FORGET Nano Banana! This is BETTER (And Lower Censorship)

Lovart + Seedream = Design insanity. Try it now!

Using Seedream 4 and Nano Banana is easiest on @lovart_ai.

Download my comparisons for free.

Thanks to Lovart for sponsoring this video.

Unlock 15% Off All Courses
Use code YOUTUBEFRIEND15 at checkout — limited time only.

Explore AI video, design, storytelling, and more.
👉

0:00 Intro – why this image model matters
0:29 High-res, censorship-light image generation
0:53 Ultra-detailed portraits and textures
1:18 Generate multiple consistent views in one go
2:07 Handling complex lighting and shadows
2:42 Professional-grade graphic design outputs
3:07 Photorealism with flaws – hands and digits
3:49 Blending graphics with photography
4:11 Perfectly consistent text rendering
4:37 Motion blur and cinematic composition
5:03 Realistic environments with subtle details
5:32 Fashion edits and photorealistic changes
5:56 Graphic design combined with images
6:23 Relighting and image editing
6:56 Object replacement and photo restoration
7:15 Complex layouts and multilingual text
7:55 Infographics blending illustration and typography
8:28 Ranked top on global benchmarks
9:41 Side-by-side comparisons with rivals
10:01 Portrait test – detail and aesthetics
11:21 City street scene object challenge
12:33 Wine glass fill test
13:35 Censorship stress tests – celebrities, violence, profanity
15:39 Text rendering poster comparison
17:48 How to use it today for free
18:39 Editing with natural language instructions
19:06 Generating Instagram posts with consistent style
19:18 Infinite canvas workflow for creatives
19:54 Who built it and why it’s important
20:18 Closing thoughts and next steps

💼Business inquiries
collabs@aisamson.io

📧Join my newsletter

36 thoughts on “FORGET Nano Banana! This is BETTER (And Lower Censorship)

  1. Over the past week I’ve been going back and forth between Nano Banana and Seedream 4. And in my use case (outdoor product photography) Nano Banana is the clear winner. It’s just more realistic and natural looking, while Seedream leans towards over saturated, over contrasted, stylized pictures that scream AI (good AI, nevertheless AI). I just wish that NanoBana had the resolution of Seedream.

    1. My thoughts exactly, i will revisit and test again purely due to the scale of the images, perhaps post production will help reduce that obnoxious over saturation and deep contrast.

    2. Also Seedream is great at editing existing pictures. Whereas Nano Banana excels at creating new ones… especially when using portraits.

    3. I find seeddream images are very ‘plasticy’ over saturated, and not very sharp. Similar to ChatGPT’s look. Not good. Maybe I’m missing something in my prompts but ii don’t get what I see these YouTube channels get.

  2. When they are new, they are always cut out to censorship, and as soon as they start making money, after a few months, then they start censoring. That’s what happened to every AI app I ever bought. At first, you do whatever you want, then when I get a little scratch, they start censoring it. It’s like these AI apps that say it’s free Just sign up then after a few images they start off in your plans . So that it’s just a carrot in front of your nose

    1. Another sad thing is that even the open source models require a computer so powerful that we will always be 10+ years behind the cutting edge if we want to run it locally.

    2. Yep, couldn’t agree more. Same happened with GPT, then Nanobanana. I can barely get any prompt to pass on Nano anymore. Really sucks.

    3. I can totally agree and relate… been doing this AI stuff for a few years now, since the early tools and they ALL go this direction. My workflow has changed as tools get locked down, move to subscription only, or just crank up the censorship to the point where even benign prompts will not proceed.
      I also find it funny where the out of date advertisements for some models / sites still say “Totally Free or No Sign Up!” only to be hit by a sign up or pay wall the moment you visit the page.

  3. I usually go through LMArena in battle mode to lower the censorship, and nano banana for everything else

  4. AI censorship isn’t prudishness; it’s a failure of context. Unable to distinguish the artistic intent of Goya’s Maja from exploitation, or an anatomical diagram from the profane, the algorithm defaults to blunt prohibition. The result is a paradox: we are building a revolutionary tool for human expression that is systematically barred from depicting the human form. It’s an engine for creativity that is blind to one of art and science’s most fundamental subjects: ourselves.

  5. Why is the 2000 years old Venus de Milo considered irreproducible by AI?
    Why is Michelangelo’s David also banned from depiction, when both sculptures are celebrated in the world’s great museums, admired by adults and children alike?

  6. Your just reposting Seedreams own example images. Try using the tools for yourself and see what results you get. The multiple views from one image does not work at all. if it did it would be a game changer, but like all of these tools they don’t do what they claim yet.

    1. And it’s bizarre the way so many YouTube channels are just blatantly lying for them.

      Paid, I guess.

  7. I just tried it. Wow. Unlike Gemini, he doesn’t spend time apologizing because he didn’t deliver the order. No, instead of apologizing, he delivered the goods. So much frustration with Gemini despite the good images it can create. But sometimes spending an hour there just to get him to do what he was created to do. Thank you for this wonderful discovery.

  8. Seadream doesn’t necessarily “blow away” Nano Banana, just gives you different styles and aesthetics like all the tools do, same with Reve (another good photo editor that uses deep reasoning) I use them all to try and get the best results. Frankly I slightly prefer the results I get with Nano banana but Reve is great for generative fill on aspect ratios. Censorship is a major nuisance and varies from site to site but it’s always something that gets worse

  9. This channel is about exploring how spicy you can go and how you can create a.i thirst traps

  10. Unbelievable, the way this man emphasizes. If I listen to him for too long, I start having difficulty breathing.

  11. at 12:14, the scene with the dogs, bus and bicycle.. something very strange is going on here, as both images have the same type of dog, bike, car and bus. the chances of this being random is unlikely..

  12. About the methodology of the censorship, I believe the standard for most platforms is two (or even more) layers. Layer one will refuse certain prompts outright. Layer two is analyzing the generated image with AI and checking for forbidden content. This second layer is necessary since image generation AI is not itself intelligent enough to actively avoid fobidden content without making it much less consistent. That’s the reason why the AI might react differently to exactly the same prompt. Somentimes the result will be deemed OK, sometimes it won’t.

  13. I don’t see anyone else pointing out where you put that circle and then smirked at the camera when doing so. Well done sir.

  14. Just think, we had nude statues and paintings hundreds of years ago and yet modern sensibilities or even medieval religious ones now deem them obscene.

  15. I like using the finger test for A.I. I just walk up to some random person and say: “Hey, pull my finger.”

  16. It’s SO CENSORED. Even with Midjurney I can create a person with a prosthetic arm but with this… “Unable to generate content due to policy limitations.” 🤣🤣🤣

  17. I noticed this when using nano banana. There are images that, just because I was using words like “claustrophobic,” were enough to prevent them from being generated. Just removing them was enough to generate them.

Comments are closed.