AI companies should pay billions to wayback machine for access

I think that'd raise serious copyright concerns, if the Wayback machine started selling other people's intellectual property.

Shouldn’t it follow that it’s illegal for the AI labs to profit off of all of that stolen copyrighted data too?

It should, but it doesn't.

They wouldn't be paying for the content, just the bandwidth. Like buying a linux OS on a CD ROM was about the cost of media not profiting off of the software.

Isn't that already a big part of reddit's business model?

It's time for copyright to end anyhow; that's what's gumming up the whole project in the first place.

I'd have a lot less of a problem with AI if everything that went into their training was public domain and made easily available to anyone for any use. It'd feel less like AI companies were just stealing the work of others and charging for it.

Seems like a reasonable norm:

* If you train AI on it, you have to afford public access to it.

* Nobody can exact violence against anybody else in response to that person providing public access to any data anymore (ie, all bytestrings are public domain).

That's the world I'd like to try in the coming years.

Maybe they should have shorter and clearer terms? Maybe it should be easy to just opt out?

I call BS on that, it's been 6 months since Opus 4.6 and we still do not see a revolutionary software release or a major increase in quality. At least we see some security bugs being finally fixed.

I think GPT 4 was meant to do that.

Can they support something like modern linux? Every single youtuber talks sh/t about davinci and they still do not care

Talking shit about Linux support or the software itself? I know at least 4 YouTubers that do all of their edits in resolve

Davinci oficially supports Linux but in the worst way possible - "it kinda works if you hack around it, we haven't bothered" There is no official Ubuntu or current RHEL support and you have to mess with libraries and stuff to make it kinda work. No proper window decoderations either

How to view the previous version to compare?

Images from their X account

While haiku is almost one year old. What a joke

Sonnet is the new haiku

Unfortunately GLM5.3 is a cheaper replacement for haiku than sonnet

who cares about c&d? just host it outside of us

An anonymous group of devs should host the source on Radicle (over Tor) and the app itself on an onion site. Bonus, it may discourage some scrapers who can’t be bothered with setting up a Tor gateway.

If the operators are in the US, they probably don't want to risk it if they were anything less than very careful in not having their personal info attached to any of the data.

I've always found it baffling that more projects by folks that disagree with copyright, access controls, etc aren't hosted in places where the cease and desist won't work.

It's like, you're doing something to annoy a large company with lots of legal resources, and you host everything in the US? Why?

Because the US is the center of the tech world. The default AWS region is in the us. All large tech companies are in the us. If you use any cloud provider the default and cheapest region is in the us and so on. Everything tech is in English and the largest English speaking country is the us.

Apparently meta can still sue foreign individuals in California if you scrape 300k accounts like a Turkish man did

gpt oss killer? this can easily run on a server cpu with its memory bandwidth

You must have hibernated for a year. Most modern 27b models can outperform gpt-oss-120b.

No they do not. In some tasks yes, but they do lack when producing output in other languages.

Not exactly surprising given it's a dense model at 2.7x the size of the experts in gpt-oss

gpt-oss is long dead though. It's been 8 month since Qwen 3.5 was released.