TheMiddleMan

Very nice!

I wish you had a page listing which vendors you currently integrate with. I was able to find it in the docs though, eventually.

I do wish they had more examples though.

Bing doesn't show my website either and my product doesn't compete with M$.

Yeah what's considered a copy or not is a grey area. Here's a good example of that: https://news.ycombinator.com/item?id=34378300

But artists have been making "in the style of" works for probably millennia. Fan art is a common example.

I suppose the advent of software that makes it easy to make "in the style of" works will force us to get much more clear on what is and isn't a copy. How exciting.

However, I don't see how the software tool is directly at fault, just the person using it.

I believe Copilot was giving exact copies of large parts open source projects, without the license. Are image generators giving exact (or very similar) copies of existing works?

I feel like this is the main distinction.

An artist can look at images for reference, and draw something new inspired by them. Why does it matter if a software tool can do this much faster?

If the artist makes the image very similar to one of the reference photos, it may be a copyright violation. It doesn't matter if the artist used a pencil or software to create the new work.

Current AI image generation does, however, make it easy to unknowingly violate copyright. If it generates an image similar to something else out there you wouldn't know.

I don't know much about copyright law though, am I wrong?

Neat, could be useful for turning mockups into html. I wonder if you could also generate some reasonable ids/classes, with the classes matching on related objects (in a list, etc).

Blind spot in the training data?

No AppSheet?

Hope to see dark mode someday for the leet programmers.

Scammers also create ads with fake support contact phone numbers for businesses. People call the scammers and they act like the company and run their scam.

Google also allows many deceptive AdSense ads, I constantly have to block ads that run on my website that are nothing but a big "Download Now" button which lead to some malware.

There's a big difference between learning and memorizing.

If the AI is "learning" how it works by studying public code then using its knowledge to create, that's okay.

But if it's just memorizing code and reciting it back, not okay. Just like if a human were doing this.

Of course we don't currently have ways to know the difference [that I know of] since AI is a black box.

Interestingly, current AI is not capable of truly understanding how code works and how it will execute, so it has to learn in it's own way. I suspect it can learn what valid syntax is, but I doubt it is aware of how the code will execute.

It's possible this is just a case of Overfitting. https://en.wikipedia.org/wiki/Overfitting

I don't think privacy is the only application. Imagine a bank using an algorithm like this with their security cameras to reduce the number of guards needed on premises. I could imagine thieves targeting their system with something like this.

Or maybe I've been watching too many heist movies...

Maybe we should just make it common practice to do: Array.prototype.flatten = Array.prototype.flat

Similar to "* { box-sizing: border-box; }" has become standard.

If they can't add an official .flatten for backwards compatibility now, we can assume they never will, right?

Would love to see this turn into a cross-platform/cross-hardware AI art GUI+backend with an open source license.

Speaking of which, what is the license on this? (the electron app)

Prompt is: monkey astronaut, bright, bright, bright, bright

(I was experimenting with repeating words, which does seem to amplify the effect each repeat with some keywords)

Seed is: 948574399

Ah I see, it makes bigger leaps each step to try to get to the same end result in less total steps. That makes sense, assuming I have it right.

Interesting. I suppose GPUs could calculate things differently. I just checked, I can rerun both 40/80 step runs and the final images are bit-identical to the first runs. So at least in my scenario the same parameters are deterministic, but changing the number of ddim sampling steps changes the result.

Maybe it's doing something fancy with the total number of steps, beyond just stopping after the count is reached.

Another gallery with 80 ddim steps: https://postimg.cc/gallery/b1kn7yd

Thought I'd share this for others interested. I've modified txt2img to save an image after each step. (actually quite easy as you can specify an img_callback to sampler)

Interestingly, both of these runs are using the same seed and prompt, yet they yield different final images, the only difference is the number of ddim sampling steps. I'd love to understand why if anyone has any idea.

Wonderful! What a great product, well done. Thanks.

This is great, I can stop emailing markdown files back and forth with my content writers.

Would this work as a sort-of wiki, for crowd-sourced documentation and translations? Or is it a bad idea to let the public edit things?

You say this like there aren't alternatives. We can find a way to grow crops that isn't harmful to people.

The iPhone still doesn't support the fullscreen api. It makes it almost impossible as a web game developer. Extremely frustrating.

Bam! https://playingcards.io/game/reversi

Good question. Both Backbone and React provide a way to build UI components. React Components are way easier for:

1. Having nested DOM elements in your components. Backbone really wanted you to just have one element per component, and then use jQuery to add children as needed.

2. Adding event handlers and other HTML attributes. JSX syntax means you can basically write HTML in your JS file. Backbone made you add event handlers manually.

3. Nested components. React lets the parent simply specify what children it wants and React automatically creates/reuses/destroys component instances for you. This was all manual in Backbone.

The bigger advantage though, is how the state is typically organized in React apps.

In Backbone, app state was sprinkled throughout the component hierarchy, and often you would have do event spaghetti up and down the hierarchy to update some state held by another component.

Typical React apps use a state container like Redux, the UI is "projected" from the state, and components just send changes to the state, aka "unidirectional data flow".

That said, React is not the best for games. Maintaining 60 fps can be difficult and we've done some fairly heavy optimization to try to keep it smooth.

The bottom area is the "hand", cards placed there are only visible to each individual player.

Note that some games like Checkers have the hand turned off by default.

Yes, you got it, Fischer-Yates shuffle via lodash. Here's their implementation: https://github.com/lodash/lodash/blob/master/shuffle.js

Creator of the site here.

We use Firebase for Auth and file storage only. Game updates are sent through Socket.io in order to have better control over them, and also Firebase charges per update which is a lot for games.

Hi, creator of the site here. Could you give a bit more details about specifically when it flips by accident?

There's a threshold that objects need to be dragged beyond to start moving, which is different for touch/mouse.

Also, what device/OS are you using?