Is that for a certain order size? Because a lot of the cost can come simply from retooling or swapping components. Like if you have 100 x1 smt components vs 1x 100 smt components it's a huge different if you get what I mean. Also anything through hole I basically just give up getting assembled as a hobbyist.
Meh it doesn't know what's "easy to understand" because it's trained on the average of the whole internet and it's mostly junk out there. Sadly the amount of junk obviously vastly outweighs the actual good writers so as long as you're just feeding it like this nothing will change. RLHF also just biases towards the average/mediocre because it's again the average of all user choices. If you want a PhD level chatbot you might need to enslave a couple hundred PhDs and milk them for a few years first. But under duress they will not perform optimally either.
Of course it matters. Suppose because tech pays better the figure in tech is more like 0.7%. The rest is shut concentrated in rural regions and other industries. Then your point would be moot wouldn't it.
Though many events are cherry picked, the trend is still quite a nice fit it seems. I would say exponential growth is rather plausible.
In fact it doesn't even need to take aliens to change our timeline. In Cosmos by Carl Sagan he often talks about so and so Greek guy that was basically correct but dismissed. They had some inkling of the experimental method but didn't pursue it. They knew about the motive powers of steam. They just didn't think much more of it I guess.
Does it actually help? What about just drawing censored versions or those covered up with a leaf kind? I'm not trying to be difficult but just curious how the whole thing works. Like I don't think drawing clothed models just makes you not understand human anatomy. Also totally understand if it's just some tradition thing people so or whatever.
It's so crude that it might not even be the way neurons work. Like just a programmed activation function? Who programs it? We inherit the weights some how? Are they compressed in DNA? Doesn't seem likely because we share 90% with fruit flies etc. So not much space is left to store weights. I don't think it's the weights doing the work in biological neurons, but the architecture. Every cell might be doing the same thing but it's how you use them that counts. Maybe a biologist can comment on this.
Just gotta hope that based on how economic theory is supposed to work, if doing these things result in saving costs during procurement of specialized equipment, the savings are passed on to the consumer… hopefully…
That's a loaded question. How about just saying it for what it is. Are you willing to spend 9 hours today (not counting the other interviews you have to do for other companies) for a X% probability of getting a job at company A which you like Y amount.
If X and Y are large enough sure. Most of the time X is hidden and Y is hard to gauge.
This is the same argument that says YouTube can play 3 unskippable ads, TVs can scan and collect data on all devices in your LAN, because they're a private company and can do what they want, and if you don't like it, just don't buy it. Well in that case, I can also smoke in front of kids, because I am a free agent and there's no laws saying I can't.
A full day is a lot unless you get compensated or something for it I feel. In today's world it is a lot of ask. I could also not be spending any time there, in which case that day is wasted. Imagine if all interviews of all your applications took a full day… probably burn off half your PTO just interviewing…
Wow that's insane. For me, if I sleep too late, regardless of how much I sleep, I still feel tired. But I'm definitely not a morning person because I just hate going to bed early. It might be a work thing too, idk.
Looking forward to the write up, and would you provide some info before that happens, like how does one even start analyzing your own genome, where do you even sequence it? Isn't it supposed to be huge?
Indeed human speech by pure bits is incredibly lousy. But you know why it's so hard to do STT and even TTS? Emotions, accents, interrogatives etc. You don't capture just counting information theoretic bits in words.
I mean Ryan gosling face has 0 bits of information in blade runner and yet we know he's pretty sad and emo.