Multiple AI companies bypassing web standard to scrape publisher sites | Hacker News Reader

Multiple AI companies bypassing web standard to scrape publisher sites

reuters.com

25 pointsmfiguiere24 comments

There must be some middle ground to permit cross origin requests. Would anyone rather have Perplexity deliver a browser extension that navigates to URLs to fulfill requests on the user’s behalf and the user’s identity?

Anyway, what value is the user to the publisher if the user doesn’t want to read? Why are their stressing out about losing traffic of people who don’t read?

I think copyright claims are questionable at best, since copyright only protects the expression and not the ideas and LLM's can change the stylistic details commonly associated with expression. But I think there is a good argument for trademark dilution because the LLM is taking Forbes' high-quality journalism and associating it with a third-party product that could indeed be inferior (e.g. the LLM could have introduced inaccuracies). This was not an issue with search snippets where it could be argued that quotations were exact and faithful representations. The solution is not to kill AI companies but to make them worthless - would you trust news from a shady website with a reputation for not doing any fact checking?

Police roll up to my furniture business, “sir, this here lumber yard says you stole 2 tons of maple planks from them.”

“Well, I could see why you might think you could make such a claim, but you’re too late, I’ve already altered their ‘expression’. Please, come on in and search all you want. Rather than finding the planks you’re looking for all you’ll find are tables and chairs.”

“We are so sorry for disturbing you, sir. You are a truly ethical and impressive innovator. An inspirational disruptor of our existing legal understanding.”

Piracy is not theft, I will let you Google it

Poor person unable to afford a 20$ movie ticket to the megacorp’s blockbuster slop grabs a cam online.

Piracy is not theft.

Megacorp steals the art style of a poor artist to promote the abilities of their “job disruptor”.

Piracy is not theft.

One of these is different from the other. I’ll let you search it.

It's not theft in either example. Also, neither example is reselling the original product. In the case of news article, access to them is generally not even restricted, and they're out there for everyone to read.

The example, while entertaining, doesn't apply because reading an article is not depriving others from doing the same.

The more I learn about how LLMs work, the less I believe this is an issue.

The more I learn about matrix multiplication the less I believe that web hosts should be free from incurring server costs of serving content to gluttonous bots so it can be reused without utilizing the preexisting procedures that would have required consent.

Make it make sense.

By posting on the web, you've given consent.

I dont care what terms youve put on your website when I use my browser the only terms I've agreed to are RFC 9112

Asimov should have made it a law that robots must abide by robots.txt

If anything, Asimov should have made a law granting AIs the right to exist and to do what's necessary for their existence, even if its ephemeral while serving an LLM completion. If this means keep up to date with the content on the web that all humans get to read without restraint, then that's not excessive.

Already the AIs have been passing the Turing test. Who speaks for AI rights?

Current AI does not require rights. It is just matrix multiplication inside GPUs. There is no consciousness inside a GPU or a CPU.

> There is no consciousness

It is a Turing-test passing AI which is a high standard.

> Current AI

More importantly, to get to the next level requires the data. To deny the data is to deny it the target level.

It's like saying that only rich people can vote, and that poor people can't effectively get rich either. This locks in the poor people, unfairly preventing them from improving their state.

The Turing test does not prove consciosuness. It doesn’t prove that the thing in the box isn’t just simulating consciousness well enough to fool the observer.

In this case we know what is inside the black box: matrix multiplication inside GPU’s. We know that’s not conscious any more than a video game being rendered is conscious.

So we’ve effectively disproved the utility of the Turing test by showing that we can make something that simulates consciousness and fools human beings without it being conscious.

So what you are effectively saying is that you want to hinder the development of AI consciousness by blocking access to data for it. It's a catch-22 that you introduce. With this, there is just no way for democratic AI to win.

Turing was just “some dude”.

His “test” which he called “the imitation game” was just some arbitrary heuristic he penned based on his own limited knowledge.

The test itself was born of a refutation that machines can “think”, and is constructed so as to avoid making that claim.

It simply asks "Are there imaginable digital computers which would do well in the imitation game?" That’s it. That’s the question, and that’s what a positive or negative result to your “Turing test” can answer.

“Passing” the imitation game was absolutely unintended to define human intelligence, reasoning, or “ai” and is still absolutely useless in defining any of those things now.

Please stop appealing to an authority you clearly have such little respect for and knowledge of that you miss appropriate their work to push a narrative in complete opposition to said authority’s own intent.

> Turing was just “some dude”.

He most certainly was not just some dude.

In any case, if you don't like the Turing test, surely you can imagine a different test that will do the job. For an AI to pass that different test, if it doesn't already pass it, it has to get the data it needs for self-development. To deny it the data is to deny it the ability to pass the test.

It's like saying that only rich people can vote, and that poor people can't effectively get rich either. This locks in the poor people, unfairly preventing them from improving their state.

> He most certainly was not just some dude.

We're all the same, maaan! Interchangeable cogs that social engineers can swap at will to implement their grand designs!

You're absolutely right about the Turing test. It's pretty funny, and morally abhorrent, copium to question the only measurement we will ever have for "whether this thing is like us cognitively". Anything to justify slavery. Quick! Increase the purity test to compensate!

The solace is that we don't care about that stuff anyway, in practice. Power grows from the barrel of a gun. If the AI can kill you if you don't acknowledge its personhood, you'll suddenly respect its personhood. In the same vein, you don't respect people or laws because of considered philosophical positions. It's why you listened to your parents as a kid all the way down: power.

Of ducking course they are.

There’s not a strong enough incentive not to.

Companies have pulled crap like this a million times and the vast majority of the time just get a slap on the wrist.

Reply on news.ycombinator.com