throwawayjava

I'm confused.

If the test data was collected in the same way that the training data was collected, this was a rather ridiculously round-about way of writing what should've been a 5 line perl script, and I'm really curious where the 5% error could've possibly even come from :-\

I mean the free speech of anyone who doesn't have $N million needed to fend off a civil suit. If cases like this and the MJ one become common-place, there's a very real risk of chilling effect.

Incidentally, I think it's probably possible to allow Thiel the ability to sue Gawker into the ground without also having a system where genuinely legal free speech is reserved for people who can afford massive financial risk. So there's absolutely no reason for you to disagree with me just because you think Thiel in on the right side of the Gawker issue...

Citizens United itself was a 5-4 decision, so it's fair to say that CU was also a far-from-clear result.

Pointing out problems with the current system is not a logical fallacy, and having a working solution in-hand is not a pre-requisite to observing the existence of problems.

Also, accusing me of using the Nirvana fallacy while also accusing me of not providing an alternative policy is literally a contradiction...

From the article, it does seem that the training and test sets were both generating from a dubious data collection process: scraping URLs they knew were fake/satire/real and then applying a label based upon the domain.

But, they had more features than just "source" in their model. so while it's possible their data collection method means that their model is way over-fit and basically just tests of proxies of "source", it's not prime facie obvious that this is the case...? Or am I missing something?

Citizens United is also terrible.

But designing a justice system around "maybe you'll find a charitable person willing to help" seems like a bad idea.

That needs to change, and you can find your way to this conclusion via sympathy with Hogan or with Gawker, but perhaps not via sympathy with Thiel (for the obvious reason).

Our civil justice system needs to be cheaper and more accessible. It shouldn't be possible to wage economic war on a company or individual via the judiciary.

>> Personally, I'm concerned that the individual factors of this case (i.e. Gawker being Gawker) will overshadow the principle at stake, which feels a lot more important than a gossip site.

Thiel isn't on trial in this thread, but our justice system perhaps is.

There is a difference. It's a microcosm of the difference between a god-king and an elected president.

The general concern is that money provides one the ability to silence other people's speech.

Well, I'll happily admit that I'm not guilty of doing ornate argmentative gymnastics to justify my position... I consider that a feature, not a bug, though ;-)

AEI and other conservative think tanks were throwing fits about over-reaching interpretations of the commerce clause throughout Obama's entire presidency.

AEI's current position on the commerce clause is motivated by their partisan preferences.

If the answer is "no" -- and, seriously, the answer is most definitely "no" -- then your "AEI isn't partisan" claim goes out the window.

Because the internet is a global network and China/Canada/etc. aren't under the auspices of the FCC.

And yet, AEI is advocating for the current interpretation of the Commerce Clause because it empowers deregulation. You can bet your last horse that they'd be making the opposite argument and rambling about King George and unelected regulators if Democrats controlled the federal government.

I can't help but feel that you've reinforced awinder's point... AEI reasons backwards from what it wants on a case-by-case basis to justify its policy preferences, which change depending on which party is in power. If that's not the definition of partisan, then your definition of partisanship is broken.

> If anything, they're smart, by framing a Republican-sponsored policy within the context of case law that Democrats have consistently supported for a century.

1. Yes, exactly, their constitutional justification is an insincere post-facto justification for policy they like, and their policy preferences when it comes to the commerce clause largely depend on which political part is in control of the federal government.

2. But they're totally non-partisan, right? ;-)

The NYC and federal advisory councils would be tasked with providing lawmakers with impartial advice on A.I. writ large. That's reasonable. Bias in algorithmic credit decisions and bias in algorithmic sentencing are different problems, but share a lot of the same underlying technological realities/problems. These areas do have cross-cutting concerns, so a legislative brain trust dedicated to disseminating knowledge about these cross-cutting concerns is a wise use of resources.

No one is suggesting "regulating" general AI, so the author's red herring question about "what is AI anyways?" is completely off-topic. Even "alarmists" like Musk and Hawking are proposing much more concrete legislation, such as banning algorithmically generated kill orders in weapons. And to the extent that those people are interested in safe general AI, they're currently trying to achieve that goal by throwing research money at it (rather than proposing concrete legislation regulating general AI). Which is hardly a reasonable thing to criticize!

Furthermore, the proposed approaches in NYC and at the federal level aren't incongruent with the author's proposed sector-by-sector approach toward regulation! These are, after all, simply proposing advisory councils.

Therefore, the author does not provide a compelling reason to oppose any of the concrete legislation mentioned at the beginning of the article. And the arguments that the author does make betray, IMO, a superficial understanding of the proposed legislation.

"The results demonstrate that intelligence is a powerful predictor of success but, on the whole, not an overwhelmingly better predictor than parental SES"

I.e., intelligence is about as likely to explain success (or lack thereof) as parental income. This is exactly what I was saying -- if you want to guess whether a 5 year old is going to succeed, take a look at their parents.

In short, I don't think this paper is the home-run for your case that you think it is. If what you're claiming is true, we'd expect intelligence to be radically better as a predictor than SES. But it isn't.

The paper also notes some important criticisms that cut that right to the core of your use of this paper in the context of the "nature vs nurture" debate.

Edit: the second component of your edit (" In the long run...") is absolutely NOT justified by the paper you cite, and is wild conjecture at best.

This makes no sense. A smart but poor kid does not have the option to move to a rich suburb with excellent schools.

> so if you have a really smart kid you aren't nearly as likely to find that kid still in the trailer park when they are 30 as one of their more average peers.

1. But, you're just as likely to encounter that extra bright kid smoking pot in a trailer home in 20 years. Poverty is a hell of a burden and barrier. Intelligence can help you stay out of poverty, but it's far less powerful when you want to get out of poverty. And can even function as a disadvantage.

2. Furthermore, lots of dumb but rich kids will go on to six figure incomes for the rest of their lives.

Intelligence isn't as strong a determinant as you claim, in either direction.

There are a few important observations that the article and the IPPR report don't stress enough.

1. The jobs most likely to be automated, will be automated using technology that's inaccessible to the employees whose jobs are being destroyed.

This point is often optimistically papered over using observations like "coal miners can go to coding bootcamps!" I know a few truckers and I have no doubt that each of them could go through a relatively short coding academy for web development work if it became a "do or starve" situation. But I doubt that most of them could become much more specialized autonomous vehicle software engineers, either due to raw intellectual ability or -- more often -- due to an inability to make a substantial half-decade investment in retraining (most of these folks would need to (re-)learn a lot of high school algebra).

So just because we're replacing truck drivers with programmers, doesn't mean that the folks displaced by self-driving are somehow getting jobs that were created via the destruction of their prior employment opportunities.

2. "Partial Automation" could mean less hours worked, but is more likely to mean lower pay and more hours worked, and not all "partial automation" is equal. I.e., to destroy trucking as a middle-class profession, you don't need a self-driving truck. You just need to automate enough of the hard stuff so that anyone can learn to drive truck in a week.

I think it's useful to separate the partial automation that de-skills work (level 4 autonomy for trucks) from the partial automation that simply means less equally-skilled work (web dev). The former creates more job opportunities, but turns a middle-class job into a low-class job. The latter just eliminates from middle-class jobs altogether.