Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Knowledge Should Not Be Gated

formaly.io

Post reply on HN
Print view View original

21–30 of 57 posts

Re: Knowledge Should Not Be Gated

#21

Post by dofm »

coffeefirst wrote: post #16
prodigycorp wrote: post #9

Earlier quoted context omitted.

because corporations are using providers with ZDR in the contract. If OAI or any of the cloud providers violate this they're getting sued to oblivion.

These are the same people who performed the largest scale breach of copyright in history on the theory that they could get away with it. I’m not making any accusations, but we should not underestimate their tolerance for legal and financial risk. It may be a little paranoid to insist on self hosting based on that, but I’m not so sure that it’s crazy.

Trade secrecy is all anyone has left. It's not paranoid at all. You would hope that most serious companies have a tier of corporate knowledge protection that is somewhere between Coca-Cola/KFC herbs recipe secrecy and Stringer Bell's note-taking exhortation: "is you giving the LLM notes on our unique advantage?"

Re: Knowledge Should Not Be Gated

#22

Post by WarmWash »

dofm wrote: post #12
prodigycorp wrote: post #9

Earlier quoted context omitted.

because corporations are using providers with ZDR in the contract. If OAI or any of the cloud providers violate this they're getting sued to oblivion.

The problem is that there is an enormous, nearly unignorable incentive to work around it. So they will. As the customer base becomes more and more corporate (which it will), they end up with disproportionately more customers whose experiences cannot be used to train the model to make it better for those customers. Either way, corporate customers cannot leach off the training from consumers handing over their personal…

Conspiracies are for the chronically online

Re: Knowledge Should Not Be Gated

#23

Post by dofm »

WarmWash wrote: post #22
dofm wrote: post #12

Earlier quoted context omitted.

The problem is that there is an enormous, nearly unignorable incentive to work around it. So they will. As the customer base becomes more and more corporate (which it will), they end up with disproportionately more customers whose experiences cannot be used to train the model to make it better for those customers. Either way, corporate customers cannot leach off the training from consumers handing over their personal…

Conspiracies are for the chronically online

This is not a conspiracy theory. It's futurology, maybe, but pretty basic stuff at that.

At some point, where does the training advantage for specialist LLMs come from, if not progressively encroaching on customer data for the benefit of equivalent customers?

Re: Knowledge Should Not Be Gated

#24

Post by WarmWash »

coffeefirst wrote: post #16
prodigycorp wrote: post #9

Earlier quoted context omitted.

because corporations are using providers with ZDR in the contract. If OAI or any of the cloud providers violate this they're getting sued to oblivion.

These are the same people who performed the largest scale breach of copyright in history on the theory that they could get away with it. I’m not making any accusations, but we should not underestimate their tolerance for legal and financial risk. It may be a little paranoid to insist on self hosting based on that, but I’m not so sure that it’s crazy.

It has been ruled that training on copyright is not a breach of copyright unless you subverted payment for it.

Which they did do, but scale is relatively miniscule to the full dataset.

Re: Knowledge Should Not Be Gated

#25

Post by jdw64 »

Personally, I think the ability to distinguish between all the knowledge that's overflowing is becoming a characteristic of the current establishment. In reality, the number of sites where you can get good information is extremely limited. It feels like we're in an era where discernment matters more.

Most of it is just misinformation, after all. People say knowledge shouldn't be restricted, but now we have the opposite problem. There's so much information that just skimming through it takes too much time. On top of that, as we shift from text to video, getting information has become even harder. Compared to text, YouTube videos feel like they have much lower information density. I've heard that the TikTok generation's text literacy is declining, but maybe that's actually a social adaptation to process as much data as possible from low-density sources

In that sense, the efficiency of RAG ultimately comes down to what kind of good knowledge you're feeding into the AI.

Re: Knowledge Should Not Be Gated

#27

Post by api »

rightbyte wrote: post #5

It seems beyond naive, rather malicious, to upload any useful private data to SaaS LLMs. Like, you are letting them data mine your business. Why are corporations not panicing over this?

Imagine someone comes to you and says: "You must remove your door locks. Anyone can come into your house any time. You also need cameras across most of your house. But in exchange, magic elves will do all of your home chores: washing, dishes, folding laundry, cleaning, minor home repairs. All of this will be done for pennies on the dollar compared to any current option."

How many people would take it?

I know I'd actually be tempted. Con: total loss of privacy. Pro: it folds laundry, and I f'ing loathe laundry with the intensity of a billion suns.

Every business has similar trade-offs they'd be tempted to take.

Re: Knowledge Should Not Be Gated

#29

Post by Herring »

dofm wrote: post #20

We are at the breathless-but-low-information-posts-about-plain-text-formats point in the cycle.

This comment is another example of "Nobody ever gets credit for fixing problems that never happened."

There's a massive push to add unnecessary complexity to everything out there, because complexity pays all our bills.

Re: Knowledge Should Not Be Gated

#30

Post by em-bee »

api wrote: post #27
rightbyte wrote: post #5

It seems beyond naive, rather malicious, to upload any useful private data to SaaS LLMs. Like, you are letting them data mine your business. Why are corporations not panicing over this?

Imagine someone comes to you and says: "You must remove your door locks. Anyone can come into your house any time. You also need cameras across most of your house. But in exchange, magic elves will do all of your home chores: washing, dishes, folding laundry, cleaning, minor home repairs. All of this will be done for pennies on the dollar compared to any current option." How many people would take it? I know I'd actu…

i believe that in the future technology will be so advanced that protection of privacy is impossible. the only way to counter that is education to respect peoples privacy and very harsh punishments for violations.

i also believe that we will live in a post scarcity world, which means profit is no longer interesting, so any business case for invading your privacy will go away and therefore it will only happen for personal interest.

the key in any case will be education, because without it abuse will be rampant and progress will halt because everyone is going to be suspicious of everyone else.

Post reply on HN