They destroy the books because it allows them to scan them bette

That’s definitely one reason, but the copyright argument is also a part of it. The court explicitly said that their digitization is legal only if does not increase the number of copies of the book.

There’s no need to do this legally - they could donate used books. The argument shows that the final product is transformative, so it doesn’t matter whether they keep the books or not.

They have to keep the digital copy of the book because they add it to the training corpus for the LLM. Selling or donating the original physical book after doing that would void the fair use argument.

LLMs can reproduce quite long passages of books, though I don’t think this changes the argument much, because it’s not reliable or useful.

One of the tests that determines if it’s fair use or not is whether it can serve as a replacement for the original book. Pirated copies can, which is why they’re illegal. A summary like CliffNotes can’t. Even if the LLM can reproduce long passages, you can’t do that reliably (like you said) and it won’t work for all books.

I always thought there was in obvious win where companies doing this could be forced to archive the scan publicly (after some period of time)

I definitely agree with this. I think copyright law needs to be modernized to handle cases like this. I think the AI companies should be allowed to donate the digital copy to a library (like the Internet Archive) while still being allowed to keep their copy in their training corpus.

I don’t see how this would go any differently to the book lawsuit they settled.

With that lawsuit, the court found that training the LLM on books is covered under fair use, as long as the books are legally acquired and they do not make illegal copies. (obviously, they can’t use data fron Anna’s Archive any more)

The argument is that them training on the books does not impact sales of the books themselves - someone that was going to buy the book would still buy it even if an LLM was trained on it, just like they’d still buy it even if summaries/CliffNotes and reviews exist. The training is considered transformative as the LLM learns the relationship between the words rather than retaining the text of the books as-is, and transformative use is allowed under fair use.

This is why the AI companies are destroying books - they have to destroy any books they scan in order to comply with US copyright law. Transforming the book from one form to another (eg from print to digital) is OK as long as only one copy remains.

I’m not saying I agree or disagree, just that this is the current state of copyright caselaw in the USA.

Having said that, the music industry is significantly more litigious than the book industry, so we’ll see what happens…

Now that’s a throwback. I remember the migration from ipchains to iptables taking forever, kinda like the migration to nftables now.

There’s places that still use Styrofoam for food? I haven’t seen it in a long time.

Stuff that’s out of copyright isn’t an issue though.

The reason AI companies are destroying books is because current US copyright caselaw (most recently the Anthropic lawsuit) doesn’t allow books to be copied, but considers it fair use if you transform the format of a legally purchased book (like from print to digital) without creating a new copy. By destroying the original book, they haven’t made a copy of it, and so they operate within US law.

This doesn’t apply to public domain books (books that are no longer copyrighted) as you can do whatever you want with public domain content.

At work, I use AI for some things. Right now I’m rewriting some legacy spaghetti code that’s had a bunch of things hacked into it over the years. I spoke to the person most familiar with the expected behaviour and used AI to combine his info plus the existing code and unit/integration tests into a list of requirements.

I wrote the new code and tests based on the requirements rather than based on the old code. After each commit, I used AI to check for parity between the old and new code, and it keeps a Google Sheet up to date with the progress (which features were fully implemented, and which ones were missing or had gaps). I had AI write some tests cases too - given the list of requirements, write integration tests for them based on the style of a few tests I wrote by hand.

It has some quirks (eg for tests it loves over-mocking even though our skills tell it to mock as little as possible) but it definitely speeds things up.

I use AI for small side projects at work too. Tweaking and adding features I want to shared libraries, internal tools to help our team debug stuff and automate triaging of bug reports (they’re all still reviewed by a human), etc.

The entire reason I can trust its code is because I can read it and tweak it myself. I sometimes need to go through a few iterations to get AI code into an acceptable state. AI writing machine code directly, like what’s been talked about recently and what this post is referencing, is such a dumb idea.

There’s other people at work that use AI for absolutely everything. Writing code, reading code, writing posts in our internal groups, etc. That’s something I don’t understand. Some people that are all-in on AI produce so much low-quality AI slop.

“You’re absolutely right!”

Also

The ICE agents arrived after she was gone. Laurie recognized two of them from an earlier encounter. About 10 minutes later, one approached her, showed his badge and asked whether she could confirm that the passenger had checked in. Laurie refused, telling him airline employees could not provide passenger information.

As the flight neared final boarding, another ICE officer accused Laurie of interfering with the operation and pressed her to help locate the passenger. Laurie declined to say whether the woman was on the flight’s passenger list and told the officer he would need to contact the airline’s corporate security department for information. “You’re going to need a subpoena,” she said. No arrests were made at the gate that day.

Legend.

Unfortunately I worry that ICE is going to target Laurie in some way now :/

The title of the page I linked to is “Websites using Cloudflare in the Top 100k Sites by Traffic” so I think my interpretation was correct.

Also, one domain can use multiple CDNs (eg one subdomain uses Cloudflare while another uses Bunny) so I’m not sure how their percentages work.

I’m also not sure how they count sites that use Cloudflare but not the CDN product (for example, if they disable caching and only use the WAF). That looks identical from a client side so there’s probably no way to tell.

Maybe I shouldn’t trust their data.

AWS(just 8% seems sus),

The chart wouldn’t be showing all of AWS, just Cloudfront (their CDN). There’s plenty of services that use AWS EC2 or Lightsail but front it with Cloudflare.

If you’re considering hosting Github Enterprise on-prem then it might be worth considering Forgejo too. It’s the system that powers Codeberg.

Codeberg doesn’t allow proprietary code, only open-source. You’d have to self-host Forgejo if you want to use it for private projects.

Something like 40% of the top 100k sites use Cloudflare. Builtwith says 44%: https://trends.builtwith.com/websitelist/Cloudflare/High-Traffic-Volume

The pricing on their paid plans is very competitive. For small to medium sites, it’s orders of magnitude cheaper than Akamai. The WAF and DDoS protection is pretty good too.

The largest sites like Google, Facebook, Netflix, etc have their own CDNs, including servers they provide to large ISPs for free since it significantly reduces upstream bandwidth usage (Netflix call this Open Connect; Meta’s is Facebook/Meta Network Appliance). Outside of the largest sites, you’d be surprised how many companies rely on Cloudflare. It’s kinda scary.

I think that’d make sense. The trackers for public torrents are often at seemingly random domain names and nobody really pays much attention to who runs them.

The US military created Tor, so I don’t doubt the US government still runs a lot of the nodes.

I doubt it. It’s likely custom code that’s better suited to the task.

The *arrs watch RSS feeds to find new torrents. It wouldn’t be difficult to write a script that pulls the RSS feed and polls the tracker to collect data on the initial seeds/peers. They might use an actual torrent client for some of the data collection, since I’d guess they have to prove the person is actually uploading, rather than just being connected but not actually uploading or downloading.

cat is misunderstood a lot. The purpose of cat is to combine multiple files together (it’s literally short for “concatenate”), like cat *.log to combine all log files together.

If you’re just using it for a single file, then you should just redirect the file to stdin. These two command lines behave similarly:

cat foo.txt | some-command

some-command < foo.txt

For head, tail, grep and some other commands, you can just pass in the file name directly as an argument (e.g. grep whatever foo.txt).

This looks like an AI edit rather than an actual renovation.

The shelf thing on the left has exactly the same items on it. Why would they keep the hole with the CRT TV in it? Also I’d hope that nobody would actually put a square table in a round space.

hold hold

Maybe he just needs to rest his arm.