codekansas

This is spot on. Currently there is very little accountability for predatory student lending. For my mortgage I had to talk to multiple people on the phone, show proof of income, and have a background check. For my student loan I just had to give my parents' socials and prove I was attending college. I know nothing about the laws around this, but I suspect it was set up this way with good intentions. Unfortunately changing this will probably require a nation-wide acceptance that not all degrees are created equal.

It would be interesting to see insurance companies have more skin in the game - we would probably start getting much more accurate breakdowns of ROI by major, degree, and school.

I think this is a different issue. It is more of an attack on some fundamental principles. Like, even if some older people in the 70's thought hippies and beatniks were lazy no-goods, the fundamentals - free speech, the right to protest and disagree - protected everyone. I think the worry today is that the next beatnik kid is too scared to say anything outside the status quo.

Personally I feel like the traditional defenders of liberalism have gotten too limp dick and internet-unsavvy. Ultimately I hope that as internet-native people age and mature they will realize the importance of the loftier philosophical ideals required to make progress actually happen.

Seems obvious when you consider that Udacity, Coursera, MIT OpenCourseWare were all started at universities

Easier said than done

https://time.com/5739688/facebook-hate-speech-languages/

The naming, "US Digital Service Academy", seems to suggest that it will be tied to the US Digital Service - which would be great. I think that might be the only government organization that I would actually be excited about working for.

Side note, I really hope the tour-of-duty model becomes more popular with techies, and more encouraged by employers. I can see it having numerous bidirectional benefits

Hard to tell without a free press

An under-appreciated aspect of this "war" is how bad US companies have been at building ML-first products. ByteDance in particular is just fantastic at it - their products feels designed with the algorithm in mind, as if everyone involved in the development has some understanding of machine learning. This exposes some weaknesses with existing social media platforms that were built before the deep learning boom, as he says in the article. It is probably worthwhile for American innovators to think about what otherwise-impenetrable areas this could expose.

Interestingly, this is basically by design.

> No two countries that both had McDonald's had fought a war against each other since each got its McDonald's. [0]

[0] https://en.wikipedia.org/wiki/The_Lexus_and_the_Olive_Tree

Flash holds a special place in my heart. I didn't take CS classes until college (probably because it's so hard for middle and high schools to find CS teachers), but my middle school did have (back then) Macromedia Dreamweaver and Flash installed. I have such fond memories of sitting in the library during lunch and after school and following YouTube tutorials to figure out how to make dinky little Flash game. I still think that Macromedia Flash was the best introduction to programming I could have had and I can't for the life of me see why Scratch has become the defacto introductory thing for programming for kids

Is it just me or does this happen for every article about anything China-related?

Reminds me of this passage from Cory Doctorow's book Walkaway:

"You realize that they're exactly right," Sita said. She smiled, stopped. "You know, when the first walkaway prostheses project started, most of the people contributing had lost an arm or a leg in Belarus or Oman, and were tired of paying loan sharks for something that hurt and barely worked and could be remotely repossessed by an over-the-air kill-switch if they missed a payment. But once they got here and started living, realized how much had been left on the table by conservative companies that didn't want to get into a patent fight and didn't see any reason to add advanced functionality to something that you didn't have any choice about, they got radicalized.

"They stopped saying 'I just want to make an arm that'll get through the day,' and started saying 'I want an arm that does everything my old arm did.' From there, it was a short step to 'I want an arm that's better than my old arm.' And from there, it was an even shorter step to 'I want an arm that's so outrageously awesome that you'll cut off your own to get one.'

A somewhat confusing point: the author of this is Chris Cox the former US Congressman [1], not Chris Cox the Facebook CPO [2]

[1] https://en.wikipedia.org/wiki/Christopher_Cox [2] https://en.wikipedia.org/wiki/Chris_Cox_(Facebook)

Do you have a more reliable source than this website? I worked on content moderation at Facebook for several years, I highly doubt that the accounts were deleted because of government pressure

I wonder how much of this is because of physiological aspects that have to change to keep up with your other lifestyle changes. Brain exercise takes energy, and you might just not used to eating and sleeping enough to maintain that level of exercise.

My girlfriend (who is in medical school) actually brought this up to me when I was talking to her about how it seemed that not only the rate of R&D was increasing, but the time to market was also falling dramatically. It's cool to see these kinds of developments being utilized so much faster

There's at least one large tech company that has located themselves totally outside the typical tech hubs - Epic Systems, in Madison. I interviewed for them after graduation but ended up taking a job at Facebook in Seattle.

The offer was for half my current salary. They tried to argue that with the cost of living adjustment I could buy a house, but what 22-year-old wants to buy a house? Locating in Madison is kind of a physical manifestation of their business strategy - avoid any competition, pay sub-market rates and issue strict, possibly non-enforceable non-compete clauses.

It works for people who really want to live in the Midwest. I think there's a spectrum between "live to work" and "work to live", and I think most ambitious college grads actually want many opportunities, for a while at least.

People underestimate how much of a problem is... There are so many qualified MDs who are underutilized or unable to go into the residency that they'd be most productive in simply because the supply of programs is kept artificially low

There's a cool package called Nengo from Chris Eliasmith's group at UWaterloo which hasn't received as much attention as I think it should: https://www.nengo.ai/nengo-dl/examples/tensorflow-models.htm...

It's cool how articles like this are able to take a vague feeling that probably a lot of people have and put it in more concrete terms. It is well-written.

They mentioned that some post had gotten big enough that it had merited commentary by Snopes. When the vast majority of political discussion takes place over a distributed mess of memes the nature of fact-checking changes tremendously. On top of that, a lot of these pages are building up the ideological armor to rebuff attempts by the platforms themselves to provide more rigorous filters (e.g. "corrupt Twitter / Reddit algorithms"). Maybe this is just the way the world works now, but if so that's kind of depressing.

Beyond feelings about the verdict itself, it is interesting that a single lapse in journalistic integrity has the potential to put a major news organization like Rolling Stone out of business. There are massive incentives to publish sensational articles, but as this case and the Gawker case have shown, there are also massive incentives to not publish specific details. The BuzzFeed model (which is perhaps shared by social media at large) is to simply provide a platform, and let other people stir the pot. I hope that doesn't become the only way to circulate ideas.

On the contrary, the concept of a "neural network" is beautifully cohesive. Many of the architectural differences are only iterative; the fundamental units are very similar. Like how there's a lot of diversity in biology, but the fundamental units are shared.

From an ML perspective, recently there's been a lot more work done on language modeling (e.g. billion word corpus). Good language models can usually fix those types of perspective. So hopefully you can expect to see that change in the future!