It's really interesting to think about this. Very different types of complexity.
The smallest possible correct program implementing a tax calculator is going to be a LOT bigger than the smallest possible correct program for controlling a 60's era manned flight to the moon. The latter is more complex in other ways, of course.
Complexity can be "deep" or "wide". software for rockets is "deep and thin complexity", but calculating hyper-local minimal tax bills is extreme "wide and shallow complexity".
I think the point was that people who currently choose not to work are not harmed by automation of low-barrier-to-entry jobs that they are already choosing not to take.
And it's a valid point! If people really are choosing not to take jobs that are readily available, then automating those jobs is a net benefit.
Now, we should not ignore the fact that this is stated as a hypothetical! Are there a mass of unemployed who are choosing not to work? And would all of the people who lose their jobs due to burger flipping automation be able to find new jobs or be happy with no job?
> If you want to claim you know what fair is in any given situation, then go and hardcode all your own fair rules
This strikes me as eerily similar to the argument that type systems are impossible because of the halting problem. It's sort of true in some sense, but not in an even remotely useful way. So it mostly functions as a way of derailing the conversation away of the more subtle distinctions that do matter (e.g., could we design an easy-to-use type system that rules out this particular type of non-termination/other class of bugs).
There's a large middle-ground between "hard code all your own rules" and "completely unconstrained learning". Learning under constraints is not a new idea.
A classical programming analogy to your argument might be "well the halting problem is undecidable so ignore all this high-level language stuff and just go code up your own turing machine; it's the best you'll ever be able to do".
> because you aren't going to find "fairness" in machine learning.
Why not? The human notion of fairness is fuzzy, which is why researchers have provided various formal notions of fairness in machine learning tasks. Obviously, these formal definitions may or may not correspond to your own gut instinct about what is "fair". And there might be friction between different notions of fairness. None of that should be surprising; otherwise, fairness wouldn't be something that philosophers continue to bleed ink about.
But it is equally obvious that for some notions of fairness, there will exist machine learning algorithms that learn well under the given constraint.
I agree with you when you say fairness is fuzzy. Though this implies that there are no set of rules that define it, that also means there is no way to train on it. By choosing the right features and utilizing it in the right way I believe you can avoid putting people in bad situations for bad reasons. I'm saying, if you train an algorithm to guess if someone is involved with crime, it's going to be incredibly stereotypical with it's answers. Same as if you rely completely on statistics.
I'm not actually saying hardcore all your rules. I was making an example that if you really know exactly what fairness is, then program it. But we both don't know there are no definite rules. So why hardcode at all?
I think it is never the ML that is "unfair", it's the one who made it who is responsible. If you're finding yourself running into issues in your ML with unfairness, I think you're just using it wrong.
The problem here is that modern machine learning is rarely more than a computerized way of building statistical models based on observed data. And the bias mitigation techniques I've read about all revolve around (indirectly) manipulating those models based on what results they produce. This can lead to various problems in the long run.
> What universe are you living in that you think modern American children want to learn? They don't.
That's sad. You should spend time with some different kids. The ones I spend time with have an insatiable appetite for learning new things. Each student is passionate about something different, of course.
> They don't look at public schools as refuge, or as empowerment.
That seems mostly orthogonal to the question of whether they want to learn and whether they realize that learning can be empowering.
> What universe are you living in that you think modern American children want to learn? They don't.
Woof...
I know when I worked around the US (mostly midwest) at a different school a week, there were always some kids who wanted to learn.
There were always some that didn't.
However, they may not have wanted to learn every single thing that the school taught, but they did want to learn about stuff that could be applied to a future career.
> “While we regret that Google did not want to continue a long and fruitful tradition of collaboration between the military and technology companies, we are even more disappointed that Google apparently is more willing to support the Chinese Communist Party than the U.S. military,” they wrote.
The amount of cognitive dissonance in this quote is ridiculous. When I originally read this qoute, I was expecting to find further down that Google is doing defense research with China.
But nope. Google is contracting with a Chinese manufacturer to build cell phones, not performing weaponized AI research with the People's Liberation Army.
Apparently, it's unpatriotic and borderline traitorous to say "you know what? I'd rather focus on building some decent cellphones instead of finding new ways to kill people from afar."
Miniaturizing electronics, increase battery, signal processing, improved cameras, ML on phones, voice and command recognition on phones, computational photography, object detection on phones - are just some of the stuff that can be used to improve phones, but are really easy to find military uses for.
...and all those things are developed by literally hundreds of other companies that often interact with Chinese businesses. The US could prohibit export of advanced battery technology or image processing or...
This is not about national security. This is very explicitly a response to Google's desire not to cooperate with the US military on explicitly defense-related research. Two senators literally drew that line for us.
It's worth noting that US Senate Intelligence Committee members (specifically, Mark Warner (D)) have previously stated concern about the apparent closeness of Huawei to the Chinese Communist Party.
Even assuming Huawei is wholly controlled entity of the CCP, Google's collaboration with them is for generally peaceful purposes. Obviously, widespread use of Android phones manufactured by Huawei could present a security risk. But peaceful collaboration inherently creates such risks.
Surely, give and take is the simplest way to put favourable global relations. This unfortunately doesn't seem to be the mode as of late. I'd wager the risk was recently increased with the tariff one-upmanship the US and China are currently engaged in.
I’m assuming you’re employed? Not everyone will have a brita filter or the inclination to pay out their ass for bottled water.
EDIT: I don't mean to be a dick or even argue, I'm just pointing out that these problems do affect people, just maybe not most people through an effective combination of factors that reduces risk (where you live, whether you can afford a filter, etc). I certainly grew up on tap water and I am mildly disturbed people are ok with it causing massive mental health issues.
EDIT 2: Finally, I don't filter water I cook with; I imagine lead would still be a problem there. I probably should filter that water too in Oakland....
I always wonder if these "toxins in the tap water are tolerable" people ever eat at restaurants. Do they know where the water used to prepare food and make coffee and tea comes from?
>>> If that wasn't the case, all of these other nations wouldn't be upset about the US change in behavior... The fact is, all nations should pursue their self interest.
>> This leads to hugely destructive wars.
> Nonsense. Self-interest != hostility.
Right. Just because someone is looking out for their own self interest doesn't mean that they are constantly "winning" a zero-sum game with someone else. In fact, sometimes cooperating is the most self-interested move you can make...
Which is exactly why all of these other nations might be "upset about the US change in behavior" without necessarily prescribing any course of action that isn't also in the US's best interest.
Kind of like how "cookie" is not web specific but unless mentioned in an explicitly nonweb context, it means browser cookies.