An activitypub is a pub you go to for activities. It’s fairly self-explanatory.

Get yourself a fairphone next time you need a smartphone. They’re not perfect, but they’ve tried to disentangle their supplier chain from the orphan crushing machine as much as possible.

That sounds like a just so story.

Spanish raiders could and did very easily talk to the locals they were subjugating to find out more about the plant. It was often described at that point as “a kind of truffle”. It was considered for a long time to be suitable for peasants and slaves, but not anyone with the option to eat ‘better’ food. The real reasons it took so long for potatoes to catch on are deeply rooted cultural norms around farming, class stigma around this ‘peasant food’, and worries around it being poisonous.

At every stage of it’s introduction into european diets, they knew how to cook it.

I actually prefer the smaller squad sizes and more barebones strategic level. Having a more RPG-like structure and more focus on a small number of characters is my jam. I looked at xenonauts, I think it’s absolutely amazing for a specific kind of player, and I know that I’m not that specific kind of player.

Menace’s solution to this is to basically remove the option for most weapons to miss entirely. A squad in that game is treated like a single unit, the equivalent to an XCOM/Xenonauts soldier. When you attack, it’s the same percentage roll applied to every bullet in the burst for each soldier. This means a single attack might have 40 individual dice rolls. The question is not if you hit, it’s for how much. Seems like a nice way of solving that issue without adding a whole bunch of units.

Cyber Knights does something similar: an attack with most weapons is a burst with each bullet’s chance to hit being individually rolled. This normalizes the chance to hit so it’s not all or nothing.

I love that game, played the crap out of it, waiting for another major patch to come back.

I was considering buying this as a massive fan of XCOM 2.

Was.

I’ll just keep playing Cyber Knights: Flashpoint thank you very much.

Do not redeem. DO NOT REDEEM!!

I like bread as much as the next guy, but do we really need to conquer it? Can’t we make a diplomatic agreement with bread, preferably one where we trade sugars for gluten development?

Why are you wasting time on Lemmy then? Shouldn’t you be posting on Facebook instead?

Defenestration wasn’t invented by the Russians. You can blame the Czechs for popularizing the term.

I wonder: if the EU-US divide continues deepening to the point where the EU kicks out US tech companies, and the EU continues supporting Ukraine, will campists support Ukraine because they’re allied with enemies of the US, or continue supporting Russia because they’re enemies of the US?

Campist math is hard.

This isn’t universally true and depends heavily on what the user uses their PC for and what hardware they have, but I have family members who would constantly ask for tech support for Windows. Just neverending issues. I installed Ubuntu for them, and tech support requests dropped by an order of magnitude.

Nowadays they barely even turn it on, it’s all phones and tablets.

Salting the water doesn’t help with that as far as I know. Salting pasta water helps with exactly one thing: it tastes better.

Aww, I’m not on the list. Is this a personal failing of mine? Am I just too agreeable?

As someone physically capable of playing games with the blinky shiny lights, they aren’t all they’re cracked up to be. My top two games atm are Manor Lords and MENACE, both of which should be perfectly accessible to you.

I’m very happy to buy games from small indie developers. They absolutely need the money to survive, and I enjoy their work.

I think you may be misunderstanding the situation. He’s telling AI critics to stop complaining or fork. If that happens, the main branch will still have AI submissions, and the fork will be the AI-free one.

Important to note: Torvalds is not an LLM evangelist, nor does he just accept AI slop uncritically. The official policy is that LLM written/assisted code is fine, but the quality must hit the bar expected of Linux kernel submissions, and the submitter takes full responsibility for the code correctness whether they personally wrote it or not. Just submitting LLM code with no verification or tests is a great way of getting banned from submitting pull requests and getting a very angry email from Linus.

The point is I don’t think LLMs normally use copyrighted code in a way that would hurt open source projects.

I don’t know. I’m not a lawyer, and copyright for code was a hot mess even before LLMs got involved. With how many opportunistic copyright/patent trolls there are and how easily convinced judges have been in the past, it could go either way.

Lol, so how do humans code in comparison?

The good programmers normally code by breaking down the problem into constituent parts and logically working through the problem, step by step. What differentiates this from tokenization is that instead of just looking for code that is similar for a similar problem, programmers can usually understand the effects of each line of code, visualize what the state of each variable will be in that step (or dump out the variables to look directly if unsure), and then move on to the next step. This logical problem-solving approach is fundamentally different from a tokenization+noise looking for a similar-looking problem approach. For one thing, you can solve problems that haven’t been solved before.

That would rely on the contact of that user using the admin’s public key. In most systems I’ve seen that’d result in a big flashing warning that the user’s key has been changed. So, dangerous for people contacting you for the first time, much more obvious if the 2 users have been talking before that point.

Other people have raised the much more interesting and potentially dangerous point that it’s very difficult in this context to make sure that a particular public key corresponds with a particular user. I’m way more used to sysadmin style issues where you have a small number of known keypairs, while in this context it’s a large number of mostly unknown keypairs, so you need some way of confirming that. I’m starting to understand why this is a much thornier issue than it appears on the surface.

LLMs don’t “create”. Under the hood, they’re tokenizing the queries, looking for “clouds” of tokens that are similar to the query, then returning a sequence of tokens (with some random noise thrown in) that match what their training data says the answer should be.

In short: all LLM code is an amalgamation of their training data by definition. If there’s nothing similar in there, it’s literally not possible for it to be part of any response.