7 ms·
>Altman took a non-profit and vacuumed up a bunch of donor money only to flip Open AI into the hottest TC style startup in the world. Then put a gas pedal to co
by LMYahooTFY 3y ago
>Altman took a non-profit and vacuumed up a bunch of donor money only to flip Open AI into the hottest TC style startup in the world. Then put a gas pedal to commercialization. It takes a certain type of politicking and deception to make something like that happen.
What exactly is the problem here? Is a non-profit expected to exclusively help impoverished communities or something? What type of politicking and deception is involved in creating a for profit subsidiary which is granted license to OpenAIs research in order to generate wealth? The entire purpose of this legal structure is to keep non-profit owners focused on their mission rather than shareholder value, which in this case is attempting to ethically create an AGI.
Edit: to add that this framework was not invented by Sam Altman, nor OpenAI.
>Then in the past week, he's going and taking money from the Saudis on the order of billions of dollars to make AI accelerators, even though the single greatest threat from strong AI (according to Hinton) is rich and powerful people using the technology to enhance their power over society.
Thus the legal structure I described, although this argument is entirely theoretical and assumes such a thing can actually be guarded that well at all, or that model performance and compute will remain correlated.
- jelling 3y ago> What exactly is the problem here? Is a non-profit expected to exclusively help impoverished communities or something? Yes. Yes and more yes. That is why, at least in the U.S., we have given non-profits exemptions from taxation. Because they are supposed to be improving society, not profiting from it.
- arrosenberg 3y ago> That is why, at least in the U.S., we have given non-profits exemptions from taxation. That's your belief. The NFL, Heritage Foundation and Scientology are all non-profits and none of them improve society; they all profit from it. (For what its' worth, I wish the law was more aligned with your worldview)
- aurareturn 3y agoFYI, the NFL teams are for profits and pay taxes like normal busineses. The overwhelming majority of the revenue goes to the teams.
- arrosenberg 3y agoI know that, does that change what I said?
- aurareturn 3y agoI don't know if it does but my point is to prevent others from thinking that a giant money making entity like the NFL does not pay any taxes.
- hluska 3y ago[flagged]
- johncampbelljr 3y agoWould you not object if someone characterized google as a non-profit because part of the org (the Google foundation) is non-profit? (Not a perfect analogy (nothing ever is, really).)
- twelvechairs 3y ago> OpenAIs 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. OpenAI believes that artificial intelligence technology has the potential to have a profound, positive impact on the world, so the companys goal is to develop and responsibly deploy safe AI technology, ensuring that its benefits are as widely and evenly distributed as possible. From their filing as a non-profit https://projects.propublica.org/nonprofits/organizations/810861541/202243199349314989/full https://projects.propublica.org/nonprofits/organizations/810...
- todd3834 3y agoI don’t think OpenAI ever reported to be profitable. They are allowed and should make money so they can stay alive. ChatGPT has already had a tremendous positive impact on society. The cause of safe AGI is going to take a lot of money in more research.
- shafyy 3y ago> ChatGPT has already had a tremendous positive impact on society. Citation needed
- todd3834 3y agoFair enough, I should have said, it’s my opinion that it has had a positive impact. I still think it’s easy to see them as a non profit. Even with everything they announced at AI day. Can anyone make an argument against it? Or just downvote because you don’t agree.
- sgt101 3y agoI think ChatGPT has created some harms: - It's been used unethically for psychological and medical purposes (with insufficient testing and insufficient consent, and possible psychological and physical harms). - It has been used to distort educational attainment and undermine the current basis of some credentials as a result. - It has been used to create synthetic content that has been released unmarked into the internet distorting and biasing future models trained on that content. - It has been used to support criminal activity (scams). - It has been used to create propaganda & fake news. - It has devalued and replaced the work of people who relied on that work for their incomes.
- vbo 3y agoand so has the internet. some use it for good, others for evil. these are behaviours and traits of the user, not the tool.
- 3y ago
- nmfisher 3y ago> Is a non-profit expected to exclusively help impoverished communities or something? What type of politicking and deception is involved in creating a for profit subsidiary which is granted license to OpenAIs research in order to generate wealth? OpenAI was literally founded on the promise of keeping AGI out of the hands of “big tech companies”. The first thing that Sam Altman did when he took over was give Microsoft the keys to the kingdom, and even more absurdly, he is now working for Microsoft on the same thing. That’s without even mentioning the creepy Worldcoin company. Money and status are the clear motivations here, OpenAI charter be damned.
- sumedh 3y ago> OpenAI was literally founded on the promise of keeping AGI out of the hands of “big tech companies”. Where does it say that?
- stavros 3y agoIn their charter: > We commit to use any influence we obtain over AGI’s deployment to ensure it is used for the benefit of all, and to avoid enabling uses of AI or AGI that harm humanity or unduly concentrate power.
- nottheengineer 3y agoRight here: https://openai.com/charter https://openai.com/charter
- ascv 3y agoIt seemed to me the entire point of the legal structure was to raise private capital. It's a lot easier to cut a check when you might get up to 100x your principal versus just a tax write off. This culminated in the MS deal: lots of money and lots of hardware to train their models.
- foota 3y agoWhat's confusing is that... open AI wouldn't ever be controlled by those that invested, and the owners (e.g., the board) aren't necessarily profit seeking. At least when you take a minority investment in a normal startup you are generally assuming that the owners are in it to have a successful business. It's just a little weird all around to me.
- quickthrower2 3y agoMicrosoft get to act as a sole distributor for the enterprise. That is quite valuable. Plus they are still in at the poker table and a few raises from winning the pot (maybe they just did!) but even without this chaos they are likely setting themselves up to be the for-profit investor if it ever transitioned to that. For a small amount of money (for MS) they get a lot of upside.
- xinayder 3y agoI like to read that, besides the problems others have listed, OpenAI seems like it was built on top of the work of others, who were researching AI, and suddenly took all this "free work" from the contributors and sold it for a profit where the original contributors didn't even see a single dime from their work. To me it seems like it's the usual case of a company exploiting open source and profiting off others' contributions.
- saiya-jin 3y agoOr any other say pharma company using massively and constantly basic research done by universities worldwide from our tax money. And then you go to pharmacy and buy medicine that costed 50 cents to manufacture and distribute for 50 bucks. I don't like the whole idea neither, but various communism-style alternatives just don't work very well.
- tokai 3y agoPharma companies spend billions on financing public research. Hell the Novo Nordisk Foundation is be biggest charitable foundation in the world.
- sgt101 3y agoPersonally I don't think that the use of previous research is an issue, the fact is that the investment and expertise required to take that research and create GPT-4 were very significant and the en-devour was pretty risky. Very few people five years ago thought that very large models could be created that would be able to encode so much information or be able to retrieve it so well.