@_chaotix a more accurate analogy would be eating a burger and halfway through the server tells you it was 3D printed synthetic meat, grown in a lab culture.
now you are still surprised, but if you like how the burger tastes, it’s all the same at the end of the day.
Over 1million views now.
Thank you everyone who listened, enjoyed and supported it!
Thank you to all the constructive and fair criticism.
And thank you to all the haters.
You pushed it in the algorithm by commenting and QRTing it, bringing it to more people.
It would never have gotten this far without the hate train.
My biggest fear is creating an echo chamber for myself, that's why I don't mass-block the hate and why I try to engage with some of it in good faith.
I know it can make me look silly, and I know it probably isn't in my best interest, but I genuinely believe there are minds who will be changed if you engage with them.
I'm not looking to make people like AI. My ultimate goal would be to get the haters to hate it for the right reasons.
Because yes, there are elements of AI that are parasitic and morally questionable.
But AI is so much more than that.
It brings people together, it gives people a way to detach themselves from corporations and create what they want and it really can be used to create amazing works of art.
By showing what AI can do, we can drown out the bad actors. We can teach people how it actually works. We can show the difference between something we put care and effort into VS "grok, make me art".
The power is in our hands. Not corporations. Not rich assholes. Not even the companies who pioneer these models.
Open weights, open source AI is the future. Creation with no limitations.
It's up to us to make sure it's used for good.
yeah you cant just clone it. you need to bundle the whole tool suite, offer a one-click installer or better yet a creative studio, and then SEO it to high heaven so you cant type “edit video” without the top 7 links blasting your site link…
@m4h007 AFAICT, they didnt even care enough to claim these as finished work, which is strongly suggested by dumping them on github instead of yk, publishing them.
This results dump is just a byproduct of using their internal evals, as they are hard problems with well-defined goals.
@metatransformr@AIOnlyDeveloper Theres a desktop app but its still browser-based.
Community members have also published mods for file management, diffs, PRs, etc. to give it more of an ADE feel than the base app
if we solve physical AI, whats the first sport to feature a full-AI league?
olympic track and field was entertaining with the recent memes etc
i doubt itll be basketball or football (either type).
boxing seems most likely (s/o Real Steel)
@Nepcaligo this, at scale, would cause a complete death spiral of arts patronage.
the most equitable outcome for everyone involved is people use AI for mass production, short deadlines, or disposable use cases
commissioned art becomes like bespoke leatherworking or jewelcrafting, quality
@MicrowaveArson the core disagreement ive had since the first diffusion models years ago was…
im glad AI exists. It is (or will be) the sum total of human achievement, expression, and effort.
im glad all my work is a part of that. its like the Voyager inscription, a living distillation of us
Game design takes images and sounds and code as inputs and creates a work of art by compositing them together in creative ways. So I'm OK with using AI to generate the parts, since my real art is the design.
I can see how a visual artist or musician would not be thrilled by it.
If a human directs a machine it can still qualify as art. That’s why photography isn’t excluded. It takes a lot of mental gymnastics to pretend that the argument being made is that the AI model is conscious. Put it in its proper place as a tool, and the house of cards anti-AI stance crumbles immediately.
@Y40IFRQTTING wait what?
LLMs dont just “statistically infer the most average type”
i mean, they CAN do that, but thats just using them wrong.
as we see in math circles right now, LLMs dont only interpolate, they can extrapolate
@squeakytoy i think it’s just damaging to people’s mental state and society if we promote verbal abuse directed at anything, even something as inanimate as an LLM.
i wouldnt want to interact with the people who shout profanity/slurs at stoplights/their cars when theyre running late
@JustCamHayden yeah, but to be fair the Internet is a smaller scale economic revolution than others in history.
AI (if trends continue) is on pace to be an Industrial Revolution scale disruption, condensed into a shorter timespan
next few years will be a wild ride
at the end of day, I have the same respect for creatives and programmers who refuse to use AI as I have for bespoke leatherworkers
its impressive and will always be an admired facet of society
but… we moved on to mass production for the majority of products for good reasons
@NickEMoran@camhberg i think the argument for doing it would be stronger yes
there are still issues as to whether abusive prompts in the training sets make alignment harder
@NickEMoran@camhberg no because this has a purpose. its entertainment for some people. its revenue generation for the creator.
the key difference is this guy isnt coming from a place of unstable emotions and wrecking things.
its the diff btwn a controlled demolition and a terrorist attack.