AFAIK all the info is from leaks, and there’s no guarantee that leaked stuff ends up being actually launched.

I don’t think this is even legal in some jurisdictions. Many places have regulations that restrict/prohibit drip pricing, and I’d argue that automatically preselecting options that cost more is a dark pattern that’s adjacent to drip pricing.

Interesting. I’ll take a closer look and read the article you linked to. I haven’t used Kagi much, but I did try the trial and their search results are definitely better than Google.

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.

Their mission has always been about making the web better for humans, though. Their AI stuff is all opt-in and they have a decent philosophy about it: https://help.kagi.com/kagi/why-kagi/ai-philosophy.html

Kagi are trying to do something similar with their smallweb project, where they’re building a list of personal sites/blogs. These types of sites usually don’t rank highly in search engines, but Kagi does rank them above other results since they’re almost always better than the SEO slop that ranks highly in other search engines.

They have a site to randomly browse them too, kinda like StumbleUpon if you ever used that. https://kagi.com/smallweb/

The list is in this repo: https://github.com/kagisearch/smallweb

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…

Meanwhile there’s rumours of Apple working on airpods with built-in cameras 👀

That would be an issue, but I’m a person of habit and it’s part of my routine now, so I don’t forget.

All my medicines are in pill bottles, so the way I solved this is by flipping the bottle upside down after I take it, then flipping it back around when I get up in the morning. If I see it the right way up then I know I haven’t taken it yet for the day.

Week starts on Monday in ISO8601, and developers like ISO8601 so this is also the case in most programming languages by default. In some programming languages, it varies by locale.

edit: also, Sunday is part of the weekend, not the weekstart.

The start of the article says:

The lazy answer is always “weaker hardware at home.” And sure, that’s part of the story, but it’s a small, almost surface-level part.

yet nearly the entire article is about consoles having weaker hardware, and the fact that console controllers are not specific to each game (which I’m also including in the “weaker hardware” bucket).

The only point that’s not about weaker hardware is that arcade games generally have a higher difficulty since they want you to keep dying and inserting more coins, whereas console games are more balanced.

I don’t usually get involved in internet drama so I wasn’t aware of this.

I started with Conduit. I hit some bugs though, and development was very slow. I looked at all the forks. Conduwuit was the most active, and all the issues I had were solved, so I used it. It got archived because of some other internet drama, and the readme said to use Tuwunel instead, which is what I’m using now.

Should I use something else?

I dunno. Use whatever I guess, as long as it’s not Synapse.

In my area, they removed the data caps last year, when they also rolled out the higher upload speeds (eg the gigabit plan now has 350Mbps upload speed instead of 25Mbps). I don’t think they’ve rolled that out everywhere though.

They also have a habit of keeping people on old plans that are either slower or more expensive. If you’re still using Comcast/Xfinity, check their site and see what they offer to new customers at your address. (you may need to use an internet connection other than your Comcast one, otherwise they may detect that you’re an existing customer)

Their pricing can differ by region. Not every region got the data caps back when they rolled them out.

It also depends on competition - if there’s a lot of competition in your area, Comcast will have better deals compared to if you live somewhere where Comcast is the only option.

I used to live in an apartment that had an exclusivity deal with Comcast (residents were only allowed to use Comcast for internet service). Now I live somewhere in the same county where there’s three different fiber providers in addition to Comcast, and their prices here are noticeably cheaper vs the old apartment.

They used to charge you more for unlimited data if you didn’t use their modem. Renting their modem plus getting unlimited data was $30/month, but just getting unlimited data with your own modem was $50/month.

Not relevant any more though, as (at least in my area) they completely removed the data caps.

I’m so glad I have a local ISP that provides 10Gbps for cheaper than Comcast’s 1Gbps.

rcrypt can handle encryption transparently: https://rclone.org/crypt/. You can use rclone mount to mount storage from practically any provider.

This is going to happen with every large centralized service, especially if they let you use it for free. These services end up being expensive to run.

The only real solution is to use something decentralized. You’re already on Lemmy so I assume you already know this. Matrix is decent (not perfect, but it’s pretty good) and it’s easily self-hostable using Tuwunel.

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

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.