Meta should go and pound sand.

my big thing is honestly storage space which is terible. I have 256 GB but thats not enough with the size of some of my games. The model i got had no 512 Option back then.

Security is not a fixed state, a closed system is not fundamentally more secure as the most vulnerable component is still within the system. The user.

Closed systems are fundamentally more secure. Principle of least privilege demonstrates this

Finally John Apple

Quite alot richer than the 4th richest Economy in the world?

In a per capita sense of course. But it does pose an interesting question. To what degree does the quality of rail scale with absolute investments vs per capita investments. On the face of it rail is almost entirely fixed cost. The capital investment for tracks, trains and the operating cost for staff and energy are fixed to matter how many or few people the trains are serving. The crux is that for a given cost the quality scales inversely with the population and covered area. Transplant the Swiss rail expenses to Japan and it would make for a pitiful experience. Stretching over an area 9x larger and having to serve 12x more people, simply requires a lot more rolling stock, track and personell to offer a similar level of experience. In particular when talking about rural rail service it's most apt to compare areas of similar size and population and in this local sense Switzerland is not just richer per capita but absolutely.

wsl2 is literally just a linux vm isn't it?

Yep, running on Hyper-V.

this looks nice but is macos only.

thanks for opening my eyes uninstalled ollama, currently trying out LM Studio will probably try the othe ones aswell. the .cpp always scared me from using llama.cpp directly.

I am so perplexed what exactly where people thinking they were. Its nothing else than highly sofisticated statistics.

From that perspective, which is totally correct, it makes you wonder what other domains of knowledge look like when pushed to the boundaries of our capabilities as a species.

That is a genuinely thought provoking idea.

Do you know of any other statistical model that can "hallucinate". They clearly have emergent capabilities that come from scale that are absent in any other statistical model we've ever dreamt up.

We know that LLMs build complex internal representations of language, logic, and concepts rather than just shallow word-counting.

If you deny that then you probably have an elementary understanding of how they work. Not even Chomsky denies that. The real argument imo is whether those internal representations constitute an actual "understanding" of the world or just flatten out to something much less interesting.

> Do you know of any other statistical model that can "hallucinate".

Actualy most statistical models can "hallucinate", specifically those that are capable of interpolation.

I have witnessed this for example in Gaussian Processes. In my own scientific work.

Well other than the limits of Copilots usefullness.