841–850 of 1001 posts
At first I thought "why would Elon choose to slow down progress like this?" And then I realized that he probably knows some things that I don't. He's always one step ahead, even if he doesn't seem like it. I wouldn't be surprised if he's in control of OpenAI a year from now. Having Elon as the brains and Altman as the salesman would be the ultimate dream team.
Earlier quoted context omitted.
From their 2015 introductory blog post: “OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.” Today’s OpenAI is very much driven by considerati…
Right. But it seems like everyone agreed that they'd need a lot of money to train an AGI.
Earlier quoted context omitted.
The primary source of superiority in a war is "not being within reach of a enemy munitions". US does great in European wars for that reason, and is historically mostly free from North American wars because US dominates the whole continent.
Dominating the whole continent isn't enough, or Britain would have won WW1 much earlier due to the support of Australia - who dominate the whole continent.
Earlier quoted context omitted.
Morals, ethics, and not being a jerk aside, Musk has now proven that he is just plain not thinking things through, when he: 1) Released the original Boring Company idea about tiny tunnels under LA to alleviate road traffic, which could easily be shot down by anyone modeling it on a single paper napkin. 2) Got caught up in ontological (theistic) arguments about alien scientists who had surely created a computer simula…
This reminds me of a quip I heard on twitter and found quite funny. “Elon is the stupid people’s fantasy of what being smart is like, just like Trump can be a poor person’s fantasy of what being rich is like.” Don’t agree entirely in the poor part, but the Elon part I find quite accurate.
I may be trying to compensate for having been an Elon stan pre-2018ish, but I do have to give him a lot credit for being the founder of SpaceX, which is the best launch provider on the planet. He was also the CEO who made Tesla what it is. He did accelerate the adoption of EVs. He truly believed in the physics of both of those endeavors, he was right, and I loved him for it. A true inspiration.
BUT.. around 2018 he got high on his own supply. I had watched every Elon video up to that point, and there is one old one where he said ~"I fear that I could get too ego driven, and have no one around to call me on my shit, and that really worries me." He was right to worry about that, because he sailed right through that with a vengeance. The first couple years of that were tolerable, but now, what a devolution. I hate to admit it, but it really broke my heart.
Now back to my first paragraph, Musk is worse than Trump politically, aka the part that matters now, because he is sucking a lot of intelligent people down to the lowest of human nature. It deeply disgusts me.
Earlier quoted context omitted.
> it is precisely the controls upon it which are trying to be injected by the large model operators which are generating mistrust I would prefer things were open, but I don’t think this is the best argument for that Yes, operators trying to tame their models for public consumption inevitably involves trade offs and missteps But having hundreds or thousands of equivalent models being tuned to every narrow mindset is t…
> Yes, operators trying to tame their models for public consumption *Falsify reality.
Most of the idiocy surrounding the "chatbot peril" comes from conflating these things. If an LLM learns to predict that the pronoun token for "doctor" is "he", this is not a claim about reality (in reality doctors take at least two personal pronouns), and it certainly isn't a moral claim about reality. It's a bare consequence of the training data.
The problem is that certain activist circles have decided that some of these predictions have political consequences, absurd as this is. No one thinks it consequential that if you ask an LLM for an algorithm, it will give it to you in Python and Javascript, this is obviously an artifact of the training set. It's not like they'll refuse to emit predictive text about female doctors or white basketball players, or give you the algorithm in C/Scheme/Blub, if you ask.
All that the hamfisted retuning to try and produce an LLM which will pick genders and races out of a hat accomplishes is to make them worse at what they do. It gets in the way of simple tasks: if you want to generate a story about a doctor who is a woman and Ashanti, the race-and-gender scrambler will often cause the LLM to "lose track" of characteristics the user specifically asked for. This is directly downstream of trying to turn predictions on "doctor" away from "elderly white man with a kindly expression, wearing a white coat and stethoscope" sorts of defaults, which, to end where I started, aren't reality claims and do not carry moral weight.
Earlier quoted context omitted.
A fellow Code Report viewer, I assume?
who? never heard of it. Prof Shannon said this a while ago, in the '50s I believe.
Earlier quoted context omitted.
So.. "open" means "open at first, then not so much or not at all as we get closer to achieving AGI"? As they become more successful, they (obviously) have a lot of motivation to not be "open" at all, and that's without even considering the so-called ethical arguments. More generally, putting "open" in any name frequently ends up as a cheap marketing gimmick. If you end up going nowhere it doesn't matter, and if you'r…
OpenAI is literally not a word in the dictionary. It’s a made up word. So the Open in OpenAI means whatever OpenAI wants it to mean. It’s a trademarked word. The fact that Elon is suing them for their name when the guy has a feature “AutoPilot” which is not a made up word and had an actual well understood meaning which totally does not apply to how Tesla uses AutoPilot is hilarious.
Other than the vehicle, this would seem to apply to Tesla's autopilot as well. The "Full Self Driving" claim is the absurd one, odd that you didn't choose that example.
Earlier quoted context omitted.
I think he disagrees with 4: 4. Language prediction training will not get stuck in a local optimum. Most previous things we train on could have been better served if the model developed AGI, but they didn't. There is no reason to expect LLMs to not get stuck in a local optimum as well, and I have seen no good argument as to why they wouldn't get stuck like everything else we tried.
There is very little in terms of rigorous mathematics on the theoretical side of this. All we have are empirics, but everything we have seen so far points to the fact that more compute equals more capabilities. That's what they are referring to in the blog post. This is particularly true for the current generation of models, but if you look at the whole history of modern computing, the law roughly holds up over the l…
Earlier quoted context omitted.
It's actually a very clever structure! Please open the following image to follow along as I elaborate: https://pbs.twimg.com/media/F_PGOPOacAApU8e.jpg At the top level, there is the non-profit board of directors (i.e.: the ones Sam Altman had that big fight with). They are not beholden to any shareholders and are directly bound by the company charter: https://openai.com/charter The top-level nonprofit company owns ho…
Except Altman has the political capital to have the entire board fired if they go against him, which makes the entire structure irrelevant. The power is where the technology is being developed -- at the bottom where they can threaten to walk out with plush jobs from the major shareholders at the bottom. The power is not where the figureheads sit at the top.
Out of curiousity, with the benefit of hindsight; what would you have tried doing differently to prevent such a coup?
Earlier quoted context omitted.
What they said there isn't their mission, that is their hidden agenda. Here is their real mission that they launched with, they completely betrayed this: > As a non-profit, our aim is to build value for everyone rather than shareholders. Researchers will be strongly encouraged to publish their work, whether as papers, blog posts, or code, and our patents (if any) will be shared with the world https://openai.com/blog/…
“Dont be evil” ring any bells?