EternalData

I mean, I was thinking more of romance as "imbued with idealism", but if differential equations can help me find true love, even better.

Interesting work. Would be better as a pure aggregator of individual postings rather than just boards as a whole but I think this is a valuable resource as is.

I took this to be a general call to learn and practice mathematics, and not as a summons to a career in academia, so perhaps I'm seeing it a bit differently.

I have to say as somebody who through attrition always feared math in university that it is meaningful to think about reasons why people fear embracing certain types of knowledge. It's only after the fact that I realize how useful and even how romantic applications of mathematics might be.

I might never be a starry-eyed or an often-drunk academic, but I've grown to really see my initial lack of mathematical learning to be an immensely high opportunity cost. So I can relate to anybody who thinks of structural reasons people might shy away from math and indeed all forms of intimidating knowledge.

Skokie, IL. The epitome of a principled defense of civil rights.

It's fascinating to think of a counterfactual world where humans don't need sleep -- sometimes I wonder how much more could be done. Sleep seems to be a biological constraint that could be optimized around.

This seems to talk about a weak form of human-machine integration.

Nevertheless, there are some fascinating examples. I remember how a Wired article talked about how advanced computing has made chess players like Magnus Carlsen so much better because he now fights against a computer that can overpower most if not all humans. It's made the marginal cost of practice so much lower as a result: you don't need the best grandmasters in the room to play against something that thinks like them.

As a result, similar to how better nutrition and health practices have augmented sports performance, I can see even this weak form of human-machine integration leading to incredible spurts in cognitive proficiency -- though perhaps it will be domain-limited for now.

Maybe green card holders/work visas can pay taxes, build companies, hire people and contribute to American prosperity while not drawing benefits, rather then get the door slammed in their face.

More people in tech should speak up about how a blanket ban impacts American prosperity as a whole. Cutting off your country from an entire pool of human talent makes no sense to me.

Yeah, I guess it starts by getting more people to run, regardless of their chances :)

I always love technological solutions that ease the path forward to meaningful action.

I wonder if there's a way to algorithmically determine how a candidate with subpar funding could beat somebody established with many more (a playbook of kinds). That would be the next step forward -- sure, you can run, but it flips the dynamic when you're smart about running to win.

This is going to be part of a greater trend of automation starting to affect fields considered to be white collar and paths to prosperity. I think the same is going to happen with financial analysts, entry-level lawyers etc. It'll be interesting to see the political response, especially given how charged the atmosphere has become around "preserving" jobs.

So, given that the stated goals of ride-sharing startups are to automate their workforce (with Uber basically saying "automation or bust"), I think there's a need to build economic analysis of what this means.

In theory, you could correlate data between https://movement.uber.com/cities and this analysis and start building a picture of what it means when a model with less variable costs starts impacting the market.

There's probably some utility to it: a lot of problems involve hacking together datasets, sometimes in dubious ways. There's also value, especially for startups that are looking to build simple neural net applications (ex: identifying plates of food from different restaurants) which are very data-dependent. Researchers may also want to reflect the cost assembling datasets (ex: MTurk, processing power) and open up datasets that may never have been open before.

My general sense on this though is that I'd like there to be more of an incentive for people to open up their datasets to the larger public. Maybe I'm being idealistic but a crowdsourcing type function where you pay for X dataset together with other users and then it's released under MIT, forever free etc.

As others have mentioned that'll probably bump against usage rights issues, a larger problem you'll have to deal with independent of your need to sell or distribute the datasets in question.

Ugh, the general tightening of the Firewall will not have many short-term ramifications on the startup scene in China (it'll probably lead to a fair bit of protectionist lift actually for mainstream service providers) but the long-term issues of closeness will lead to inferior service providers thriving under the protection of the government.

Regardless of the research dollars China is throwing at technology, this is no way to build a successful/enduring technology infrastructure.

I understand that -- as a parser, of course it's not its job to understand sarcasm or wit.

Just referencing the old Chomskian bit about statistical correlates and how relevant it is to understanding the fundamental principles behind the expression of language (rather than merely its syntax).

I guess it was a bad joke gone over. ah well

Ah, but can it tell when I'm trying to be witty?

My solution to this is to set myself weekly goals in terms of steps. It changes your routine when you do that. At 12.7k steps a day, I basically have to elongate my walking commute, and find create ways to stand up and move all day.

I think we agree on the larger point, just to the degree of the damage. I think I'm on a 7.5, you might be on a 9 in terms of damage AI can do to employment.

"And this post doesn’t even look at the text of the Amazon product reviews or the metadata associated with the products!"

time to bust out the sentiment analysis

Been using PyTorch for a few things. Love how it integrates with Numpy.

Machines are bad at transferring domain knowledge and generalizing/making unconventional linkages between topics -- what some would regard as creativity, they're also not going to be very good at empathy without essentially solving for strong AI and regularized data on human emotions -- there also are going to be human jobs because of protectionist instincts. ex: A machine right now would actually probably do a more optimal job of determining interest rates if given more economic data and more data on outputs/outcomes, but replacing the Fed and analysts under it with a machine is a political football and a half to even say.

If we solve for strong AI, we might as well kiss the species goodbye, never mind just jobs -- but the AI we have in between is not going to completely eliminate every category of work in the present, never mind future ones that might arise.

The problem with AI is that it's dynamite when it comes to capital-labour relations. You can't compete with something with perfect agency that never gets sick. Now, are there going to be new categories of work and new economies? Yes. But in the meanwhile, for a category of work that defines a plurality if not majority of the functioning economy (basically anything that doesn't require creativity across multiple domains or empathy), there are going to be SEVERE income share implications.

Keep in mind that labour share of income has gone down drastically in the last few decades and that as a result, real wages have stagnated -- which has produced significant unrest in developed countries and unseated the established political order. If that trend continues...

With most developed nations looking to focus on tech and an inverting age curve threatening pension systems, I wonder how long it'll take for everything to shift from rampant nationalism to more open borders (at least economic borders).

One of the co-founders is pretty big in the Montreal startup scene.

A lot of advances in deep learning have come from providing more data and semi-structured information, but there are a finite amount of fields for which this can apply. Using game theory dynamics can move AI to transfer of domain knowledge and can create more adaptable deep learning architectures. It could also be a pathway to "stronger" forms of AI.

Go Montreal! woot.

I've noticed that the number of applications MI can be embedded into is nowhere near its full potentiality. There's a lot of potential for people who are able to think on their feet product-wise and incorporate MI, even if MI is a black box to them. Echo your opinion.