hafriedlander

I have tried to explain "you're not mad at generative AI, you're mad at late stage capitalism" before.

Most people aren't really willing to smash the state though (I understand, that's where all my stuff is) so look for less drastic ways to protect themselves.

Somewhat tangential, but the Krita community and core team have been pretty explicitly anti-AI. https://krita-artists.org/t/change-in-policy-for-topics-rela...

(I am part of a group that builds UI on top of open models, but we stopped working on our Krita version for that reason.)

Don't discount the value (or difficulty) of a great user interface

There are multiple servers that provide the backend "meat" already: gyre.ai (my personal project); Automatic1111 webui; InvokeAI; plus all the hosted APIs to name just a few.

There are no UI interfaces I'd consider great. I think Flying Dog aistudio[1] is the best (although I'm biased), but still early days. The UI part of WebUI is ... functional(ish).

DreamStudio / StableStudio is super basic in capability at the moment, but it's a space with tons of room for improvement.

[1] https://www.youtube.com/watch?v=FcGlA38M52M

(Edit: markdown doesn't work here)

Yeah, neither view is a perfect fit. Another example is vision transformer backbones, where a common generic base weight is used to fine-tune all sorts of different processes (segmentation, image to text, etc). The terminology (and licenses) haven't really kept up.

A properly unencumbered model would be my preference too. The community generally seems a bit laissez-faire with license compliance though, so the restrictions currently don't generate much push back. (Plus it's not totally clear that you can copyright model weights at all, given they're the output of an automatic process).

Model weights can't really be described as source code though. The equivalence isn't exact, but I'd describe the weights more as the compiled binary, with the training data & schedule being the source (which is sort of under an open source license, with the complication of LAION's "it's just links"). The fact it costs $1 million to "compile" isn't relevant.

This isn't to defend Stability particularly though - they've been getting increasing slow and restrained in their model releases. Charitably because they're attracting a lot of heat from political and anti-AI aligned groups. Uncharitably because they've taken a lot of funding now.

(Edit: typo)

Since I re-implement the Stability APIs with gyre.ai, it should "just work" for a local inference option.

I haven't tested though, and the API spec has lots of under-specified bits, so could be some stuff I need to tidy up.

(Alternatively, despite the post content, in the github repo there is already a webui plugin, no idea of current state).

Since everyone's jumping on the plug wagon, allow me to throw on my project, www.stablecabal.org.

We've got a GRPC server that's backwards compatible with the official grpc.stability.ai but adds advanced features, a Flutter based infinite canvas web client in heavy development, a Discord server and a Krita & Photoshop plugin (all except the last are open source under various licenses).

The outpainting the server supports is (IMHO) the best I've seen, and recently got some a tiny glimpse of attention when combined with the Photoshop plugin - https://twitter.com/NicolayMausz/status/1577767106384433156

They aren't making any more land. New Zealand has very lax laws regarding property ownership compared to many countries. No capital gains tax. Very limited ownership restrictions.

To pick on a common scapegoat, Chinese citizens can buy a house in New Zealand, so as their economy improves it drives our housing prices up (same level of supply, greater demand), but I can not buy in theirs.

I think it probably only has a small impact on housing prices (a much bigger problem is the political impossibility of any measure that drops housing prices) but I don't think you can say it has none.

I would personally be in favour of introducing a capital gains tax (but, as a renter, I would say that).

I'm a New Zealander, and strongly pro-immigration. Given how many emigrants there are (including, currently, me), if nothing else we'd run out of people pretty quickly without it. And I don't believe integration is currently that much of an issue in New Zealand. However:

> "Of those who do choose to move to New Zealand permanently, analysis of the New Zealand General Social Survey show immigrants integrate well. They are less likely to claim a benefit, more likely to be employed, and their children have better education outcomes than native born New Zealanders."

None of those show integration in a way I recognize as valuable culturally. They would be equally true if we were measuring a segregated expat enclave in a developing nation, but that's simply a factor of leveraging existing power structures.

If you don't believe Jeff Atwood, I'm not sure why you'd believe me based just on my anecdata, but at a previous job I gave a simple filtering programming task to interviewees (not exactly fizz buzz, but very close. Single loop, multiple 10-line solutions possible).

Out of 10, only 1 solved it well, 2-3 solved it acceptably (only edge case bugs), the rest never solved it, or had obvious bugs or code that would never pass peer review.

Of course, I could just be a shitty interviewer, but I always had a (changing) peer present during interviews, and I hope we were an honest enough company that they'd tell me if they thought I was.