Hacker News Reader

Comment by rkagerer | Hacker News Reader

This highlights a significance today's cloud-addicted generation seems to completely ignore: who has control of your data.

I'm not talking about contractual control (which is largely mooted as pretty much every cloud service has a ToS that's grossly skewed toward their own interests over yours, with clauses like indemnifications, blanket grants to share your data with "partners" without specifying who they are or precisely what details are conveyed, mandatory arbitration, and all kinds of other exceptions to what you'd consider respectful decency), but rather where your data lives and is processed.

If you truly want to maintain confidence it'll remain private, don't send it to the cloud in the first place.

View on HN

Framing this as a moralized issue of “addiction” on the part of consumers naturalizes the structural cause of the problem. Concentrated wealth benefits from the cloud capital consumers generate. It’s this group that has most of the control over how our data is collected. These companies reduce the number and quality of choices available to us. Blaming consumers for choosing the many conveniences of cloud data when the incentive structure has been carefully tailored to funnel our data into their possession and control is truly a superficial take.

Well put. And generalizes to most consumer-blaming.

And this kind of consumer-blaming ultimately serves the interests of the very people who benefit most from the status quo, but shifting attention away from them and toward people who are easy to pick on but ultimately have very little control over the situation. For most people, opting out of the cloud is tantamount to opting out of modern society.

I can't even get important announcements from my kids' school without signing up for yet another cloud service.

> Blaming consumers for choosing the many conveniences of cloud data when the incentive structure has been carefully tailored to funnel our data into their possession and control is truly a superficial take.

Couldn't have said it better.

Just consider Apple as an example: Some time ago, they used to sell the Time Capsule, a local-first NAS built for wireless Time Machine backups. Today, not only has the Time Capsule been discontinued, but it's outright impossible to make local backups of iOS devices (many people's primary computing devices!) without a computer and a USB cable.

Considering Apple's resources, it would take negligible effort to add a NAS backup feature to iOS with polished UX ("simply tap your phone on your Time Capsule 2.0 to pair it, P2P Wi-Fi faster than your old USB cable" etc.) – but they won't, because now it's all about "services": Why sell a Gigabyte once if you can lease it out and collect rent for it every month instead?

Yeah, I see a ton of people all up in arms about privacy but ignoring that OpenAI doesn't give a rats ass about others privacy (see: scraping).

Like why one good other bad?

If something is able to be scraped, it isn't private.

There is no technical reason for chats people have with ChatGPT or any similar service to be available on the web to everyone, so there is no way for them to be scraped.

It’s not zero sum. I can believe that openai does not take privacy seriously enough and also that I don’t want every chat I’ve ever had with their product to be entered into the public record.

“If one is good the other must be good” is far too simplistic thinking to apply to a situation like this.

I personally just can't fathom the logic that sending something so private and critical to OpenAI is ok, but to have courts view it is not? Like if it's so private, why in hell would you give it to a company that has shown that it cares not at all about others privacy?

Interesting. It seems obvious to me.

I’ve asked ChatGPT medical things that are private but not incriminating or anything, because I trust ChatGPT’s profit motive to just not care about my individual issues. But I would be pretty irritated if the government stepped in and mandated they make my searches public and linkable to me.

Are you perhaps taking an absolutist view where anything less than perfect attention to all privacy is the same as making all logs of everyone public?

> But I would be pretty irritated if the government stepped in and mandated they make my searches public and linkable to me.

Who is calling for this? Are you perhaps taking an absolutist view where "not destroying evidence" is the same as "mandated they make my searches public and linkable to me"? That's quite ridiculous.

Discovery routinely leaks. Handing over every chat from every user to opposing council has both human, technical, and incentive issues that make it far more likely that something I told ChatGPT with an understanding of its privacy limitations will appear in a torrent.

This seems like an unhelpful extension of the word "privacy". Scraping is something, but it is mostly not a privacy violation.

We have shifted over to SaaS so much for convenience that we have lost sight of “our control”.

I imagine a 90s era software industry for today’s tech world: person buys a server computer, person buys an internet connection with stable ip, person buys server software boxes to host content on the internet, person buys security suite to firewall access.

Where is the problem in this model? Aging computers? Duplicating computing hardware for little use? Unsustainable? Not green/efficient?

> who has control of your data

As frustrating as it is, the answer seems to be everyone and no one. Data in some respects is just an observation. If I walk through a park, and I see someone with red hair, I just collected some data about them. If I see them again, perhaps strike up a conversation, I learn more. In some sense, I own that data because I observed it.

On the other other hand, I think most decent people would agree that respecting each other's right to privacy is important. Should the owner of the red hair ask me to not share personal details about them, I would gladly accept, because I personally recognize them as the owner of the source data. I may possess an artifact or snapshot of that data, but it's their hair.

In a digital world where access controls exist, we have an opportunity to control the flow of our data through the public space. Unfortunately, a lot of work is still needed to make this a reality...if it's even possible. I like the Solid Project for it's attempt to rewrite the internet to put more control in the hands of the true data owners. But, I wonder if my observation metaphor is still possible even in a system like Solid.

It's not developer's common interest to develop local services first. People build cloud services for a profit so they can be paid. However, sometimes developers need to build their portfolios (or out of pure interest) so they make software that runs locally anyway. A lot of websites can easily be ran on people's computers from a data perspective but it's a lot of work to get that data in the first place and make it useable. I don't think people are truly "cloud-addicted". I think they simply do not have other choices.

the post does not reflect the reality that it is not 'your data'*. When you use a service provider, it's their data. They may give you certain rights to influence your usage or payment of their service, but if it's not on machines you control then it's not your data.

*my legal argument is "possession is 9/10ths of the law"

How do consumers utilize expensive compute resources in this model? Eg, H100 GPUs.

You don't have control over your data in the eyes of these guys... this was clear as soon as they started training their LLM on it without asking you

> If you truly want to maintain confidence it'll remain private, don't send it to the cloud in the first place.

I mean yes. but if you host it, then you'll be taken to court to hand that data over. Which means that you'll have less legal talent at your disposal to defend against it.

> but if you host it, then you'll be taken to court to hand that data over.

Not in this case. The Times seems to be claiming that OpenAI is infringing rather than any particular user. If one does not use OpenAI then their data is not subject to this.

Reply on news.ycombinator.com