flexagoon

They were doing great game updates until Caves and Cliffs, where they promised too much and had to split it over like 5 next updates. After that they decided to mostly just do useless "feature" drops.

There is an unofficial mobile port of Java edition, so no

https://github.com/AngelAuraMC/Amethyst-Android

Isn't this just Opengraph but in json? What's the advantage?

There are tons of existing Skills/MCPs for Google/Kagi/whatever search, and making your own is trivial. I gave DeepSeek in Pi a link to Kagi API docs and asked it to add a web search skill, and it did that easily.

> The companies that offer web hosting could get into this.

They already do. DigitalOcean is one of the providers on OpenRouter, for example

OpenCode Go is $10/month and the limits are much more generous than those or Codex

Why "etc."? Isn't it just node, bun and deno? Genuine question

Slurs are targeted towards specific groups of people (or I guess just anyone? "Clanker" is considered a slur, and you can probably come up with plenty of slurs for animals, they just won't be considered offensive and socially unacceptable like the ones for humans)

Sure, there are plenty of things to dislike about DMCA or any other modern copyright law, but in this context the discussion was specifically about using DMCA to take down an unauthorized republishing of a copyrighted book online, which is a pretty normal application of copyright law. So I don't really understand how "AI is bad because it's trained on stolen work" but also "it's bad to take legal action against people distributing your work".

Not sure what your point is; the title doesn't claim "AI stole the book", it says "an agency stole the book and used AI to republish it", which is true.

The fake website also re-published the entire 1:1 text of the original book.

So you're against copyright protection, but also against AI using copyrighted work? I can understand both of those positions separately, but how can you combine them in the same statement?

1. Age-gate social media

2. Children start using VPNs to bypass the ban

3. Age-gate VPNs

4. Repeat steps 2-3

Truly a masterful plan.

> HN recognizes what's going on

Maybe the pre-2024 users do, but I've seen plenty of those exact "frontier models never hallucinate" comments on HN as well

I agree, but I was responding to the question of why people expect LLMs to be like the star trek computer, and the answer is "because people making and promoting those LLMs claim they are like that"

Because AI company executives and devoted vibecoders constantly make egregious claims like "programming is fully solved" and even straight up "hallucinations don't exist on frontier models"

Not necessarily. Primary sources are allowed on Wikipedia in many cases, they're just not preferred. In fact, a direct quote from an interview would be as much of a primary source as a Tweet.

Depends on what the poster is. If it's still mostly text content, just arranged in a certain way, then by all means you can. For example:

https://typst.app/universe/package/poster-syndrome

And here's one for foldable cards:

https://typst.app/universe/package/keepsake

(Not endorsing those specific packages, they're just the first ones I managed to find)

For more free-form layouts, I just use Inkscape. It has built-in PDF export which works great. I assume other similar tools like Figma or Illustrator would also work well.

No, you can request your personal data as a part of GDPR (and most other privacy laws). That's why things like Google Takeout still exist.

> I've found graphene's keyboard far more error-prone than the stock android keyboard, but I also don't care to learn swipe to type.

Graphene's keyboard is the stock AOSP keyboard. Most Android systems ship with their own one instead of it, but that's the one that is built into the system by default.

They list their exact criteria for supporting a device. So far, only Pixels fit all of them (and I guess the Motorolas will soon)

https://grapheneos.org/faq#future-devices

AFAIK most malware like this first sends the contents of your environment variables, ssh keys, passwords, etc. to the server, and then sets up a persistent process that executes arbitrary commands received from the attacker's server at any time, allowing them to run whatever else they want

Not exactly what you're looking for, but OWID has a bunch of great visualizations about cancer, including ones that really show the progress we've made (and how much we've yet to make!)

https://ourworldindata.org/cancer

Yep, that sucks. But on the other hand, overbooking is a thing, so if you buy a ticket without seat selection for a popular flight there's a chance you literally won't be let into the plane you paid for because all the seats are already taken. Happened to me once, I had to agree to a free replacement ticket to a different city in an entirely different European country (Brussels instead of Cologne) and spend about 200€ on a late night taxi to my destination.

No, I mean they don't show it at all in their API to prevent distillation. The reasoning is encrypted and only accessible by their server. They sometimes show a summary of the reasoning, but it's not the same as the actual contents of the reasoning.

Is seat selection really a dark pattern? Exit row seats are the most worthwhile add-on to buy when flying IMO, sometimes having even more leg room than business class seats for a fraction of the cost.

> Essentially $20 for unlimited GPT 5.5 agents entirely deployed in the cloud.

How does it make sense financially for you?

Netherlands has always been pretty firmly against Chat Control (except for one political party, EPP)

If you're on Linux, the easiest way IMO is to just run the agent in bwrap

I do it like this

https://github.com/flexagoon/dotfiles/blob/main/dot_config/f...

But I'm sure it's simple enough that you can just ask the agent itself to make you a command for it with proper bwrap configuration

What does QGIS have to do with this? It's not even an OSM editor.