Hacker News Reader

Comment by wolverine876 | Hacker News Reader

The value of data declines over time, often rapidly. Think about what a business would pay for current user data, data one year old, data five years old, etc.; who buys five year old data? I suppose it depends on the application. Imagine how your life changes over time, an accumulation of small and large changes. Data on your health, finances, social network, interests, politics, activities, etc. was different 1 year ago, 5 years ago, etc.

The question seems like a justification for what many want to do, which is just close their eyes to the problems and not put the energy into resisting the tide toward compliance and giving power to government/corporations; it's easier, right now, to give up, and despair and powerlessness are normalized - instead of quitting being at least a bit shameful, it's cynically embraced, flauted as rebellious.

View on HN

You are just wrong. Some of your data such as health data stays constant throughout your life. Especially for guys living with chronic conditions. Your key identifiers stay the same too. E.g your phone number etc. These are then used to re-establish your advertising identity. This theory that old data declines in value couldn't be more wrong given that the "old" data is never deleted. Just waiting to find its original purpose.

Yes, but the value of the data has declined. If I am a business I would not pay the same amount or more for five-year-old data. Some portion of that dataset will be bad data. Some of the subjects will have had died, for example.

Decline does not mean zero, but the point is older data is less valuable. So, making your data leaking footprint smaller today will still provide you value (not infinite value though) as time goes on.

Sure, the more recent data will always have a higher value if you are looking at massive datasets.

But I think if you look at this problem from an individual perspective. Key data that is unique to them doesn't change most of the time. Say a target person has dental issues and you have their phone number. There is a high chance that in 5 years from now, they still may require dental services and their phone number hasn't changed. They may have moved jobs, neighborhoods etc. in between but their need for dental service has remained constant and probably will into the future.

Let me give you another example. Say you want to run for public office, and in your young adult life some of your social media pages had questionable posts/media. Is this data any less valuable to your competition running for the same office?

I may be taking a different approach with this but a lot of nuances get lost in the numbers.

My point is once the data is out there on the internet, it can be hard to control it and/or assign value to it.

That cherry-picks examples, phone numbers and chronic conditions. We can think of plenty more that change. (Even phone numbers change sometimes.)

Sure. But what if those two data points are the only things I need to tag and target you.

Cherry picked examples are proof that the OP is wrong, at least to an extent, and that there exists data that doesn't change or changes slow enough to be re-tagged (and this poses the issue in question).

Proof of existence is enough to invalidate a claim of non-existence, even if it doesn't hold for any condition in the world.

No comments were made in the context of logic, so the logical forms don't apply. Personally I'm interested in the long-term value of data and not a logic contest - I'm interested in the territory, not a particular map of it.

You can change your phone number though. What if you change it so frequently that it's no longer a key identifier? Why not have a system where you can receive phone calls without a phone number? I believe some of the solution here is to stop having globally unique identifiers.

> Why not have a system where you can receive phone calls without a phone number?

How would you do that?

Burner numbers are close, but if you use the same one for all your contacts it still becomes an identifier, even if you rotate it. An alternative is a proxy service where you get issued a unique pin for each person-to-person connection, so the same phone number gets used by multiple parties. Those both use the existing phone network in better ways. With a VoIP system you don't need numbers, you just don't want to trade them for usernames or some other global identifier. I implemented one approach with Severus, but I'm also looking for alternatives that increase privacy.

What if you go even further and pollute the data to such an extent that you start fitting into all buckets.

You could be male, female and any other gender, any race, any job, work as a CEO and a janitor, live both in Russia and the US.

You could probably do this to Google search.

Edit: Apparently somebody already built an extension which does exactly this for Facebook. https://floorvandenberghe.be/derivedscrolling/

Interesting thought. I wonder if the value of the time in a person's life spent without having their data tracked, sold and used to influence their thought and decisions doesn't increase over time like an investment. So let's say they were not under heavy surveillance and sway from such a system for one year 5 years ago... Could that year away have given them a base formation of better adjusted mental health habits that helps them more today because it happened earlier in life than it would have if they had been free of it for only the past year, after 4 years more of continued privacy and attention degradation? I bet this is especially poignant when neuroplasticity is considered.

Optimistically, this would mean any resistance made today hurts the state-corp alliance hard now, and benefits the human being especially hard later, giving courage to make efforts for the long game.

Reply on news.ycombinator.com