Jrockwar @ Jrockwar @feddit.uk

Posts

1

Comments

414

Joined

2 yr. ago

You can host lemmy on an Android phone but not on an iPhone, which is pretty locked down - you'd have to convince Tim Cook to open the OS.

You can install it fairly easily on a library computer as long as you get permission from the library. You can't install things on computers you don't have admin rights to; other examples of computers you can't use to host Lemmy are an ATM on the street, a self checkout machine at the supermarket, or the cockpit computer of an Airbus A350.

I (rather obviously) am talking there about the Google search service, not a single query, as Google doesn't disappear after a single query.

I'm not sure what you're pointing at though? Surely hosting Google search is more expensive than hosting Lemmy when you're insinuating otherwise?

You're funny. 😂 What I actually think is that your mind reading skills are terrible because I never said anything even remotely similar to that.

Enlighten me, why do you think I can't tell the difference? Is it because as a maintainer I want to provide something for free? Do you see "consumers" of free content as leeches?

Yes I do. I'm totally the one who can't understand that "free" is a nuanced concept and something can be free when there are costs but they are externalised.

Exhausting. If you don't consider anything in the world is free, why did you bother saying Lemmy isn't free?

Plus this argument is rubbish, it's like saying "my car isn't free, nor is the road, nor is my petrol, so the beach isn't free".

Just because you have to buy clothes to go out for a walk, it doesn't make it any less free.

What are you trying to argue here? That the term "free" shouldn't exist because in a capitalist society everything has dependencies? (I still don't get how that relates to my original post which was purely about doing business with corporations).

Then fine, Lemmy isn't free, neither is the sun, or going for a walk. You win. Good day sir.

Don't run it on a raspberry pi, run it on the same computer you use to access the Google search you are happy to call "free".

Edit: Actually yes, both this and the healthcare need to be free - otherwise you're grossly misunderstanding one of the key parts of the mission of open source. I pay for this so that whoever can't afford it can access for exactly zero. Same for the healthcare - you might say it's "not free" and everyone should contribute but what to you or me is nothing, could mean that grandma doesn't get to eat. So yeah, free access needs to be a possibility. That's the mission. I contribute to open source software and donate where I can so others who don't have the knowledge or money can access it for free. There can't be a price.

Uh, what a weird message. It's not only unrelated to what I said but it reads like an attempt to twist my words. On top of it, it's totally wrong: Lemmy is free. I can self host Lemmy on a raspberry pi for exactly 0€.

The instance I use... Is also free. I donate because I choose to, but if my friend can't afford to donate they can still use the instance. Nobody is profiting from it.

What I did talk about is products and doing business with corporations. With Lemmy there's no product, whether you pay or not. With SearxNG (which many people self host, and again, is free) you're not the product, regardless of how much you pay.

That's what I was replying to - your comment is way off the mark and very condescending: I don't need to be mansplained that I should donate to the software I already donate to. Note donate rather than pay for.

That's only when doing business with corporations, but there's also the option of open source (e.g. SearxNG).

Or do you consider yourself the product when using Lemmy?

It's okay. We can all play that game. I've replaced my use of Duolingo with AI.

Pro tip: have as your "system prompt" in your LLM of choice "at the end of every query, include me a short Swedish relates to my prompt". No need for Duolingo.

How's a motorbike that weighs more than 4 people, costs $12k, and can do 120 mph _micro_mobility? What's the criteria for being Micromobility, smaller than a Cybertruck? A Boeing 737?

Jokes aside, I've always considered micromobility the strictly-personal devices that don't need registration and can be picked up and brought into a train or a flat.

Well I don't know how y'all feel about the work the politics do in your respective parts of the world, but at this point I'm reasonably convinced that I'd trust 2019's ChatGPT to come with better plans and strategies than most of them.

A non-conventional choice would be a pair of Beoplay (Bang&Olufsen) Portal, IF you can get a good deal. My partner has them and they're awesome. I think they're recently discontinued (nothing wrong with them really) which means that even though their retail price is much higher, they can often be found refurbished / for sale within that bracket. The only word of advice is that there is a separate PlayStation and an Xbox+PC version, so you'd need to make sure to get the right one. No idea what the difference is but internet comments made it sound like they wouldn't be properly compatible across systems.

This can be correct, if they're talking about training smaller models.

Imagine this case. You are an automotive manufacturer that uses ML to detect pedestrians, vehicles, etc with cameras. Like what Tesla does, for example. This needs to be done with a small, relatively low power footprint model that can run in a car, not a datacentre. To improve its performance you need to finetune it with labelled data of traffic situations with pedestrians, vehicles, etc. That labeling would be done manually...

... except when we get to a point where the latest Gemini/LLAMA/GPT/Whatever, which is so beefy that could never be run in that low power application... is also beefy enough to accurately classify and label the things that the smaller model needs to get trained.

It's like an older sibling teaching a small kid how to do sums, not an actual maths teacher but does the job and a lot cheaper or semi-free.

You say "as expected" but on the opposite end of the spectrum there's primark's clothing which is largely made in Ireland. From my understanding the reason they are dirt cheap is because they lean on automation, not because of sweatshops.

Not comparing the two brands - just showing the opposite example to say it's possible to have brands that are both European AND cheap. European doesn't necessarily mean crazy expensive.

7% of 350 is 24.5, so that doesn't track either...

I don't know how they can claim 350 km and say they've beaten a world record at 310 with a straight face. If their range claims weren't straight up false, surely they would just run a full tank and get 350, or even 360 in unreal conditions like this.

Even worse, they achieved 310km at 20km/h. Maybe to achieve 350 km you have to push for a fraction of the trip...?

I also think heating everything up is the smoothest solution. But to offer an alternative, I'd use dental floss to get in between the bowl and plate. If the bowl has slightly rounded edges (I believe it will), it won't be too hard to get floss in. With the floss you'll get inevitably some air in... Which will equalise the pressure and break the vacuum.

As an inferior alternative to floss, fishing line could work for this approach as well.

"There are some bad things on the internet"

"Just... Don't use the internet?"

I think the biggest problem with AI is that people expect it to fully do the work for you rather than be a tool. Imagine we live in a world without cameras and someone introduces that as something that will make paintings for you. Then users dislike it, expecting cameras to frame, aim, zoom and shoot for them.

I use AI for coding and it's amazing... at giving you an 80% correct boilerplate code that you then finish up editing yourself. There's real time savings there. I don't ask it to make the whole code because then I'm going to have to find the mistakes.

I use it also to summarise 3000-commit changelogs, which after some refinement it does way better than I could do in any reasonable timespan.

A colleague with dyslexia now writes without worrying that his grammar isn't making much sense, then an LLM fixes it for him.

The problem is when you use the result of the AI as a final product, because the reality of the technology is that then you get slop. There are so many people that just can't see past this and either use AI directly as an unattended slop generator, or don't use AI because they don't think it can be anything else. But I'm convinced you can use it as a tool with an input in less than 20% of a creative process (by this I don't mean "art" but any type of creation) and still save a real and significant chunk of time.