There is such a thing as general intelligence which differs between different people. Arguing that IQ isn't real because IQ tests are imperfect, is like arguing in the year 1500 that temperature isn't real because all thermometers are imperfect.
Our lack of ability to precisely measure something does not mean the underlying thing is not real. There is such a thing as general intelligence which correlates strongly with almost every type of performance and life outcomes.
IQ tests are useful for measuring features of populations, but they're a very noisy measure of an individual's "general intelligence" (if such a thing even exists), with several confounders: whether you've trained to pass IQ tests, TDTPSATDIBCA [1], how well-rested you are, how stressed you are, how hungry you are, whether environmental conditions are distracting you… Many of these are also a factor in group averages, although in the context of measuring children's educational attainment, this is a feature rather than a bug: in that setting, IQ tests are a good measure (to the extent that educational attainment is something we want to be optimising for, which is another question entirely).
However, in this thread, we were discussing "the typical racist's use of IQs". Nobody was "arguing that IQ isn't real": you brought that up, unprompted. The lady doth protest too much, methinks.
But its frequency still rose by over 300x since the start of the chart, which might suggest that for some people it is a basic color term.
"English" isn't one monolith. Every English speaker speaks their own version, some closer to each other than others, and new features are constantly being added or removed.
The further you have to narrow down the set of speakers for whom it's a basic color term, the less of a basic color term it is for English as a whole. We don't have to have this argument about e.g. orange.
To use math as an example, you can always look up formulas. But after more than 1 "layer" of looking up, that quickly becomes impossible. Like, when I had to learn to calculate derivatives and primitives, I could look those things up. But when I got to linear algebra, I couldn't progress until I deeply internalized derivatives and primitives, because looking up formula A only for it to contain unknown formula B just becomes a mess.
Yup. Even if Anthropic spends that $40B back at Google and thus the money ends up there, Google is still out of $40B worth of compute or whatever the profit margin is on that.
It's why eating at your own restaurant is cheaper, but still not free.
I had this exact realization. Taking chords from a chord progression generator, putting it into FL studio, adding a random melody that stays in the key from a cool synth preset, some random drum loop, and end result? I guess it could be called music. Its a combination of sounds that doesnt sound actively bad.
I noticed the problem when I realized I couldn't make music in a specific mood or genre. Sometimes I'd finish my song and think "oh wow, a happy rock song" or "a sad edm song" or whatever but it was always just random chance where I ended up. With music theory knowledge I could always add more instruments or notes that could exist in that place but with 0 direction, whatever I made was always listenable but never more than that.
Over hundreds of hours of actively using AI for basically every area of my life, it has just never actually achieved anything besides giving me the feeling of productivity.
Ideas are mediocre. Plans are arbitrary. Research is untrustworthy. But telling it "generate me 100 ideas for X" feels really productive.
I think a version of me with no access AI will not just stay competitive, but even outcompete the version of me with unlimited access to AI.
I wouldn't say no one cares about it, and I am not at all sure it will fail - nor that it should; there are better futures available from here, also.
Honestly, if I'd cottoned on quicker to the guy's real problem, I would've treated him more gently sooner. It's not quite true that an addict can't help himself, and in the place where that's false is the hope of recovery. But to blame people for getting hooked on the shit Silicon Valley is pushing, would be like blaming people for getting an opioid habit when the hardest imaginable versions of that drug were handed out like candy for decades.
Exactly like, in fact. Some people on this forum have BOP numbers waiting for them. You know who you are. In time, so will everyone.
This is only true if the new technology is actually significantly useful. Which so far AI has not proven to be. Theres no reason to assume people using AI will, on the long term, outcompete those who don't.
Our lack of ability to precisely measure something does not mean the underlying thing is not real. There is such a thing as general intelligence which correlates strongly with almost every type of performance and life outcomes.
reply