3 ms·

There are a couple of different motivations a company could have around blocking API access to prevent AI scraping: A) scraping itself is too expensive. I susp

There are a couple of different motivations a company could have around blocking API access to prevent AI scraping: A) scraping itself is too expensive. I suspect that's probably not the case with SO because they blocked backup. Downloading the database from the Internet Archive doesn't cost SO any money. B) the AI is going to replace the original creators (or more likely, devalue their work and push wages lower) and they'd like to prevent that negative social consequence. This is the charitable interpretation, and I understand writers/programmers/artists being concerned about it, even if I'm slightly more cynical myself about how AI content generation is going to work out once the "shine" has worn off. Note that I'm not saying that this concern is necessarily right or that there aren't positive uses of AI that have nothing to do with replacing jobs; just that it's a concern that a site/community could reasonably have. C) companies are realizing that there's a lot of VC money in AI right now, and they would very much like to be in the business of selling shovels, and their feeling is that if anyone anywhere is making money off of "their" content then they are morally deserving of some kind of cut no matter what. This is obviously the case for some companies, but is (charitably) probably not the case for all of them. One test we could use to try and distinguish between B and C is -- if a company is blocking API access, are they then turning around and licensing that data or opening up paid API access, and if they are, is any of that money going to the users that made the content? If SO turns around and makes API access paid and continues to not pay any of the volunteers writing answers, at that point it's much easier to argue that they're trying to sell shovels, not trying to protect users. This is also part of why I take a cynical view of what Reddit is doing with its API (although Reddit claims they're in camp A more than B). Reddit is probably not doing this to protect its users from theoretical AI displacement because it's still planning to license the data. It's just pricing it so high that only giant companies would be able to afford it.

shagie 3y ago
On the B/C test: From Jody Bailey (SO staff) https://meta.stackexchange.com/a/390040 https://meta.stackexchange.com/a/390040 (in full) posted 7 minutes ago (as I write this) > Stack Overflow senior leadership is working on a strategy to protect Stack Overflow data from being misused by companies building LLMs. While working on this strategy, we decided to stop the dump until we could put guardrails in place. > We are working on setting up the infrastructure to do this correctly in the age of LLMs --- where we continue to be open and share the data with our developer community but work to set up a formal framework for large AI companies that want to leverage the data. > We are looking for ways to gate access to the Dump, APIs, and SEDE, that will allow individuals access to the data while preventing misuse by organizations looking to profit from the work of our community. We are working to design and implement appropriate safeguards and still sorting out the details and timelines. We will provide regular updates on our progress to this group. --- On reddit, for the A/C test... yea, I'm going to be cynical there that they're looking to sell the information (and it isn't so much trying to protect users). But also that 3rd party clients are not showing ads and may be poorly behaved when provided a free API with (what at one time) was generous rate limiting.
- danShumway 3y ago
  I looked into this a bit more and I'm somewhat doubtful of this justification. There's at least been discussion within SO about charging access: https://meta.stackexchange.com/questions/388551/is-se-going-to-be-selling-our-content-for-ai-model-training-and-what-exactly/ https://meta.stackexchange.com/questions/388551/is-se-going-... > First, I'd like to say that the intent of what Prashanth is saying is very simple: to return value to the community for the work that you have put in. The money that we raise from charging these huge companies that have billions of dollars on their balance sheet will be used for projects that directly benefit the community. This is worded very specifically. Is SO planning to give money to users? They don't say anything like that; instead they say that they'll be "spending that money on the platform." Well what does that actually mean? Every feature that SO builds could be characterized as "for the benefit of the community." It's hard not to read that response as just another way of saying "we're going to profit from this as a company, but don't worry because we use our profits to fund product development." Heck, Reddit could make exactly the same claim, and in fact the linked Wired article actually makes that comparison: > "Community platforms that fuel LLMs absolutely should be compensated for their contributions so that companies like us can reinvest back into our communities to continue to make them thrive," Stack Overflow’s Chandrasekar says. "We're very supportive of Reddit’s approach."

shagie 3y ago

On the B/C test: From Jody Bailey (SO staff) https://meta.stackexchange.com/a/390040 https://meta.stackexchange.com/a/390040 (in full) posted 7 minutes ago (as I write this) > Stack Overflow senior leadership is working on a strategy to protect Stack Overflow data from being misused by companies building LLMs. While working on this strategy, we decided to stop the dump until we could put guardrails in place. > We are working on setting up the infrastructure to do this correctly in the age of LLMs --- where we continue to be open and share the data with our developer community but work to set up a formal framework for large AI companies that want to leverage the data. > We are looking for ways to gate access to the Dump, APIs, and SEDE, that will allow individuals access to the data while preventing misuse by organizations looking to profit from the work of our community. We are working to design and implement appropriate safeguards and still sorting out the details and timelines. We will provide regular updates on our progress to this group. --- On reddit, for the A/C test... yea, I'm going to be cynical there that they're looking to sell the information (and it isn't so much trying to protect users). But also that 3rd party clients are not showing ads and may be poorly behaved when provided a free API with (what at one time) was generous rate limiting.

danShumway 3y ago
I looked into this a bit more and I'm somewhat doubtful of this justification. There's at least been discussion within SO about charging access: https://meta.stackexchange.com/questions/388551/is-se-going-to-be-selling-our-content-for-ai-model-training-and-what-exactly/ https://meta.stackexchange.com/questions/388551/is-se-going-... > First, I'd like to say that the intent of what Prashanth is saying is very simple: to return value to the community for the work that you have put in. The money that we raise from charging these huge companies that have billions of dollars on their balance sheet will be used for projects that directly benefit the community. This is worded very specifically. Is SO planning to give money to users? They don't say anything like that; instead they say that they'll be "spending that money on the platform." Well what does that actually mean? Every feature that SO builds could be characterized as "for the benefit of the community." It's hard not to read that response as just another way of saying "we're going to profit from this as a company, but don't worry because we use our profits to fund product development." Heck, Reddit could make exactly the same claim, and in fact the linked Wired article actually makes that comparison: > "Community platforms that fuel LLMs absolutely should be compensated for their contributions so that companies like us can reinvest back into our communities to continue to make them thrive," Stack Overflow’s Chandrasekar says. "We're very supportive of Reddit’s approach."