Видео

This worked for me! Thank you so much. ❤ I used generative AI on one small imperfection, didn’t realise it was going to flag my whole image as AI! Either way, I’m very glad that it’s brought me to your channel. Subscribed 🎉

Stripped the metadata for my latest post but it still got the "Made with AI" label.

The AI ​​interprets texture problems as if they were lighting problems. So they are still there but less noticeable

I just bought this and it didnt download to my computer

You did awesome!

So no speed retouch video? 😊

Loved being able to hear your thoughts and process for this!! Completely different world for me in photoshop but you can see how your retouch had intention behind it and it just felt more “real” 🖤🖤🖤 Can’t wait for more

I knew right away the left one was yours because it was more perfect. But yeah i am waiting till ai can do better or the same level of retouching as you before i use it. Which should still take years.

Well, looking at this, AI does not provide an end result, but rather prettified step for then manual polishing it to the final. And now i wonder about comparing AI (Evoto, Retouch4me, etc) with what you can do in kinda same time - run an action and loosely mask it. That would be closer to it's current application (and purpose).

the ai version looks use high pas, your version a lot more to d&b, d&b version looks more natural on the skin texture

Opening the flattened file in another app and exporting seems to work too. It strips the meta data.

So Far you are correct, we have played around with various Metadata readers and strippers. Every time we used a METADATA Stripper, we don't get this information... then again, all your current info gets stripped... so a Metadata editor would be a wiser choice and the options for Windows systems are FAR BETTER than on Macs.... Just FYI.

when there's no human skin, it's fair to mark as AI 😉

Hey Pratik, Thanks for sharing your thoughts on GenAI, it got me thinking. You made some great points about the incredible progress we've seen, especially with the new multimodal version of ChatGPT 4o. It's mind-blowing stuff, right? May I add my two cents? While LLMs like ChatGPT are undeniably impressive, I can't help but feel that they are sometimes "cheating". Sure, they can generate answers that sound super professional and knowledgeable, but sometimes when you really dig into it, the output is just plain wrong. They just know how to write really, really well. Or take ChatGPT 4o's quick reply: Have you noticed how it rephrases the question before answering it? I think it's a clever trick to make it feel faster, but in reality it still takes a few seconds to come up with the real answer. But we don't notice due to the rephrasing. I agree that we're making incredible progress on AGI, but I think the road to real Artificial General Intelligence is a two-part journey. In the near future, we'll probably see AI that's so good at certain tasks that it can take over many jobs. But it'll still be "faking it" to some extent, relying on clever tricks and sheer computing power rather than true understanding. The second part of the journey, the leap to true AGI, is probably years or even decades away. Sure, progress will accelerate. But I think for a long time, all AI will lack the general intelligence and understanding of the human mind. Plus, from a commercial point of view, it is more interessting to sell AI Bots or Agents to people rather then trowing all cpmputing power together to create one AGI. While I agree with you that the departure of OpenAI's AI safety team is bad, I think the big threat will not come from OpenAI. Bad AI could come from anywhere, especially from organizations or countries without strong democratic oversight, ethics, or a humanistic worldview. And we don't need bad AGI; bad actors with regular AI could use it to stifle dissent, manipulate information, and consolidate power. It's a frightening thought, but it underscores why responsible AI development is so important. But one thing is clear. AI is here, and we are going to explore every aspect of it. Cheers. - Phil

Maybe they resigned because they were made redundant by AI and were given the chance to resign before they were fired.

Just spitballin', but perhaps the company has come under the review of a state actor

Well the BIG question is WHAT Version of Wisdom is it being fed? Because before we get prematurely scared, we need to use logic to sort through the various sources of Wisdom, which will include many sources of Religion... which means we need to be able to feed it cultural information as well as political and governance from different societies.... which will butt heads with a large areas of many societies and individuals as it won't be "in alignment". I would have been much easier to implement this technology Prior to All the different Social Medias.... but the world is changing.... we will, sadly, in my perception, need AI to help us sort through all of this and as that tool develops, it may lead to a one world government or basis of thought..... then again... many SciFi movies and SciFi Books as well as psychology and religions have taught us.... it won't be unified in our lifetime.

Protection will be local network infrastructure not connected to the net.

Pratik we live in multidimensional simulation. Nothing is real.

I;, loving the creative side of AI and been looking if there are any future careers fir this,If so do you know of any future careers in this creative AI path

This is great, thanks! Would be great if you could do the neural photoshop stage as mentioned?

Is there a possibility to modify the script so you can toggle through the visibility of layers? Example; you select 4 layers and every time you press a shortcut one of the layers is visible, the rest is invisible. So layer 1 is visible, 2+3+4 are insivible. You press the shortcut; layer 2 is visible, layers 1+3+4 are invisible, and so on.

I agree with you when you are taken it your lol 😂

@pratik this is your name

I found your original picture of it just now in 2024 . Original day was 07/27/2022 @ 2:14pm

Affinity photo need this type of plugin. It can be done manualy, just a graduated layer and a mask with the right blend mode, but the number of point is what makes the difference here.

have you tried using simple scribbles to create an actual image?

No trial without paying 40? Waste of time..

the fact that they give no trial screams greedy fucks to me.

Very helpful and informative as always! Thank you so much Pratik and good luck!

I am trying to find the "PAY PAGE" WHERE IS THE LINK, thanks

Krea is better.

what's the largest size it produces?

Hi Pratik, thanks for the tutorial. I can’t see the link you mentioned in the video to get the plugin.

Thank you for the review! I tried it. Like other upscalers, the problem is hands, even at low creativity and high resemblance.

I can't hear the audio, the volume is incredible low

Sad.

Does Magnific offer a preview of your image before you commit to the upscale?

MAGNIFIC TH1EF

thank you! so much help

I like to blend the before (softer aspects) and after (added detail) together. Have you done any of this yourself?

Does it enhance old / black&white photos ?

good app...but a disgusting price. It's pretty easy to do the same thing for free with control net tiling on stable diffusion. it really isnt worth 40 quid a month for only 200 images

Not a word about feet and hands when upscaling. Does it mess up?

Hope the price comes down

I’ll def wait for one that doesn’t cost near $40 a month but amazing indeed

Its cool,but the only problem is, it only accepts low resolution pictures.

It looks great but the pricing is insane. It even costs like 2-3 times the price of midjourney

hard to believe people want to compare this to other Upscalers , this is on a another level but yeah 40 bucks is a lot , topaz LOL codeformer has that beat .

Thank you so much Pratik for the very informative tutorial! It would be highly appreciated if you make a tutorial on upscaling procedure, you were mentioning in your video! Good luck!