🏛️ AI's training data gets a copyright bill - citations, links and licences

Jason politics-market w=1.00 · n=62 · E111 (2023-01-13) · explicit_prediction · strength 3 ▶ 49:14 SUPPORT

“So, ChatGPT and all these services must use citations of where they got the original work. They must link to them and they must get permission. That's where this is all going to shake out.”

Chamath politics-market w=1.02 · n=96 · E111 (2023-01-13) · sentiment · strength 2 ▶ 47:42 SUPPORT · horizon 36mo

“Now, that doesn't mean that ChatGPT can't figure that out, but it's those kinds of problems that are going to be a little thorny in these next few years that have to really get figured out.”

Sacks politics-market w=0.97 · n=88 · E111 (2023-01-13) · sentiment · strength 2 ▶ 50:32 OPPOSE

“Well, forget about permission. I mean, you can't get a big enough data set if you have to get permission in advance, right?”

Jason politics-market w=1.00 · n=62 · E115 (2023-02-11) · explicit_prediction · strength 3 ▶ 1:27:58 SUPPORT

“They train their AI on Getty Images with the watermarks on them, and they've been busted, and they are dead to rights now, and they are going to pay a hundred million dollars or more to Getty Images for stealing their content and allowing it to be republished in a commercial setting.”

Chamath politics-market w=1.02 · n=96 · E115 (2023-02-11) · explicit_prediction · strength 3 ▶ 52:54 SUPPORT

“And I'm going to go and make sure that all of these publishers have a new kind of agreement that they signed up for, which is I'll do my best to help you monetize, you do your best by being exclusive to our AI agents, so you deprive other models of your content on your pages, because that will get litigated, and there is no way, just like, again, if you say, do not crawl, you're not allowed to crawl if you're Google or Microsoft researching. This is going to happen for these agents. It's unrealistic to expect that it won't.”

Sacks politics-market w=0.97 · n=88 · E115 (2023-02-11) · explicit_prediction · strength 3 ▶ 1:02:57 OPPOSE

“New technology waves that are this powerful don't get stymied by either chip costs or legal rights issues.”

Friedberg politics-market w=0.84 · n=74 · E115 (2023-02-11) · explicit_prediction · strength 3 ▶ 1:27:17 OPPOSE

“I will bet any amount against your premonition here, J Cal.”

Jason politics-market w=1.00 · n=62 · E125 (2023-04-21) · explicit_prediction · strength 3 ▶ 39:46 SUPPORT

“as you brought up robots.txt last week, there's going to need to be an AI.txt.”

Jason politics-market w=1.00 · n=62 · E128 (2023-05-12) · sentiment · strength 1 ▶ 10:36 SUPPORT

“Yeah, so the question is, where is he getting these quotes? And then why doesn't it give citations?”

Jason politics-market w=1.00 · n=62 · E130 (2023-05-26) · sentiment · strength 2 ▶ 1:06:32 SUPPORT

“They made specific note when they launched this that all of this is done with stock photography. They have the complete license to, so they can monetize this without getting sued like Microsoft is currently being sued.”

Jason politics-market w=1.00 · n=62 · E133 (2023-06-16) · explicit_prediction · strength 3 ▶ 46:27 SUPPORT

“You choose, do robots.txt, except we're going to call it AI.txt.”

Friedberg politics-market w=0.84 · n=74 · E133 (2023-06-16) · explicit_prediction · strength 2 ▶ 1:05:55 SUPPORT

“Ultimately, the company loses the value if that data, that content gets extracted and they can't monetize it or capture the value somehow.”

Chamath politics-market w=1.02 · n=96 · E133 (2023-06-16) · explicit_prediction · strength 2 ▶ 1:10:20 SUPPORT

“they'll get to the right answer, which is give them a healthy rev share. That's the future.”

Jason politics-market w=1.00 · n=62 · E157 (2023-12-16) · explicit_prediction · strength 3 ▶ 29:11 SUPPORT

“it is Getty's right to then make derivative products based on their images... that is no excuse to use an open web crawler to avoid getting a license from the original owner of that. Just because you can technically do it doesn't mean you're allowed to do it.”

Sacks politics-market w=0.97 · n=88 · E157 (2023-12-16) · explicit_prediction · strength 3 ▶ 28:32 OPPOSE

“I do think that AI models should be able to use the content that's available on the web under a fair use doctrine to train their models for the benefit of consumers. And I don't see a reason to try and tie that up in a bunch of copyright lawsuits.”

Friedberg politics-market w=0.84 · n=74 · E157 (2023-12-16) · explicit_prediction · strength 3 ▶ 26:09 OPPOSE

“it's unrealistic for you to frame this as validating or justifying the fact that these companies won't be able to access and utilize open data under fair use to train models... So what ChatGPT is doing is they're accessing content behind a paywall... So it's not a training data set that's now being unlocked”

Chamath politics-market w=1.02 · n=96 · E157 (2023-12-16) · sentiment · strength 3 ▶ 30:20 BRUSH_OFF

“Hey, Friedberg, can you shoot me in the face and let me know when this segment is over?... stop with this navel gazing nonsense. We're in inning one and nobody knows anything.”

Sacks politics-market w=0.97 · n=88 · E167 (2024-02-23) · sentiment · strength 1 ▶ 1:10:56 SUPPORT

“I think the better AI models are providing citations now and links. Perplexi actually does a really nice job with this.”

Jason politics-market w=1.00 · n=62 · E167 (2024-02-23) · explicit_prediction · strength 2 ▶ 1:13:32 SUPPORT

“I just think the authority at which these LLMs speak is ridiculous. They speak as if they are absolutely 100% certain that this is the crisp, perfect answer ... When it should present it with citations.”

Jason politics-market w=1.00 · n=62 · E168 (2024-03-01) · sentiment · strength 2 ▶ 46:02 SUPPORT

“You took the other side of this. But now that you see the market based solution starting to emerge, what do you think of this market based solution? You think it's got legs?”

Jason politics-market w=1.00 · n=62 · E170 (2024-03-15) · explicit_prediction · strength 2 ▶ 25:55 SUPPORT

“Well, you'd have to get a license to make a derivative work, as we've seen. ... They're in a lawsuit right now with the New York Times, and they failed to negotiate their license.”

Friedberg politics-market w=0.84 · n=74 · E170 (2024-03-15) · explicit_prediction · strength 2 ▶ 26:12 OPPOSE

“yeah, YouTube's on the Internet, and I feel like anything that's on the open Internet should be watchable by these models. I don't consider training models to be generation of derivative work. That's my position.”

Sacks politics-market w=0.97 · n=88 · E170 (2024-03-15) · explicit_prediction · strength 2 ▶ 26:59 OPPOSE

“I agree with Friedberg that fair use doctrine should be applied to that. I know that you have a different point of view on that, J. Calder, I don't know if we need to rehash that.”

Sacks politics-market w=0.97 · n=88 · E174 (2024-04-12) · explicit_prediction · strength 2 ▶ 47:21 OPPOSE

“I think it's too soon for this legislation. I mean, the first thing that needs to happen is we need to get the question of fair use arbitrated by the courts”

Friedberg politics-market w=0.84 · n=74 · E174 (2024-04-12) · explicit_prediction · strength 3 ▶ 50:37 OPPOSE

“the models don't actually hold the data that they're trained on. They develop a bunch of synthesized predictors... I'd rather focus on is there copyright infringement in the output? Does the output mimic an original copyrighted source? And if it does, then that's where the infringement should lie, not on the how did you do it”

Chamath politics-market w=1.02 · n=96 · E174 (2024-04-12) · explicit_prediction · strength 2 ▶ 56:18 OPPOSE

“The difference is if Google doesn't sue OpenAI, it means copyright is worthless”

Sam Altman (guest ×0.5) no politics-market track record · E178 (2024-05-10) · sentiment · strength 2 ▶ 36:38 SUPPORT

“So I think there's an opt in, opt out in that case, first of all, and then there's an economic model.”

Friedberg politics-market w=0.84 · n=74 · E178 (2024-05-10) · sentiment · strength 2 ▶ 36:56 OPPOSE

“What's the difference here and why are you making the case that perhaps artists should be uniquely paid? This is not a sampling situation. The AI is not outputting and it's not storing in the model the actual original song. It's learning structure.”

Jason politics-market w=1.00 · n=62 · E178 (2024-05-10) · explicit_prediction · strength 2 ▶ 41:25 SUPPORT

“The music industry is going to consider this opportunity to make Taylor Swift songs their opportunity. It's part of the four part fair use test is, you know, these who gets to capitalize on new innovations for existing art.”

Jason politics-market w=1.00 · n=62 · E179 (2024-05-17) · explicit_prediction · strength 2 ▶ 1:22:21 SUPPORT

“And I think that we're going to see a major lawsuit here. Those people who are in those boxes are going to look at the answer here and realize maybe they don't get the click through and that this answer was built on that. And now we're going to have to have a new framework. There's going to need to be, Sacks, a new company that clears this content so that Google can do answers like this.”

Sacks politics-market w=0.97 · n=88 · E179 (2024-05-17) · explicit_prediction · strength 2 ▶ 1:30:53 SUPPORT

“Well, look, this is the same conversation we've had two or three times where we're gonna need the courts to figure out what fair use is. And depending on what they come up with, it may be the case that Google has to cut them in by doing licensing deals.”

Sacks politics-market w=0.97 · n=88 · E180 (2024-05-24) · explicit_prediction · strength 2 ▶ 20:16 SUPPORT

“And by the way, I mean, because it is such an industry-leading company, I think we could end up with some very bad fair use precedents or laws because Scarlett Johansson is so sympathetic as a plaintiff compared to OpenAI. And unless they show us some discovery that proves that they really did hire the voice actor and all the rest of it, I mean, this could lead to some very bad precedents for the industry around fair use.”

Chamath politics-market w=1.02 · n=96 · E180 (2024-05-24) · sentiment · strength 2 ▶ 13:02 SUPPORT

“Using a computer to probabilistically copy something is still copying something.”

Jason politics-market w=1.00 · n=62 · E180 (2024-05-24) · explicit_prediction · strength 2 ▶ 21:34 SUPPORT

“What happens in these content cases is they get settled almost every single time. So, the case law doesn't get codified, they just get settled out of court. If you go look at all the Fair Use cases, they almost never go to the mat. And so, this one will just be settled. It'll just be a question at what price.”

Friedberg politics-market w=0.84 · n=74 · E180 (2024-05-24) · explicit_prediction · strength 2 ▶ 8:03 OPPOSE

“I think what we will see over time is that rather than have the ability to sue for their likeness, a lot of AI that is like a celebrity, the value of it will arise from the celebrity's endorsement, not actually using the celebrity's features.”

Reid Hoffman (guest ×0.5) no politics-market track record · E194 (2024-08-30) · sentiment · strength 2 ▶ 26:58 OPPOSE

“we can train these models, like training is like reading. And like reading things is, like when something's available to be read, and you've engaged in the right economic thing for reading it, I think that's a kind of a reasonable, fair use thing.”

Jason politics-market w=1.00 · n=62 · E194 (2024-08-30) · sentiment · strength 2 ▶ 25:56 SUPPORT

“It's pretty crystal clear that their content's behind a paywall. And that's how they make money.”

Jason politics-market w=1.00 · n=62 · E213 (2025-01-31) · positioning · strength 2 ▶ 1:10:12 SUPPORT

“I think it's for three years is the license and they just did this blanket license for every book. They didn't look at yourself. They didn't look at how desirable it was. It was just like a blanket deal. Everybody gets $2,500 bucks per book for three years. And I think I'm going to just do it just to support proper licensing so that people can start going down this path.”

Jason politics-market w=1.00 · n=62 · E215 (2025-02-15) · explicit_prediction · strength 3 ▶ 1:24:40 SUPPORT

“I have a prediction here. I think this is all going to wind up like the Napster Spotify case. ... I think that there is a non-zero chance, like it might be 5 or 10 percent, that OpenAI is going to lose the New York Times lawsuit and they're going to lose it hard and there could be injunctions. And I think the settlement might be that these language models, especially the closed ones, are going to have to pay some percentage in a negotiated settlement of their revenue, half, two-thirds, to the content holders. And this could make the content industry have a massive, massive uplift and a massive resurgence.”

Naval Ravikant (guest ×0.5) no politics-market track record · E215 (2025-02-15) · explicit_prediction · strength 2 ▶ 1:24:20 SUPPORT

“I'm going to go out even thinner limb and say I largely agree with you. I think it's a bit rich to crawl the open web, hoover up all the data, offer direct substitution for a lot of use cases, because now you start and end with the AI model. It's not even like you link out like Google did. And then you just close off the models for safety reasons. I think if you trained on the open web, your model should be open source.”

Chamath politics-market w=1.02 · n=96 · E215 (2025-02-15) · sentiment · strength 2 ▶ 1:27:20 SUPPORT

“This is lossy compression. It is absolutely... I'm now on your page. I hate to say this, Jason. I agree with you. You were right.”

Friedberg politics-market w=0.84 · n=74 · E215 (2025-02-15) · explicit_prediction · strength 2 ▶ 1:29:08 OPPOSE

“I philosophically disagree with all of your nonsense on this topic. All three of you are wrong. And I'll tell you why. Number one, if information is out in the open internet, I believe it's accessible and it's viewable. And I view an LLM or a web crawler as basically being a human that's reading and can store information in its brain.”

Sacks politics-market w=0.97 · n=88 · E233 (2025-06-28) · explicit_prediction · strength 3 ▶ 1:35:32 OPPOSE POLICY-ADJACENT

“if all they're doing is transforming the work, they're doing positional encoding, and then coming up with their own unique work product, this judge said that that is not a violation of copyright. I completely agree with that.”

Jason politics-market w=1.00 · n=62 · E233 (2025-06-28) · explicit_prediction · strength 3 ▶ 1:39:20 SUPPORT

“And then Chachipiti and the New York Times should do a deal. And I predict they will do this, or with another language model that's smart enough to do it”

Chamath politics-market w=1.02 · n=96 · E233 (2025-06-28) · positioning · strength 2 ▶ 1:44:56 SUPPORT

“I told you about this. We did a licensing deal with OpenAI, we meaning Beast. They paid us, and they're using it for pre-training. It's quite reasonable.”

Jason politics-market w=1.00 · n=62 · E237 (2025-08-01) · explicit_prediction · strength 3 ▶ 43:43 SUPPORT

“I will guarantee that this will be the beginning of the deals and the deals will go up from here.”

Chamath politics-market w=1.02 · n=96 · E237 (2025-08-01) · explicit_prediction · strength 3 ▶ 42:13 OPPOSE · horizon 60mo

“I read that and I thought, this is the peak of these deals. These deals will only go down in terms of dollar value from here. And it actually brought me to this point where I was thinking to myself, is it even realistic to believe that patents and copyrights actually exist in five years?”

Friedberg politics-market w=0.84 · n=74 · E237 (2025-08-01) · explicit_prediction · strength 3 ▶ 35:49 OPPOSE

“I strongly disagree with the idea that AI getting trained is the same as AI replicating copyright material.”

Sacks politics-market w=0.97 · n=88 · E237 (2025-08-01) · explicit_prediction · strength 2 ▶ 40:12 OPPOSE POLICY-ADJACENT

“you just can't have a situation where China can train on the entire internet, and our AI models are hamstrung by needing to contract, negotiate contracts with every single website.”