4 ms·

At this point, the microsoft CEO [1] is de facto right and even a bit minimalist in his stance: If it is on the internet or connected to the internet, it's up

At this point, the microsoft CEO [1] is de facto right and even a bit minimalist in his stance: If it is on the internet or connected to the internet, it's up for grabs by any reasonably powerfull organization. They will use it to train AI or derive some kind of privacy violating model from it. They might even use your cpu power and electricity for it. I'd love the governement to take a more powerfull stance against it, but they clearly aren't interested. If small you and me want some privacy, you have to shield it from the internet, govs and bigcos: No more Microsoft, Google, Adobe, .... products. [1] https://www.androidauthority.com/microsoft-ai-ceo-interview-3455794/ https://www.androidauthority.com/microsoft-ai-ceo-interview-...

try_the_bass 2y ago
[flagged]
- mpweiher 2y ago
  > They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? Er...no? As in "not even close"? The maker of the typewriter Hemingway used has no rights to the works of Hemingway. The maker of the shovel has no rights to the ditch dug with it. The maker of my kitchen knives has no rights to the food I prepare with them.
  - try_the_bass 2y ago
    But a typewriter isn't a product that evolves over time in the way software does? If you want a newer, better typewriter, you have to find and buy it. The typewriter you have isn't going to get periodic updates that improve it's functionality. And the company that made the typewriter isn't paying for the hardware every time you use it, unlike with a SaaS product (in most cases). You're also not paying a periodic fee to continue using said typewriter. They're fundamentally different ownership/usage models, so why wouldn't the data ownership model change to accommodate? [E] It would be helpful if you replied to my whole comment, rather than just the first line. I did clarify what I was saying further on into the comment.
    - jononor 2y ago
      Desktop OS like Windows one would also need to buy new versions of to get improvements. Nor did it require a periodic fee. But most importantly, the AI trying by default is changing the nature of the purchase/usage agreement, without consent of the buyer.
      - try_the_bass 2y ago
        Well for one thing, this thread was about Figma, not Windows. For another, how is changing to opt-in by default "without the consent of the buyer"? In this situation, Figma is opting in folks who have signed up for a waitlist to use the AI features. Those users have opted in to a new feature, and that new feature comes with its own set of opt-in preferences. They have clearly communicated that this doesn't take effect immediately, and are giving users the ability to opt out with plenty of advance warning. Are more philosophically: if a product is allowed to change over time, why is the purchase/usage agreement not allowed to also change over time? It seems irrational to demand that one thing change (the SaaS product should improve over time!) while simultaneously demanding the purchase/usage agreement not change.
        jononor 2y ago
        According to the linked documentation, Figma is opt-in AI features for all plans, and AI training on all plans except for Organization/Enterprise. Regardless of whether people have indicated interest in such features or not.
        try_the_bass 2y ago
        Also according to their documentation, all the AI features are in limited beta (hence the wait-list I mentioned earlier). It isn't clear to me on re-reading the documentation if everyone is opted into training the AI, rather than just those who have opted into using the AI features while they are in beta.
- StrLght 2y ago
  > They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? I am sorry, what? Should every SaaS product have the same rights to customers data as customers themselves?
  - try_the_bass 2y ago
    I didn't clearly state it, but I don't mean "the same", strictly speaking. But they do have some right to it, and what rights they have are up to the both of you to decide. If Google Docs said in their Terms of Use that they wanted full rights to every document created in the product, I assume very few people would use Google Docs. But if they reserve the right to train an LLM on the data to create better suggestions in various contexts, I assume many people would be fine with that, and will continue to use the product. Ultimately it comes down to a) what the company wants, and b) what people are willing to give. None of these commercial products we're taking about here are monopolies, so users always have a choice whether or not to agree to these terms. There's nothing inherently wrong with that setup?
- joelanman 2y ago
  so if you wrote a book in Google docs, Google should have just as much right to it as you? Or Microsoft should have just as much right to the app you write in VS Code?
  - try_the_bass 2y ago
    The sibling reply here gets at the point I was trying to make. Immediately jumping to the extreme interpretation is not exactly in good faith. There exists a middle ground in which you do not have full rights to your data generated while using SaaS products, but neither does the company who makes the SaaS product. This is why Terms of Use and Privacy Policy documents exist: to outline what rights the company wants to your data. If you don't agree to them, don't use the product. If you cannot find a product that does what you want and has a data ownership model you agree with, and you think many other people agree with you, then you have identified a market for a product. If you are unwilling or unable to build said product, perhaps your demands are unreasonable? I find it to be a childish argument to demand full control and ownership over all your data, including usage data. Like I pointed out in my original post, most people who want to control all of their data do nothing with it. They willfully destroy valuable information rather than share it with those who could use it. Sure, you might not like what they're using it for, at which point I'd be willing to agree that's a reasonable stance to take. Clearly not all data should be shared with all people. But this thread is about Figma wanting to collect data to train an AI so they can improve their product and make it easier to use. Why would you use Figma, but disagree that they should improve said product? I would find it hard to believe that you don't actually want the product to improve, which essentially boils the argument down to "I want you to make the product better, but I don't want to help you do it in any way. I don't want you to look at how I use the product to find out how it could be improved, either." That just seems wildly self-centered and antisocial.
  - rented_mule 2y ago
    Saying Google has the same rights (including publication, etc.) as you do to your book written in Google Docs is certainly over the line. I can't speak for the grandparent, but I have to assume they didn't have this extreme in mind when writing their comment. The far more interesting question is, "Where is the line?" Is it okay for Google Docs to store your documents on Google's drives? If you want that product to exist, then it better be okay! What if Google analyzes font usage in all docs? No docs out of millions have ever used SillyFont, so they decide to remove SillyFont from the default list to declutter the interface. I don't think that's over the line. How about detecting that everyone is ignoring/undoing the spelling correction of a specific word and using that to realize they have a misspelling in their dictionary? I'd like that. How about noticing that your documents all use British spellings rather than American spellings (e.g., "colour" rather than "color"), so they adapt spell checking and autocompletion? Is that over the line? How about using everyone's text and location to build an ML classifier to determine that "go on holiday" is British and "go on a vacation" is American and using that to adapt autocompletion? What if 99% of all occurrences of the phrase "It was a dark and" end with "stormy night", so they autocomplete that? If you're writing your new book in Google Docs and I type the first five words, should it autocomplete the rest of your book for me? I think it's easy to agree this is over the line. Where is the line? I suspect we won't all agree. Is it possible to come up with somewhat standardized concepts and language around this, so that vendors can publish data usage policies that let consumers make informed decisions?
- GeoAtreides 2y ago
  > They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? Wait till Fendocaster and Canon (among many others) hear about this, they'll be ecstatic
  - try_the_bass 2y ago
    A different product ownership model seems like it would dictate a different data ownership model. Last I checked, buying a Canon is a one-time purchase, you don't get periodic updates to the feature set, and Canon isn't paying some small amount of money every time you use their product. With SaaS products these attributes are all different, and it seems fine that there also be a different data ownership model. Otherwise you (as a customer) are just trying to have your cake (unrestricted usage of the product) and eat it too (full ownership of all the data generated, even usage data). That hardly seems like a fair trade?
scosman 2y ago
It is possible to learn from private data while preserving privacy. Intro here (there are many other forms): https://machinelearning.apple.com/research/learning-with-privacy-at-scale https://machinelearning.apple.com/research/learning-with-pri... Lots of issues come with privatized ML though: - It's pretty close to impossible for a consumer to judge if the methods used are actually privacy preserving, or just lip service. It's just too technical. - It's much harder to implement than non-private learning. - Governments will likely not be able to regulate at the level of technical detail needed to allow privacy preserving, and not the non-private learning - You complain about using the consumers CPU/electricity, but that's often very helpful for privacy. The private alternative is taking DP data off the device, in which need to collect a lot more data for same privacy levels.

try_the_bass 2y ago

[flagged]

mpweiher 2y ago
> They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? Er...no? As in "not even close"? The maker of the typewriter Hemingway used has no rights to the works of Hemingway. The maker of the shovel has no rights to the ditch dug with it. The maker of my kitchen knives has no rights to the food I prepare with them.
- try_the_bass 2y ago
  But a typewriter isn't a product that evolves over time in the way software does? If you want a newer, better typewriter, you have to find and buy it. The typewriter you have isn't going to get periodic updates that improve it's functionality. And the company that made the typewriter isn't paying for the hardware every time you use it, unlike with a SaaS product (in most cases). You're also not paying a periodic fee to continue using said typewriter. They're fundamentally different ownership/usage models, so why wouldn't the data ownership model change to accommodate? [E] It would be helpful if you replied to my whole comment, rather than just the first line. I did clarify what I was saying further on into the comment.
  - jononor 2y ago
    Desktop OS like Windows one would also need to buy new versions of to get improvements. Nor did it require a periodic fee. But most importantly, the AI trying by default is changing the nature of the purchase/usage agreement, without consent of the buyer.
    - try_the_bass 2y ago
      Well for one thing, this thread was about Figma, not Windows. For another, how is changing to opt-in by default "without the consent of the buyer"? In this situation, Figma is opting in folks who have signed up for a waitlist to use the AI features. Those users have opted in to a new feature, and that new feature comes with its own set of opt-in preferences. They have clearly communicated that this doesn't take effect immediately, and are giving users the ability to opt out with plenty of advance warning. Are more philosophically: if a product is allowed to change over time, why is the purchase/usage agreement not allowed to also change over time? It seems irrational to demand that one thing change (the SaaS product should improve over time!) while simultaneously demanding the purchase/usage agreement not change.
      - jononor 2y ago
        According to the linked documentation, Figma is opt-in AI features for all plans, and AI training on all plans except for Organization/Enterprise. Regardless of whether people have indicated interest in such features or not.
        try_the_bass 2y ago
        Also according to their documentation, all the AI features are in limited beta (hence the wait-list I mentioned earlier). It isn't clear to me on re-reading the documentation if everyone is opted into training the AI, rather than just those who have opted into using the AI features while they are in beta.
StrLght 2y ago
> They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? I am sorry, what? Should every SaaS product have the same rights to customers data as customers themselves?
- try_the_bass 2y ago
  I didn't clearly state it, but I don't mean "the same", strictly speaking. But they do have some right to it, and what rights they have are up to the both of you to decide. If Google Docs said in their Terms of Use that they wanted full rights to every document created in the product, I assume very few people would use Google Docs. But if they reserve the right to train an LLM on the data to create better suggestions in various contexts, I assume many people would be fine with that, and will continue to use the product. Ultimately it comes down to a) what the company wants, and b) what people are willing to give. None of these commercial products we're taking about here are monopolies, so users always have a choice whether or not to agree to these terms. There's nothing inherently wrong with that setup?
joelanman 2y ago
so if you wrote a book in Google docs, Google should have just as much right to it as you? Or Microsoft should have just as much right to the app you write in VS Code?
- try_the_bass 2y ago
  The sibling reply here gets at the point I was trying to make. Immediately jumping to the extreme interpretation is not exactly in good faith. There exists a middle ground in which you do not have full rights to your data generated while using SaaS products, but neither does the company who makes the SaaS product. This is why Terms of Use and Privacy Policy documents exist: to outline what rights the company wants to your data. If you don't agree to them, don't use the product. If you cannot find a product that does what you want and has a data ownership model you agree with, and you think many other people agree with you, then you have identified a market for a product. If you are unwilling or unable to build said product, perhaps your demands are unreasonable? I find it to be a childish argument to demand full control and ownership over all your data, including usage data. Like I pointed out in my original post, most people who want to control all of their data do nothing with it. They willfully destroy valuable information rather than share it with those who could use it. Sure, you might not like what they're using it for, at which point I'd be willing to agree that's a reasonable stance to take. Clearly not all data should be shared with all people. But this thread is about Figma wanting to collect data to train an AI so they can improve their product and make it easier to use. Why would you use Figma, but disagree that they should improve said product? I would find it hard to believe that you don't actually want the product to improve, which essentially boils the argument down to "I want you to make the product better, but I don't want to help you do it in any way. I don't want you to look at how I use the product to find out how it could be improved, either." That just seems wildly self-centered and antisocial.
- rented_mule 2y ago
  Saying Google has the same rights (including publication, etc.) as you do to your book written in Google Docs is certainly over the line. I can't speak for the grandparent, but I have to assume they didn't have this extreme in mind when writing their comment. The far more interesting question is, "Where is the line?" Is it okay for Google Docs to store your documents on Google's drives? If you want that product to exist, then it better be okay! What if Google analyzes font usage in all docs? No docs out of millions have ever used SillyFont, so they decide to remove SillyFont from the default list to declutter the interface. I don't think that's over the line. How about detecting that everyone is ignoring/undoing the spelling correction of a specific word and using that to realize they have a misspelling in their dictionary? I'd like that. How about noticing that your documents all use British spellings rather than American spellings (e.g., "colour" rather than "color"), so they adapt spell checking and autocompletion? Is that over the line? How about using everyone's text and location to build an ML classifier to determine that "go on holiday" is British and "go on a vacation" is American and using that to adapt autocompletion? What if 99% of all occurrences of the phrase "It was a dark and" end with "stormy night", so they autocomplete that? If you're writing your new book in Google Docs and I type the first five words, should it autocomplete the rest of your book for me? I think it's easy to agree this is over the line. Where is the line? I suspect we won't all agree. Is it possible to come up with somewhat standardized concepts and language around this, so that vendors can publish data usage policies that let consumers make informed decisions?
GeoAtreides 2y ago
> They're training on data that you generated while using their product. It seems like they should have just as much right to it as you? Wait till Fendocaster and Canon (among many others) hear about this, they'll be ecstatic
- try_the_bass 2y ago
  A different product ownership model seems like it would dictate a different data ownership model. Last I checked, buying a Canon is a one-time purchase, you don't get periodic updates to the feature set, and Canon isn't paying some small amount of money every time you use their product. With SaaS products these attributes are all different, and it seems fine that there also be a different data ownership model. Otherwise you (as a customer) are just trying to have your cake (unrestricted usage of the product) and eat it too (full ownership of all the data generated, even usage data). That hardly seems like a fair trade?

scosman 2y ago

It is possible to learn from private data while preserving privacy. Intro here (there are many other forms): https://machinelearning.apple.com/research/learning-with-privacy-at-scale https://machinelearning.apple.com/research/learning-with-pri... Lots of issues come with privatized ML though: - It's pretty close to impossible for a consumer to judge if the methods used are actually privacy preserving, or just lip service. It's just too technical. - It's much harder to implement than non-private learning. - Governments will likely not be able to regulate at the level of technical detail needed to allow privacy preserving, and not the non-private learning - You complain about using the consumers CPU/electricity, but that's often very helpful for privacy. The private alternative is taking DP data off the device, in which need to collect a lot more data for same privacy levels.