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

Nano Banana 2: Google's latest AI image generation model

blog.google

Post reply on HN
Print view View original

121–130 of 593 posts

Re: Nano Banana 2: Google's latest AI image generation model

#121

Post by GaggiX »

zug_zug wrote: post #102

I'm sure this has been written about but here's what happens long term - images are commoditized and lose their emotional appeal. Probably about half of us here remember photos before the cell phone era. They were rare, and special, and you'd have a few photos per YEAR to look back on. The feel of photos back then, was at least 100x stronger than now. They were a special item, could be given as a gift. But once they…

You can still buy a Polaroid, there is one factory left in the world able to produce the film required but they still make them.

Re: Nano Banana 2: Google's latest AI image generation model

#122

Post by tokai »

WarmWash wrote: post #24
nickandbro wrote: post #17

These image gen models are getting so advanced and life like that increasingly the general public are being duped into believing AI images are actually real (ex Facebook food images or fake OF models). Don't get me wrong I will enjoy the benefits of using this model for expressing myself better than ever before, but can't help feeling there's something also very insidious about these models too.

It's more likely than not that every single person who uses the internet has viewed an AI image and taken it as real by now. The obvious ones stand out, but there are so many that are indiscernible without spending lots of time digging through it. Even then there are ones that you can at best guess it's maybe AI gen.

Maybe not an actual argument for anything, but even before these image models everyone that used the internet had seen a doctored image they believed to be real. There was a reason that 'i can tell by the pixels' was a meme.

Re: Nano Banana 2: Google's latest AI image generation model

#123

Post by minimaxir »

jakub_g wrote: post #108

Since talking images, are there any AI models that can output real transparent gifs/pngs? And not a (botched) fake white/gray grid background that is commonly used to visualize transparency?

You can output to a plain background and use any number of tools to mask it.

Re: Nano Banana 2: Google's latest AI image generation model

#124

Post by electrosphere »

zug_zug wrote: post #102

I'm sure this has been written about but here's what happens long term - images are commoditized and lose their emotional appeal. Probably about half of us here remember photos before the cell phone era. They were rare, and special, and you'd have a few photos per YEAR to look back on. The feel of photos back then, was at least 100x stronger than now. They were a special item, could be given as a gift. But once they…

It reminds me of the Star Wars content thats come out recently - before there was the Original Trilogy which we all watched many times and the lines became iconic. Since then it's all become a mismash and blur of mediocrity due to over-exposure.

(except The Mandalorian, and I can't believe I'm using the word "content" :/)

edit: Totally forgot about Andor & Rogue One sorry, great film and two seasons of top-notch storytelling.

Re: Nano Banana 2: Google's latest AI image generation model

#125

Post by nathan_compton »

JOJESU wrote: post #118

I’ve been exploring this exact problem space from the angle of extreme constraints (single-digit MB memory, no cloud assumptions). I documented what broke first and why here, in case it’s useful: https://github.com/nullclaw/nullclaw

So this is an ultra-minimalist software platform to farm work out to enormous energy chugging AI models?

Re: Nano Banana 2: Google's latest AI image generation model

#126

Post by soared »

fasteddie31003 wrote: post #59

I'm building my personal home right now. The AI image models have been a game-changer in designing the look of the house. My architect did an OK job, but the details that Nano Banana added really bring the house up a notch. I just do hundreds of renders from the basic 3D models and I find looks that I like and iterate from there. We are implementing the renders from Nano Banana over our Interior Designers designs. We…

Same! I redid my backyard entirely and needed ideas. Gemini took a pile of dirt and gave me countless ideas, improved my plans, recommended materials, etc. a designer gave me two out of the box ideas that Gemini didn’t come up with, but it did everything else perfectly. (Designer said, put a patio out in the yard and put your table there, and take your ugly shed and make it the center of attention, since you’ll never succeeed trying to hide it)

Re: Nano Banana 2: Google's latest AI image generation model

#127

Post by ge96 »

minimaxir wrote: post #117
ge96 wrote: post #88

My naive question, can image generation make something novel eg. "show me a DNA structure that cures cancer" can it do that, or it has to have seen something before to generate it. Just think we conceptually know what a brushless motor design looks like and it's just pixels. I guess even if it did produce the image we wouldn't know what it means.

All image models can generate images that were not in its training dataset, but it can't generate reductive extreme cases like your example.

What about it is extreme? It's a concept, like "generate an xray image" eventually hopefully the cure to cancer could be represented as a simple molecule or whatever, I'm not saying I know.

Re: Nano Banana 2: Google's latest AI image generation model

#128

Post by yakattak »

gwd wrote: post #84
rm_-rf_slash wrote: post #45

Earlier quoted context omitted.

It can also backfire. AI slop ads and marketing material imply cut corners and poor quality products. If a bakery isn’t going to bother touching up its AI slop banner, I don’t expect their cookies to be great either.

FWIW I've never seen a correlation between a small company's website and the quality of their product. Slick website? Maybe they care for their craft, maybe they're all marketing and no content. Website stuck in 1998? Maybe they're sloppy and don't care; maybe they care about their core product, not a slick marketing brochure. I don't see any reason AI would be different in that regard.

That’s true. I think it’s more of a problem of getting someone in the door. Anecdotally going to art festivals I’m much more likely to enter the booth of someone who has handcrafted marketing over the person who has generated marketing.

Re: Nano Banana 2: Google's latest AI image generation model

#129

Post by coffeebeqn »

pousada wrote: post #69
whynotmaybe wrote: post #41

Earlier quoted context omitted.

>fake OF models Soon many real OF models will be out of job when everyone will be able to produce content to their personal taste from a few prompts.

You can’t really because these powerful models are censored. You can create lewd pictures with open models but they aren’t nearly as good or easy to use.

I’ve seen some very high quality NSFW AI video in the last few months. Those models are not far behind and the search and training space for porn is smaller than being able to generate anything

Re: Nano Banana 2: Google's latest AI image generation model

#130

Post by fasteddie31003 »

bartman wrote: post #97
fasteddie31003 wrote: post #59

I'm building my personal home right now. The AI image models have been a game-changer in designing the look of the house. My architect did an OK job, but the details that Nano Banana added really bring the house up a notch. I just do hundreds of renders from the basic 3D models and I find looks that I like and iterate from there. We are implementing the renders from Nano Banana over our Interior Designers designs. We…

Can you write a bit more about your workflow? I've been thinking about doing the same, but since I'm very non-interior-design minded have struggled to ask the right things. Like... What are your inputs to the model? Empty renders of the space, or more fully decorated views/ photos? Do you have a light harness around this to help you discover the style you like and then stay consistent with it? Do you find that giving…

I started with sketchup to make basic floor plans and house shapes. I had a rough idea of the style of the home. I picked "Transitional English Estate" since the build site is out on a farm that sorta looks like the Cotswolds. I used AI in this process to get rough renders and feedback on the floorplan. I then took that basic floorplan and house dimensions to a Draftsman who did a lot of tweaking to get it up to code and fix issues. I got his plans and took it to a Sketchup Pro on Fivver . They made a detailed sketchup model. I then took that model and took screenshots from different perspectives and tweaked the prompt to get renders I liked. These changes were reencorprated into the blueprints. I did the same thing with the interior. Took screenshots from sketchup and put them into AI and tweaked the prompt. https://imgur.com/a/lSIYTYr
Post reply on HN