12,479 karma · joined April 18, 2014
Like I enjoy my apple products, and I'm sure glad Apple wasn't run by a psycho like Musk, and didn't put Ads in the OS like Microsoft. But I don't think any of this is heroic or anything. Like if anybody's a hero it's probably the open-source guys who do it for no money at all.
Also the crime I have witnessed (e.g. homeless people) would not have been in any way shape or form prevented by tracking cellphones or whatever palantir is selling.
I think it's very, very easy for an emotional citizen to conjure up images of being mugged for $200 while very easily ignoring the thousands of dollars we each lose by billionaires cheating on their taxes.
Frankly I don't know the last time anybody I know has been affected by violent crime. But scarcely a week goes by when somebody in the executive doesn't do something criminal or criminal-adjacent (e.g. insider trading).
Like are we living in the same world? IMO the Epstein-class is 100 time bigger a threat than getting mugged, and if we are going to have be scrutinizing people it needs to be measuring that our public servants aren't taking bribes, selling classified documents, meeting with Russians, etc.
Especially when EV vehicles are already working and taking over the market.
And as status-symbol or identity statement, being anti-oil (or anti beholden on America, Russia, Iran, etc) seems like a pretty good one.
Feels like "shakes things up" isn't really the clear scientific breakthrough I was hoping to read
I wonder what their statement will be on the matter.
I know people pretty reflexively downvote questioning this, but I question this. I think some people are afraid that even asking this moral question is somehow inciting violence.
I think it's quite believable that the possibility of force is actually essential to keeping institutions in-line. Certainly a lot of civil rights progress was a lot less peaceful than I was taught in school.
- a few people still specialize in it
- a lot of people are helpless if the problem exists at that layer, and just ignore or workaround it because it's not cost-effective to fix
- maybe in 20 years AI will already have closed those gaps. If not there may be a big supply/demand incentive for experts which would create high salaries and cause more people to specialize in it.
(To be clear I'm not trying to argue this is good, these are just my predictions)
I guess you're trying to draw a false-equivalency between taking a problem extra seriously and denying/perpetuating it? However taking a problem too seriously doesn't harm people, if you want to wear a mask out of an abundance of caution you won't kill anybody else.
Also nobody believed the world was going to end in two days, that feels like a disingenuous talking point. If somebody literally believed the world would end in < 10 years they'd likely quit their job, spend all their savings, etc.
If your point is that you've met ~15 individuals in your life who were obnoxious/self-righteous/unlikeable about their attempts to make the world better -- congrats every movement has that. But it can't distract from the fact that one thing is true and the other is false, and anybody who tries to focus more on the stereotypes of the individuals in a movement than whether it's true or not is only creating noise.
Most junior engineers today (70% ?) will need to switch fields.
I think for a while the test was passed. Then we learned the hallmark characteristics of these models, and now most of us can easily differentiate. That said -- these models are programmed specifically to be more helpful, more articulate, more friendly, and more verbose than people, so that may not be a fair expectation. Even so, I think if you took all of that away, you'd be able to differentiate the two, it just might take longer.
I'll also pick apart this question "The mainstream media gave Trump an 8% chance to win the 2016 election. Him winning clearly shows they were biased and wrong."
Well, defined biased. Pollsters weren't using perfectly random models, so there was systemic bias (as there always is when you can't have a truly representative sample). There was also a sense of group-think that "This couldn't possibly happen, this is beneath our country" from a lot of individuals in and out of media.
That said it doesn't prove a deliberate attempt to deceive, but it certainly was a wake-up call to many in the media.
Point here being -- almost all emotional discourse is taking very nuanced situations and trying to cram them into semi-arbitrary judgmental terms like "fair" and "legitimate" and "biased."
1. Formally calling out a concept for judgment-based skills that cannot be easily taught. I think everybody understands this, but having a word for this would be useful.
2. Opening up the conversation on the topic of which types of skills can/should be codified and how.
That said, everything else in the article is suss. "Dimensionality" is largely a distraction to try to sound smart, most of the claims in the article are unwarranted (e.g. processes & checklists can be great, even for disciplines with true experts like airline pilots).
For example saying that skill learning cannot be accelerated is just patently false in many domains -- take something like learning chess. If you have a coach and other tools you'll learn a LOT faster. But certainly I've worked places that wished they could automate away reliance on experts because it gives organization power to a few non-management individuals.
You teach the machine by asking it to solve some problems, and then whatever answer it gives you say "That's exactly right. Now we train on those answers YOU just gave me" (even if they are wrong) and repeat. Somehow THAT works over time.
You use the outputs from the first run (right or wrong) as answers for the second training run, and repeat. Magically it works. That's what's so surprising.
I guess a theory is because there are so many diverse ways to be wrong that they don't accumulate error... still seems surprising and would be interesting to see if it works in other domains.
Laymen also correctly have an intuition that the people doing these studies aren't entirely trustworthy. What they don't have is a clear picture of how much work goes into these studies, who's doing it, what their motivations are, etc.
In my opinion studies when they can, should record videos of all data and make it publicly available online. Watching somebody do 1,000 hours of research is more proof-of-work to lay-people than some semi-coherent summary-for-a-layperson article.
It almost feels to me like acting as though a famous person being gasp wrong about something is implicitly suggesting that this is somehow surprising?
We should be surprised and write essays when the smartest people we know say something silly. Just because somebody's bank account has some zeroes in it doesn't mean it should be worthy of our focus.
Given that increasing model size doesn't yield proportional increases in intelligence, there is a world where these datacenters don't have a positive ROI if we make these models even a fraction as effective as the human brain.
The main breakthrough [rotating by an orthogonal matrix to make important outliers averaged acrossed more dimensions] comes from RaBitQ. Sounds like the RaBitQ team was much more involved, and earlier, and the turbo quant paper very deliberately tries to avoid crediting and acknowledging RaBitQ.
My understanding is that the efficacy of these methods isn't in dispute, what turboquant did was adapt the method that was being used in vector databases and adapted it for transformers, and passed it of more as a new invention than an adaptation.