krzyk

If google does that, then they do a poor job at that. In Europe OSM is way more detailed and up to date than Google Maps.

Yes, it is recentish, but not that recent. Remember Spore game? It allowed 3 installations, later they changed that.

I was never in the "resell software" mindset, similarly to books, I don't sell those, unless I get a present that is mistargetted. No one gifts me software sadly.

Well, yes. Always online is a problem, but it doesn't change what one buys. A thing that is easy to copy without destroying the original. So they invented licenses to contain the copying part.

Video games ar software, this has only digital form.

And AFAIR all software is sold as a license since way back.

That is literally in any software since like 90s. You buy license.

I hope you see how this is less secure than a docker container?

Agent that can apt-get is more useful.

Or provide them.

Here there is at least same prefix.

Only issue I recall might be with female Aleksandra (abbreviated to Ola) and male Aleksander (abbreviated to Olo, or Olek).

Others usually (if I remember correctly) have similar prefix.

Is it really?

Harness is important for model performance, but weights are surely mode important, without that you would have haiku doing the work.

I was afraid that it is for Marty McFlys only.

I don't think people keep all their money on a single account. I have 20 in my bank (different savings, some foreign currency accounts, etc.), and only one is tied to my debit card. I move money there when I need it. Opening another account number is just few clicks away. There is a limit for it also.

Single account sounds more like a boomer thing.

File format for what? Text, graphics, compiled code?

No?

The only PC peripheral I have with USB-A is a mouse dongle when I'm lazy and don't connect bluetooth - and that one I connect to the monitor.

All others are usb-c.

3090 and 2x3090 are quite popular. But if you uses gigantic (for local models) context of 200k it will go south pretty quickly - any quantization of context quickly becomes the issue.

AFAIR the general consensus is (was?): - llama.cpp for single user - vLLM for multi-user (e.g. enterprises)

They are similar, but for different use cases.

We need first to reach level of Sonnet 4.x, we aren't at that level yet.

It was quite fun.

But at my first work (begining of 2000s) there was one person that made a fun email, using From of head of company (or was it head of that particular division) to his coworker with congratulations for pay increase and promotion. It would be all great, but that coworker didn't catch the joke and replied to it (person in the From wasn't amused). Author of the joke was fired (which is not easy thing to do in Europe), some people don't catch jokes.

I'm a bit new to rust or npm system.

But I always thought NPM was what the author describes - just a random set of packages with git sources, which I thought was the main issue (leftpad etc.). Isn't that the case?

What about one system that just works and is there for "ages": maven repository?

Focusing outrage on politics is pointless.

Technical people are rather good at learning new things, and ollama situation is a good learning experience.

llama.cpp gets you more tokens/s even if you ignore ollama team bad behavior.

> People say Electron is heavy and not great, but it's also the fastest way to support Mac, Windows, and Linux all at once.

But why make an app when websites is enough? And I don't need to run n web browsers for that.

Solar panel breakage doesn't depend on a graphics card.

ollama is a wrapper on top of llama.cpp, and it makes llama.cpp slower, why use it?

Also Ollama has other issues (like forgetting what it really is - a wrapper).

Why would anyone use Ollama at all (aside from obvious reasons one can look up online) - llama.cpp used directly, without this wrapper is faster.

Basically one has two real choices for local LLMs: llama.cpp (if single user) or vLLM (if multi-user/enterprise).

Is autocomplete using LLMs really useful? Even with frontier models I found it to be about 50% right, I turned it of and prefer to use IntelliJ built-in, it is way more reliable.

For me local models is all about quality, and how to achieve that - e.g. by providing guardrails that test the job done.

People are using 3090 (24GB) to run models, and it is the most cost effective way to run the. Yes, it is 2x faster, but memory wise you surely can spend 24gb on llm.

Also there are smaller, still usefull models that can run on 8GB or less.

Ollama is not a good choice - https://sleepingrobots.com/dreams/stop-using-ollama/

As for oprncode, doesn't the system prompt eat too much of the context? Local models are really constraint in regards contex, and opencode AFAIR uses a 10k of it or some thing close.

It is done in the open, but it adds complexity and it removes that made Unix/Linux great - composability, variety and replaces it with corporate introduced "stuff". And any distro is forced to support those additions because corps owning Fedora, Redhat, Ubuntu just rule the Linux world, and event Debian gives up.

As long as there are just few "normies" using Linux, it is safe from corporations adding their "security", "safety" etc.

Yeah, this is quite concerning.

And FedRamp has some issues with data being sent out.

Our corp doesn't allow usage of local models because of concern about potential "agent sends out code to the net" issues.

> As long as the company states the time period

But they don't, they have the "30 days", but just after that they add "unless ....". So the time period is vague.