I have a somewhat extreme love for toast.

Siri's AI is just piggybacking on a DARPA project that has been running for 40 years utilizing the resources of the top universities with Artificial Intelligence programs in the United States. In other words, the company that developed Siri is using technology developed by people who have been working on the bleeding edge (which also means any competitor can just piggyback on the same project).

If Apple threw their cash reserves at AI research, the best it could do is hire the researchers who basically already built the AI behind Siri. You would definitely see huge advancements, but none that wouldn't have happened anyway, and they would instead be owned by Apple rather than openly published.