Any models trained on the ill-gotten data should now be public domain.

What needs to go is advertising.

The evils of social media are not consequences of people using the internet to connect with other people, they're consequences of people using platforms where you can buy a following instead of having to earn it.

I guess it depends on what you're there to do. Sometimes I want feedback on a thought that's hyperspecific.

I could wander around town and strike up conversations with strangers and years would pass before I found somebody with the specific set of interests. But there in the depths I can find somebody else who is actually looking to talk about the same niche thing.

I don't know if friends is the right word, I only rarely meet them again later. But for a brief time we were a some kind of partner.

> Just the way how the packages are installed is guaranteed to be consistent across platforms

You get stronger safeguards than usual, but I'd stay away from "guarantee". You can easily write a derivation which does something crazy for x86-darwin and something sane for everything else.

That said, I love it too. It's so much closer to the kind of consistency non-computery people expect of computers than anything else.

> Stripe controlled blockchain could offer that a database could not.

A database cannot resist tampering by somebody with admin access to the database. It may be the only thing that blockchains have going for them, but it's a big one.

Whoever controls the protocol and network nodes can indeed unilaterally alter the blockchain just like any other data structure.

Yeah sure, if there is a single such party. I haven't looked at how stripe is implementing this, but since it's a blockchain total control is an option, not a requirement.

But this is a classical consensus setup where stripe has a front door to change anything about the network they want.

Never trust a cryptocurrency developed by a for-profit corporation.

Stripe would never create a blockchain they can't control, because if it has anything to do with fiat then it's going to need to have that control. I have an extremely high suspicion they have some sort of admin access.

You can sign records in a normal database. Implement some segregation of duty and don't give your DB admins access to any signing keys.

Which is by the DAO exploit on Ethereum was successful and not rolledback by the Ethereum Devs.

Oh, wait... I've been handed a piece of paper...

The protocol allowed it, there's nothing wrong with that. It's not like a DB admin just made a change without having to worry about what the rules were.

you get enough voting power and you can fork any blockchain to whatever state you want, no different than an admin doing an upsert

Can you name any significant economic fallout of any kind related to this attack vector?

Because I feel pretty confident that it is dwarfed by the volume of money that has been unlocked by tying crypto to ransomware.

Enron

I have similar concerns, but I think it's the payment conglomerates you have to worry about. Visa knows how much you spent, Fiserv knows what you bought.

Sometimes Visa also knows what you buy: https://www.marqeta.com/blog/data-details-what-is-level-1-2-...

It's not just Fiserv, there are often others involved as well. Even the merchant gets more data as they now have a fairly good way to identify you and correlate transactions.

Sure but that's only going to be among transactions that you made with that merchant.

If you think it's bad now, just wait until passkeys are ubiquitous and best practice is to only trust a small list of providers. The only way to prove you're human will be to prove that you're Google's human.

To an extent, I already saw ads on various fora effectively asking for pretend humans ( you sign up to a list with your info and 'they' use it in your name ). It is going to be another cat and mouse game to track and I am getting tired.

Publishing something is considered by most to be sufficient consent for it to be not considered private.

I realize there's a whole legal quagmire here involved with intellectual "property" and what counts as "derivative work", but that's a whole separate (and dubiously useful) part of the law.

That is definitely normally true but I feel like the scale and LLM usage turns it into a different problem.

If you can use all of the content of stack overflow to create a “derivative work” that replaces stack overflow, and causes it to lose tons of revenue, is it really a derivative work?

I’m pretty sure solution sites like chegg don’t include the actual questions for that reason. The solutions to the questions are derivative, but the questions aren’t.

Stack overflow doesn't really have a legitimate claim to that data either though. Nor do the users, we're just pasting error messages and documentation. It's derivative all the way down. It'll never sit still and behave like property.

Privacy makes sense, treating data like property does not.

Point taken, but it still feels like a gray area to me. The value that SO created was the curation of knowledge and high quality discussions that were well indexed and searchable.

The users did provide the data, which is a good point. But there’s a reason SO was so useful to developers and quora was not. It also made it a perfect feeding ground for hungry LLMs.

Then again I’m just guessing that big models are trained on SO. Maybe that’s not true

Replacing stack overflow has no bearing on the definition of "derivative"

I'd believe 90. 99.9 seems aggressive.

I think you massively overestimate how much the average user knows about computers

The statement was 99.9%, so that's not talking about average users; it's talking about the top 1% or less of users.

If 99.9% percent of users don't know what JS is, that means even a majority of professional software developers don't know what it is.

It doesn't matter how they feel about LLMs, ignoring their battle hardened plugin system and going native would be bad architecture.

It’s just native support for ghost text. It’s not llm specific