I’m sure Reagan deserves a lot of the criticism and he was possibly in early stage Alzheimer’s in his second term. However I mostly remember him being the president in charge when the cold war ended and Eastern Europe was finally liberated.

No screen protector but I did have a fairly solid caseology case for my Pixel.

In the discussions our project are having about how to approach LLM use in the community there has been a lot of consensus in the human taking ownership. Ultimately it’s the human who has to direct the tool and if they have little interest other than blindly hitting accept then we’re not interested in the contribution.

Some of the AI generated bug reports we get are good because the author had verified them, added a test case and at least reviewed if it is guest trigger able. Yesterday’s flood however was low effort spam.

I wouldn’t say it’s totally legacy. (v)ram bandwidth does matter and while the Apple M-series chips do well with their unified memory architecture don’t forget it’s fixed because it’s part of the CPU chip.

I don’t doubt that locally hostable models will be important for agentic tasks but I suspect they are still going to be bigger than most can comfortably host for time being. I still think there will be a place for the super large models for more complex reasoning although how much will be due to the intrinsic knowledge in the weights and how much due to the plumbing around them remains too be seen.

And if course LLM’s are not going to be the end point of the search for AGI. Whatever their architecture they will still need copious amounts of compute.

You don’t think the creators should get paid? Do you know how much effort goes into making quality videos?

Ironically it’s probably because so many of the slop videos try and copy their style. However I would have thought YouTube would be more proactive with a channel with 25M subs.

Hmm I see nothing - I guess the posts don’t federate back to .ml?

I joined the .ml server because it was the dev server. Looking at the communities I subscribe to most are on other servers. Occasionally people ascribe beliefs to me based on my home server but so far I haven’t bothered looking for a new home.

When I first joined there were a bunch of popular instances that closed down so choosing a non-dev server seemed like a bit of a gamble. Same with mastodon servers.

On the other hand it’s legitimately one of the wonders of the digital world. A communally edited source of knowledge free for all to access. While it’s good to scrutinise foundations and how they are run I’ve not seen anyone propose a better solution so far. The data is freely licensed so forking is possible but it’s not really the technical details that make Wikipedia the resource it is today.

I don’t begrudge peoples difficulties with English. It has some positives like not needing gendered conjugation for objects and a fairly wide cultural exposure. However it freely steals words from other languages making it a bit of a mongrel. Also the spelling and pronunciation is entirely non-regular so the usual answer to people looking for rules to work these things out is “just know”.

Thanks. Are they related? I don’t understand any Asian languages myself but I do recognise difference in for example the Korean symbols and Japanese ones. But Japanese and Chinese ideograms look very similar to my ignorant eyes.

It’s super common if it’s any sort of open source project because the collaborators can be anywhere and English is a good baseline.

IIRC one of the big tasks when Star Office became Open Office (later Libre Office) was translating a bunch of the German comments to make the code more accessible.

I’m sure I’ve seen some majority french code bases as well. There are certainly academic hypervisors out of China which have comments in Chinese kanji.

They have their own flags, languages, laws and legislative assemblies.

I think the two principle arguements he’s making are open weights allow for misaligned models and these models are using distillation to more efficiently train.

I don’t really care about the distillation point - it seems it’s just hoovering up the internet with extra steps. It’s really up to Anthropic to deal with by enforcing their ToS. Sauce for the goose etc…

I can see the argument about alignment but it’s not a new problem. In the FLOSS world we generally don’t put restrictions on use on the code we release even if it might get used for bad stuff. It certainly wouldn’t hurt for the authorities to test open models to find out where the guardrails are but ultimately rather than banning the tech you need to enforce the law against the people that use the tech to do bad stuff.

Quite, I think her long career in genre films was one of the reasons Tarantino was so keen to work with her.

Just an early example of AI psychosis

I suspect it’s profitable in the abstract - and their accountants would be bad at their jobs if they couldn’t work out what utilisation rate you need to pay for the server runtime.

However how aggressively you amortise the cost of the training is the key, especially if you keep releasing new models every 6 months.

Invariably it’s having to scroll that’s the issue and you don’t want to be cross contaminating via your shiny black mirror.