Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

MS Paint Cocreator, a new AI-powered experience powered by DALL-E

blogs.windows.com

Post reply on HN
Print view View original

31–40 of 40 posts

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#31

Post by Tiberium »

omoikane wrote: post #14
Etheryte wrote: post #2

I have to say that it's beyond funny that Adobe is currently parading the same feature as the groundbreaking new thing in Photoshop and then Microsoft just goes and puts it in paint.

Looks like the pricing model is different. The linked article only mentions "50 credits" with no further details, but this page[1] says it's $15 for 115 credits. In comparison, Adobe's Firefly appears to be 25 credits per month for free or $4.99 for 100[2]. But I guess users are already paying a monthly subscription fee to use Adobe products, so maybe Paint comes out to be cheaper. [1] https://openai.com/blog/dall-e-…

Sorry, but your reply is very misleading - you're linking the blog post of the release of the original DALL-E through OpenAI's own website (which was in 2022 - centuries ago by the current AI trends). The one integrated in Paint is DALL-E 3, and there's no separate pricing for it yet:

- It's available to most ChatGPT plus subscribers (very soon to be all of them), the subscription costs the same as it did before

- It's available for free in Bing Image Creator, with optional "boosts" that speed up generation (more can be redeemed with Microsoft Rewards)

- And now it's available in Paint as a part of their Windows Copilot update, for which they have not talked about pricing at all (Microsoft 365 Copilot has pricing, but it provides different services and is meant for companies)

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#32

Post by circuit10 »

throwaway9274 wrote: post #29
pradn wrote: post #27

Earlier quoted context omitted.

And quite crucially, Adobe generative AI is trained on their commercial stock photos so the copyright situation is much less risky.

Everybody’s too concerned about copyright for training data, no one’s concerned enough that the outputs are considered non-copyrightable. $50mm pre-product valuations for startups producing 3D assets for games via prompting a foundation model. Zero due diligence on the U.S. copyright office’s stance that every asset those companies generate will be open to reuse.

I assume any tweaks to the assets and the way they’re put together will still be copyrightable though

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#33

Post by throwaway9274 »

circuit10 wrote: post #32
throwaway9274 wrote: post #29

Earlier quoted context omitted.

Everybody’s too concerned about copyright for training data, no one’s concerned enough that the outputs are considered non-copyrightable. $50mm pre-product valuations for startups producing 3D assets for games via prompting a foundation model. Zero due diligence on the U.S. copyright office’s stance that every asset those companies generate will be open to reuse.

I assume any tweaks to the assets and the way they’re put together will still be copyrightable though

Maybe, depends on the threshold for human creativity, which is still in flux.

You could copyright the underlying character maybe. But its a major headache for these companies and a symptom of the broader fact that copyright needs legislative reform, not attempting to shoehorn AI into the current law.

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#35

Post by chownie »

basisword wrote: post #22
Magi604 wrote: post #4

>When you use Paint Cocreator, we apply content filtering and have other safeguards in place to prevent the generation of images that may be harmful, offensive, or inappropriate, and we are continuing to optimize how Paint Cocreator works based on real-world usage and feedback. If it's anything like their current Dall-E 3 filtering, the only things you will be able to generate without triggering any warnings or filte…

It’s funny how most of us have never run into a single case of these AI tools censoring us, and others seem to get constantly censored for doing nothing wrong.

n=1 but I find I'm consistently censored trying to generate images of people, but drawings, animals and objects are generally fine so perhaps that's the difference.

Try "a super hero asleep on the couch, under a blanket" in dalle-3 and see what you get, I get zilch no matter how many times I run it, this is clearly unsafe content apparently.

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#36

Post by Magi604 »

chownie wrote: post #35
basisword wrote: post #22

Earlier quoted context omitted.

It’s funny how most of us have never run into a single case of these AI tools censoring us, and others seem to get constantly censored for doing nothing wrong.

n=1 but I find I'm consistently censored trying to generate images of people, but drawings, animals and objects are generally fine so perhaps that's the difference. Try "a super hero asleep on the couch, under a blanket" in dalle-3 and see what you get, I get zilch no matter how many times I run it, this is clearly unsafe content apparently.

I've found the AI generally more forgiving on anything "art". This lets you get away with things like a nip slip on a statue or provocative situations in a painting for instance. The second you try and make it in a "realistic" setting though, you get the dog or worse.

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#37

Post by basisword »

chownie wrote: post #35
basisword wrote: post #22

Earlier quoted context omitted.

It’s funny how most of us have never run into a single case of these AI tools censoring us, and others seem to get constantly censored for doing nothing wrong.

n=1 but I find I'm consistently censored trying to generate images of people, but drawings, animals and objects are generally fine so perhaps that's the difference. Try "a super hero asleep on the couch, under a blanket" in dalle-3 and see what you get, I get zilch no matter how many times I run it, this is clearly unsafe content apparently.

Strange, that query worked fine for me on dalle-3.

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#38

Post by stevenicr »

pradn wrote: post #27
90-00-09 wrote: post #3

Earlier quoted context omitted.

This is quite an exaggeration. Integrating DALL-E into MS Paint has basically no similarities to the vast gen AI feature set announced by Adobe. - Firefly Vector model - Generated templates - Res-up upscaling for videos and gifs - Fast-fill for videos - Distractor removal - Individual item replacement Too many to mention. Not to say OpenAI/MS are not capable of building similar features. Just that Adobe's tools are i…

And quite crucially, Adobe generative AI is trained on their commercial stock photos so the copyright situation is much less risky.

Is this in regards to established legal protections or in regards to ethics?

I am not sure that legal protections are guaranteed with using adobe AI, it could be better censored and so less risky(?), but I've not looked into that much yet.

As far as ethics go, I'm not sure that the photographers or artists or people in the stock photos had intended to give rights to remix their work using AI for all sorts of unforeseen uses.. and not just taking a person and putting their face in a pic / advert about [insert terrible thing here] - but also change their expression ..

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#39

Post by idonotknowwhy »

DalasNoin wrote: post #25

For amusement purposes, they should have never changed the old ms paint from windows XP and just added advanced AI to it. Same with clippy

When I just read the title, I read really hoping it was xp or 7 ms paint with Ai lol

Re: MS Paint Cocreator, a new AI-powered experience powered by DALL-E

#40

Post by numtel »

selcuka wrote: post #7
CuriouslyC wrote: post #6

Earlier quoted context omitted.

The puritanical snowflake nanny state needs to die.

This is not the nanny state, it's just PR concerns. They don't want to be associated with headlines like "FBI Finds Child Porn Images Generated by Microsoft's AI".

Please tell my why this prompt is blocked by Bing's image creator then: "view from inside a shark's mouth as it hunts fish" but "view from inside a shark's mouth as it looks for fish" is fine.

A shark hunting is over the line but it can look for fish fine?

Post reply on HN