general・🧺

[nimble falcon]

Google is search, Google is a misspelling of googol

🧐

[neon pagoda]

I don't know how to use imageFX to create a picture😰

[sick belfry]

@outer wagon does veo2 fast mode sacrifice quality?

[solemn lotus]

You can ask Gemini for pointers or give it a vague idea of what you want to get.

Be specific that you want help making a prompt or it will generate the image.

If you do use Gemini for making the pic, you can ask for edits but the output eventually will lose track of what you want - whisk is a better tool for that.

What would be better? And do you mean you all, the end users, or you all, the Google team? Or both?

Or what would better look like I should say

[woeful moat]

guys adid anyone get access to the ai mode on google?

[meager laurel]

I never expected Character.AI to rank in the top 3 yet here we are 😬

Never heard of janitorAi tho, gonna take a look 👁️

Oh it’s

Yeah

Today I learned something new then LMAO

[cerulean mist]

guys from your opinion when veo 2 will be available for all people

[meager laurel]

I used to use quillbot but nowadays you can get what quillbot have even in Gemini

Check out civitAi

Ah I thought they have video generation model for use

[solemn lotus]

Every...... special interest ... certain folks ask after here have a thousand options over there.

[stuck ferry]

I can see that

[solemn lotus]

Mochi, I think, added end of last year.

[woeful moat]

Frfr a

[astral minnow]

is it me or Animate button is gone from Whisk?

[stuck ferry]

Let me check

[stuck ferry]

Yep it seems they're working on it I don't have it on my end

[astral minnow]

i was so confused

[meager laurel]

Oh

They’re cooking something 🔥

[solemn lotus]

It certainly feels like they did a bunch of in-house alpha development while we talked in circles. I'm excited for updates to involve more public facing engagement.

[vital trout]

Oh wait wrong channel

[young stream]

one of the longest ive ever had generated @notebook

[dry bronze]

Can you confirm whether this is in the early access environment? It's currently only enabled on the URL that includes "early-access"

@stuck ferry could you check too?

[stuck ferry]

Yes one moment

[stuck ferry]

Okay so it seems you have to use specifically that link in order for it to work. It's not accessible when navigating to it through the homepage

But yes otherwise it does work you just have to use the Early Access link

[dry bronze]

That's right!

Thanks for checking

[stuck ferry]

I've had longer, though at some point when it exceeds around 40 minutes it has the tendency to repeat itself a ton

[dry bronze]

No it doesn't. And for style it only uses one image at a time

[astral minnow]

I think you are right, it is available in early access link!!!!

[weary cave]

#off-topic・🐢

[nimble falcon]

[weary cave]

Don’t remind me of that please, thanks!

[nimble falcon]

Well don't be policing that's the mods job

[whole hawk]

subject also can effect style but not as hardly

[weary cave]

https://tenor.com/view/shut-up-shut-up-be-quiet-funny-gif-19370091

[nimble falcon]

[weary cave]

Stop

[narrow chasm]

🧸

[rotund lark]

Tapping on daily listen doesn't do anything I just saw the option for the first time and it's not working.

[rain warren]

quality on that came out very crisp

[woeful moat]

ik happened with me. i couldnt do anything other than deleting the whole data of google

[dry bronze]

yes, we appreciate the community helping out

though the notebookLM chatter, makes me think we might need a channel for it

also, please remember to treat each other kindly and with respect (rule #1)!

[solemn lotus]

Channel for notebookLM would be great

[outer wagon]

we had one with minimal engagement since nblm has its own server!

please emoji react to this message if you'd be interested in using a notebooklm channel ~

[whole hawk]

is it possible to add post at start here with the list of google servers? BTW @solemn lotus made a good summary which worth to post there too. 🙂

[nimble falcon]

Self moderate means watch what YOU do.

[solemn lotus]

I also found it murky around "me on me" or "us on us" so I consulted reference when it was first posted.

[frigid stump]

Is there any technique or program for generating multiple, different images based on parameters in a spreadsheet import?

Edit: that doesn't include coding. Asking for it before I try a complicated route

[solemn lotus]

Canva Bulk Create will let you use CSVs to better visualize data . User friendly is their thing so check if that lines up.

Social media manager at work is next door to me, doesn't like computers, but love that thing.

Or did you mean "I have twenty unique prompts. Make me twenty unique images." type of automation ?

[frigid stump]

I was talking about unique prompts, or a similar "base" prompt with another column for variations.

Now that you've said it, though... I think using it for visualization would be beneficial with the way my brain works... Hmmm

[solemn lotus]

Are you into coding at all? Python is pretty easy to get installed.

Last night in about an hour Gemini and friends helped me streamline my old video combiner script. I made that 3 minute video I posted with it.

[frigid stump]

I need to be with the hobbies I'm currently pursuing, methinks.

[solemn lotus]

Fair. I'll keep an ear out.

[weary cave]

Blud left the server 😭🙏

[gray saddle]

Isn’t this impressive? Making an Entire drill beat with just text prompt . No DAW involved.

[swift jolt]

Guys just asking for a friends, does Google labs accept voulunteers for google labs under 18 years?

[ivory hornet]

As an employee?

[swift jolt]

no just simple ones that they want to help

[ivory hornet]

idk

maybe not? Since you’re under 18

but you can still help people regardless if you get hired or not

[swift jolt]

https://brainatica.website/

[swift jolt]

👽

[ivory hornet]

@outer wagon scam link in all channels

[rain warren]

yeah he probably fell victim to it unfortunately

no one click that link!

[frozen jewel]

GEN AI Video / Technical Lead ($2000 USD) - ComfyUI Focus

Hey everyone,

We're a film production company rooted in short format videos (mostly ads and music videos). We've just been awarded a project for a client that we're approaching as a research initiative. Our creative team is ready, and we have a good grasp of both the potential and current limitations (consistency, physics, quality) of Gen AI. While our initial plan involved paid services, we're now eager to explore the freedom offered by locally run models inside ComfyUI.

That said, we're looking for a GEN AI Video / Technical Lead with experience setting up and optimizing workflows inside ComfyUI. The focus will be on achieving control, consistency, and high quality for image and video production.

This is a real-world project, and we have a budget for your work time, plus ample investment for cloud-run GPUs. Our primary goal is to push the boundaries of what's possible in terms of quality and creative control.

We are based in Germany but are fluent in English and can adapt to various time zones.

To express your interest, please DM or email your portfolio to [e-mail].

Portfolio: https://www.upperfast.de/
Instagram: https://www.instagram.com/upperfast.de/

Thanks everyone!

[swift jolt]

🍏

[ember atlas]

Google is very good

[swift jolt]

Gemini is good, and respective deepmind projects.

I do not know of other alphabet properties however

sad is it not

[weary cave]

@bright pendant

[swift jolt]

branch based internal tools

[swift jolt]

If you are good, and of good, then have a good day.

otherwise, like, 🪭

these keyboards are so nice. thank you codename atlas team.

[sinful ravine]

Hey! how can i use the model gemini-embedding-exp-03-07? it cannot use it on experimental nor preview :/

[carmine zealot]

Hello world

[terse locust]

hi

[swift jolt]

Yes

[swift jolt]

isolated system

[swift jolt]

hi sorry for the ping but when u request for a data export when do they usually get back to you?

@summer loom

[eternal carbon]

amazing

[blazing hamlet]

HI

[cyan bolt]

[sick belfry]

Please make veo2 work on phone

[swift jolt]

שית

[weary cave]

<@&1173646484311384114> OMG GOOGLE JUST FIX THIS, YOUR GONNA LOSE MOST OF YOUR USERBASE

[oblique plover]

I find ImageFX to be one of the best image generator at the moment, if not the best for most categories of visuals. Though indeed lately, and today a few times, I have experienced similar episodes just as 'described' by your screenshot, within android 13>chrome and Win 11>Chrome.

[weary cave]

So sad

[meager laurel]

so sad

anyway

[weary cave]

STFU

[swift jolt]

soooo saaadd

[rain warren]

looool

[neat onyx]

woaaaah

[neat onyx]

abbreviation for Student From University : }

[swift jolt]

Dude mad

[obtuse trellis]

fal just got veo2 img 2 vid hope it's coming to videofx?

[meager laurel]

It's actually on the main VideoFX site 👀 it shows up for like quick 2 sec then it changes to Text to Image to Video, it seems like they're still getting things ready for early access users.

[obtuse trellis]

Oh awesome 😎 yeah I didn't notice that, good catch

[rotund lark]

Done the same and still didn't work.

[stable matrix]

hello guys how to get videofx ?

[ivory hornet]

join the waitlist

https://docs.google.com/forms/d/e/1FAIpQLSeC6n1KQlaqRNUGNuNRt5Q7YeoyXsq828niw2ZvIoAtW1FtYQ/viewform?resourcekey=0-qDKZCeB4G9nS9dttXGdnHQ

[stable matrix]

I have already joined, how long should I wait?

[ivory hornet]

The was recently an application couple days ago but they closed it

[stable matrix]

thx Very much

[ivory hornet]

there’s no eta

[meager laurel]

And a few hundred more after the early access “ discord member edition “ was opened

I just wish more people understood that early access isn’t just for fun, it’s really about giving feedback to help improve VideoFX before the official release! 🙂

[meager spire]

The image to video is very interesting

I already had some tests with it earlier. People will love it 1000 times over

[dusky fern]

One message removed from a suspended account.

[grizzled sail]

Yepp

[swift jolt]

Can we get Dyno work again?

[swift jolt]

where'd you test it?

[meager spire]

@swift jolt originally on fal.

[sinful ravine]

Gemma 3 and the ShieldGemma 2 looks promising; will test them out

[stuck ferry]

Yes! I’ve tried Gemma through AI Studio. The multimodality in an open-source model makes a big difference

Very impressive ELO score too

[swift jolt]

Hey @outer wagon if a person applied to google labs early testers, when do we expect to get an answer back?

[real grove]

when will whisk use native image gen considering Gemini now has native gen in public preview, would be cool

[sand lodge]

We can do that. We would really love to implement this tool into our workflow and the more early we understand the logic of the model the better results we can get later on and the better your product can be. We can implement it into ongoing and commercial use projects that can give quite a good feedback to you. I have attatched a side-by-side commercial we made with mostly kling and image fx. We also work on many experimental projects as an incubation hub so we can push the boundaries of this product.

[vital trout]

i can give feedback 💯

did u guys see that gemini can modify images now

unfortunatley this isnt modified

[old night]

How did you get it to use the same image?

[frozen pendant]

thats not possible with gemini

[vital trout]

yes it is lol

[vital trout]

aistudio.google.com gemini 2.0 flash expiremental

[frozen pendant]

there isnt a way to do it on gemini

[vital trout]

then what am i doing here

[frozen pendant]

😮

i tried it said "I'm designed solely to process and generate text, so I'm unable to assist you with that."

i cant get it to do anything

[old night]

It works in the Google AI Studio. https://aistudio.google.com/prompts/new_chat

[meager laurel]

OHH so thats what the new update on aistudio was about holy cheese crack3er!!

i need to try this

This is amazing ahahaa

[vital trout]

I put a sticker on my cup with gemini 2.0 🎉

[meager spire]

How exactly do you achieve this?

[meager laurel]

https://aistudio.google.com/prompts/new_chat from here

u should see the "image editing"

[meager spire]

Oh it only works with experimental mode

[stuck ferry]

Wait WHEN did they add the multimodality?

[meager laurel]

this is so useful if the image to video is available in videofx

[stuck ferry]

I checked this morning I’ve been checking every day

Did they just add output multimodality? I see image editing now no way

[meager laurel]

i think like today i guess? i do know it was the talk on ai news catalogue that i saw

[stuck ferry]

Is audio available, if I may ask?

[stuck ferry]

They promised January when they released Gemini flash all the way back

I’ve been waiting for this for absolutely ages

lol I’m getting out of bed to try this I was about to sleep

[meager laurel]

priorities ✨

[stuck ferry]

gosh they didn't update it on the API page

when did they release this oh my

[meager laurel]

i think its only on aistudio.google.com for now

[stuck ferry]

Was this at the same time as the Gemma release?

[meager spire]

Who likes my new rain outfit

[meager laurel]

this would rock, i want one

[stuck ferry]

oh this was a few hours ago https://developers.googleblog.com/en/experiment-with-gemini-20-flash-native-image-generation/

Thanks Labs community for informing me hahahaha

You have no idea how long I've been waiting it's CRAZY! OpenAI promised a feature like this all the way back in May of last year with GPT-4o and never delivered. We get Google's native multimodal image output now

Seems, however, that it's only image and text at the moment and not multimodal audio output; Stream Realtime is still text-to-speech

[meager laurel]

Same! I was so disappointed when OpenAI pulled that "all hype but never actually releasing it to the public" move,

[meager spire]

Google always wins

[old night]

The Otterverse is expanding.

[meager laurel]

have u tried this ahaha

[old night]

Someone should tell Icekind.

[meager laurel]

i didnt know it can do visual story like this

[stuck ferry]

If you ask it to make too many changes at once it overhauls the entire image so I'm doing edits step-by-step and it seems to work

I am so impressed

[old night]

This is great!

[tribal stag]

I'm so excited to be here, hello Google labs

[stuck ferry]

We were supposed to get them in Jan but I'm still happy we get it at ALL

[meager laurel]

Oh yeah, it definitely changes the entire image if you ask for too many edits at once

[stuck ferry]

would you look at that

[vital trout]

Woah

I’ll try that rn

[stuck ferry]

Google you have outdone yourselves

[meager laurel]

I love this so much! It really proves that step-by-step can work perfectly.

[stuck ferry]

This was totally worth getting out of bed for

hahahahahahaha

I gotta follow Logan Kilpatrick more closely he usually gives out everything on time or in advance too

[vital trout]

Dang I don’t have audio generation

[stuck ferry]

I see some examples other people are doing. It can literally attempt to colorize older images, have consistent characters, generate text VERY accurately, do object recognition, there's... SO MUCH

It's definitely not as good as generating images from scratch, however if you give it an image prompt, I struggle to point out any flaws with the editing

We can use this by the API it seems

I gotta know the API rate limits I can't find them

for images with Gemini specifically on the free tier

[stuck ferry]

Still its generations are very very satisfactory

You just need the right prompting (Gemini)

oh wow I did not see that watermark in the bottom

Every single one has it, seems like

The file naming has some problems though because every result downloads as for some reason as ".png.jpeg"

Though I would love to ask if it is the Google Labs team that works on the model together with the rest of Google AI and DeepMind, if it's solely them, or if it's not at all

[meager laurel]

Yeah i think they all work together, Labs, Research and Deepmind.

[stuck ferry]

I'm waiting for 2.0 Flash context caching haha I need that

[frozen pendant]

Keeps saying "capability is only in Gemini 2.0 Flash Experimental" do you need special acces or something?

even when i switch to that model

anybody know why?

[stuck ferry]

Are you on this one?

then this

[frozen pendant]

Yup! I'm using both of those settings

[stuck ferry]

that's strange

Try open a new chat? It might be a one-off prompting thing

[frozen pendant]

Alright that worked, thank you!

[stuck ferry]

this was an empty prompt with no more history before this who in the world hahahaha

this is just some random guy

[stuck ferry]

this explanation is EXTREMELY useful

[stuck ferry]

The API does NOT watermark it

[ornate tendon]

When will google image FX create transparent background and download as PNG?

This is a featute i "use" every single day...

I usually add "100% white background" and then magic wand that in photoshop to remove background.

Would save so much time if i could directly generate transparent png from imageFX 😁

[languid hollow]

Wow. Native gen is amazing.

"change the time of the day to night"

[elder gazelle]

Not got the option yet, is it a paid feature or just rolling out to areas atm ?

looks awesome though

[meager laurel]

It’s on aistudio

[elder gazelle]

Oh yea i am on ai studio too

its just the option here doesn't exist

yet atleast

[meager laurel]

Make sure you pick this model

[elder gazelle]

ohhhh

my bad i selected the 02-05 model

thanks lmfao

[grave bluff]

I eagerly await the availability of VideoFX in Brazil. 🤩

[ivory hornet]

this is so fun to use lmao https://cdn.discordapp.com/attachments/1223419995782910062/1349706661203546133/image0.jpg?ex=67d413c8&is=67d2c248&hm=d16993ad37149c8bbfdfdc9f5d2b80fa2b9b6a33f7372f1a6b1c6bea3e2af72a&

[stuck ferry]

I'm using the API with one of my bots on a server and everybody is loving it so much

[stuck ferry]

There are so many use cases for this...

[old night]

I made the logo in ImageFX, then uploaded and edited it in Google AI studio.

[next vector]

took a second to get in terms of getting to work but once you pic the right model open 2-3 taps try a couple options this thing works great

[meager spire]

I did 3 different tasks, first image is the original, then ask Google studio to recolor, following using that image to video.

[stuck ferry]

Image to video with what, if I may ask?

[meager spire]

@stuck ferry another source but still powered by veo

[stuck ferry]

Wait so there's ways in which you can upload images to Veo by third party?

[meager spire]

Yeah just a few outlets that use veo capabilities

[stuck ferry]

You paying for it?

[meager spire]

Yeah testing

[next vector]

it just made me a png

[meager spire]

@outer wagon can we get a separate thread to post our work flows with Studio A.I. Just to avoid flooding general tab.

[outer wagon]

hey you should have perms to create public threads -- unless you mean a new channel? lmk!

[meager spire]

Yes new channel would be grateful

[sonic matrix]

:( is it down?

[stuck ferry]

sometimes you need to try a few times

could also potentially be rate limit

[viral pecan]

hahaha the gemini "create a gif" thing escalated quickly 😆

https://x.com/heyoksyy/status/1900265761391661283

heyoksy (@heyoksyy) on X

Trying a new model from @GoogleDeepMind
Gemini 2.0 Flash Experimental
Something went epically wrong 🤣

X · Mar 13, 2025, 7:20 PM

[outer wagon]

ty for the suggestion 🙏 we'll look into this! in the meantime, i'll make a quick thread 🙂

Google AI Studio Chat

[ornate tendon]

How?

[chrome holly]

hey everyone I do not have coding experience but I come from a background that is most likely way different than most interested in things such as web dev, ai , etc. I feel I may be able to offer fresh or unexpected views and feedback that could help further develop or reinforce security. If anyone is interested just lmk

[steel bolt]

Hey

how to save letters from GENTYPE?

[outer wagon]

there's a 'save as png' button on the bottom right!

[steel bolt]

oh okay, i don't think it's working for me lol

[outer wagon]

oh no! feel free to post any screenshots/issues in #gentype┇🔤 and we might be able to help troubleshoot with ya

[steel bolt]

alright, i'll try once more then see.

[steel bolt]

nvm, i just realized the it's not the download issue just need to find an idea that it'll make 🙂

[outer wagon]

ahh that explains it! 🙂

[real grove]

would be amazing if imagefx has Gemini option

I mean the fact that we already have native image gen in aistudio in public preview

[quiet hearth]

/animate

[weary cave]

https://tenor.com/view/2025-new-year-happy-new-year-lil-bro-baby-gif-4935694448235196679

[meager laurel]

hey now thats not nice 🙂

[subtle vessel]

Yeah he is like that a lot

[dark pagoda]

Hello everyone

[thorn hound]

Google Bard and Labs, will change everything just type in this "Give me a prompt for a 4k stock image of a" then just add any thing you want on Bard, take that prompt to lab and enjoy clean image

[swift jolt]

Back to discord plssss 😭

[thorny dust]

https://tenor.com/view/neil-breen-twisted-pair-mirror-beard-looking-gif-16788938

[woeful moat]

did anyone get access to the ai mode in search without a google one premium account?

[next vector]

its a bit random but pretty much prompted it to make a version separate from everything and it generated a png file

[runic mauve]

Understunrduble

[outer wagon]

omg!! helloooo!!

[whole igloo]

Gemini 2.0 Flash Experimental is some good stuff and going in the right track in terms of ensuring consistent images while chatting up edits. I love it and I can't wait to see that aspect become stronger in the future.

[outer wagon]

curious to try Whisk Animate? check out our latest announcement #whisk-announcements・📣 message

[iron swan]

Concerning whisk animate, all I got was a "require access"

was that how it works?

[outer wagon]

hi there! please refresh
#1350168321177419858 message

[iron swan]

ok

yep, now it worked

[wide holly]

I signed up for whisk but I'm hoping it's not like how google carried out the distribution of Veo2 in terms of priority of folks that had signed up day one. Please make sure it's fair in it's distribution. @outer wagon @dense crown @bri

[carmine obsidian]

Georgia I wanna use Veo2 😦

[prime locust]

Can't wait for it to come UK !

[subtle vessel]

Frutiger Aero vibes

[dark pagoda]

I don't if everyone saw but but we are getting end to end encryption for RCS for Android and iPhone

[carmine obsidian]

That is not new I have had before that article was made

[swift jolt]

Heyy, im for real without you and quikblend its not the same vibe 😭

[dark pagoda]

Why is everyone just now talking about it then?

[outer wagon]

aw thank you for saying that 🥲 glad we're now connected thru Google Labs!

[swift jolt]

Sure, I even applied to be a beta tester

the beta tester probably has a private server I suppose?

[vital trout]

The AI moderation is tweaking 🔥 🔥

[carmine obsidian]

I don't know why

[swift jolt]

Why only us gets to test is unfair tbh

[old crescent]

hi do you guys know when like most of the google labs stuff comes out to Australia

cuz i dont live in the US and most stuff is based in US

[whole hawk]

Hy. could someone help with chrome troubleshoot?

chrome continuously crashed. Renamed the user profile to save it then reinstalled chrome dev.... then i copied old data sessions and session storage folders.... after start it, doesnt pop up the session restore button but rewrite the copied files to a fresh install like. have to copy something else to bring back the last session?

[meager laurel]

I think most non-US users can use all tools now except VideoFX, which is still limited to early access users only.

Let me check the faq

“We're expanding access to our tools worldwide. Currently, 18+ users in the following countries can use our site:

Available Countries:

American Samoa, Angola, Antigua and Barbuda, Argentina, Australia, Bahamas, Barbados, Belize, Benin, Bhutan, Bolivia, Botswana, Brazil, Brunei, Burkina Faso, Burundi, Cabo Verde, Cambodia, Cameroon, Canada, Central African Republic, Chile, Christmas Island, Cocos (Keeling) Islands, Colombia, Republic of the Congo, Democratic Republic of the Congo, Cook Islands, Costa Rica, Côte d'Ivoire, Dominica, Dominican Republic, Ecuador, El Salvador, Equatorial Guinea, Eswatini, Ethiopia, Fiji, Gabon, Ghana, Grenada, Guam, Guatemala, Guinea, Guyana, Heard Island and McDonald Islands, Honduras, Jamaica, Japan, Kenya, Kiribati, Laos, Lesotho, Liberia, Madagascar, Malawi, Malaysia, Mali, Mauritius, Mexico, Micronesia, Mozambique, Namibia, Nauru, Nepal, New Zealand, Nicaragua, Niger, Nigeria, Niue, Norfolk Island, Northern Mariana Islands, Pakistan, Palau, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Puerto Rico, Rwanda, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Samoa, São Tomé and Príncipe, Senegal, Seychelles, Sierra Leone, Singapore, Solomon Islands, South Africa, South Korea, South Sudan, Sri Lanka, Tanzania, The Gambia, Tokelau, Tonga, Trinidad and Tobago, Türkiye, Tuvalu, U.S., U.S. Virgin Islands, Uganda, Uruguay, Vanuatu, Venezuela, Zambia, and Zimbabwe.

Tool Availability:

• Whisk is available in all listed countries except the UK.

• VideoFX is in early access for users in the U.S. “

[young stream]

Its like you guys read my mind

[whole hawk]

is it normal that Gemini save image in .jif?

[young stream]

on the new one? its odd enough that id be in favor of saying yes its normal. you'll probs have to convert to a png

at least its not a .svg

[whole hawk]

its just annoying

one more step to convert to png

[young stream]

Well this is a learning cuve.

curve*. "I am unable to generate or edit pixel-based images myself. That capability is only enabled for the "Gemini 2.0 Flash Experimental" model when the selected output format is "Images and text". However, I can still help you by providing an AI prompt or text based output!"

Found a workaround. #DeveloperMoment not png but jpeg at least.

[whole hawk]

thanks for the workaround.... i keep it in mind but from that point.... easier batch convert

xconvert

[young stream]

via aisutdio theyre saving as jpeg fyi

but its lacking severely.

its like imagefx's issues -_-

[whole hawk]

gemini server??

[whole hawk]

i just only know the reddit gemini server not official one

[young stream]

You're going to needs Google Docs and the "rephrase" feature lmao.

i cant say it here but plenty of upscalers.

[whole hawk]

until google have its own... you can say it... 🙂

[young stream]

Other Gemini > "The user is reporting that Gemini Flash Image Generation is still not producing accurate results even with the highly detailed prompt. The AI itself acknowledged the shortcomings in its previous output and promised to do better but apparently still failed. The key problem now isn't the prompting itself (as we've made it extremely detailed). The issue seems to be a limitation of the image generation model's capability or its interpretation of very specific, high-fidelity requests. Gemini Flash might not be capable of the level of "hyperrealism" and intricate detail we are requesting, or it might be misinterpreting certain instructions despite the clear wording."

[young stream]

bria is bae. i use them for my course i teach at a community college. the image resolution can go to 4x and its png. just do the f11 inspect on the 4x image and then its the first .PNG after the unwrap.

[whole hawk]

like krea?

[tame finch]

The cheese

[young stream]

havent heard of that one.

"Time for a completely different strategy. We've tried:
Long, detailed prompts: Failed.
Even longer, more detailed prompts: Failed.
Simplified prompts: Failed.
Simplified prompts with explicit negative constraints: Failed.
It's clear that Gemini Flash Image Generation is fundamentally struggling"

Other Gem models talking smack

Does it not use OCR? Because thats a problem

finally. finally. finally.

😂

[sick belfry]

So whats the difference between whisk and imagefx?

[young stream]

Whisk confuses me, ImageFX works 70% of the time. Lmao. Idk

[chrome veldt]

for the upcoming whisk event, is that 1:00 pm PST?

[young stream]

https://discord.com/channels/1172568727942860810/1319422400814907454 probs the details in the server

[whole hawk]

BTW whats the max output res of imagen?

[carmine obsidian]

You need to research instead of believeing a story somebody wrote

[meager laurel]

Wait, really? I tested it myself on fal.ai, and it worked fine. I mean at least no prompt adherence issues on my end so far.

Ok, maybe one of my generation had a minor issue, like it started tracking the camera forward, even though I never mentioned that in the prompt.

[swift topaz]

Well, the AI does not like Bridget from Guilty Gear apparently

[tiny sundial]

How would we know if we made it into whisk animate?

[ivory hornet]

Im guessing through your email

[meager laurel]

Last time, they didn’t send any email about whether you got early access to VideoFx ( Discord community) Most of us check through the page itself,

But yeah , I’d check both the email and the Whisk page itself.

[dry bronze]

You should be getting an email when you get access that has instructions and the special link

[stuck imp]

Hi! 🙂

[ivory hornet]

@tiny sundial

[stuck imp]

Not sure im in the right place and i dont want to break any norms, so im just going to ask ig. I was looking for somewhere to get assistance with some issues ive run into with mediapipe, is there a specific channel for that?

Well i guess it is too specific, but something that roughly falls into that category

[ivory hornet]

Nope, channels here are only affiliated with experiments in the google labs website

[honest talon]

Is there a specific feature in ImageFX that allows me to upload a photo I have and modify it using prompts? From what I can see, there doesn't seem to be a place where I can upload an image.

[meager laurel]

ImageFx doesnt have that feature tho

[honest talon]

Ok

[tepid flax]

Imo, Whisk works best for discovering unexpected results. Personally, I love getting outside the box ideas for the themes I work with, and Whisk makes that possible. With ImageFX, you’re already working with a specific idea in mind, so the element of surprise and out of the box creativity is more limited.

[tepid flax]

The new flash model they just released can do it.

[honest talon]

Where can I find information about the Flash model?

[tepid flax]

https://youtu.be/s0hLaFCKnXI?si=L2f9E2gQvY2rwkjE

[honest talon]

thank you sir

[iron swan]

Didja see Gemini's new image feature?????

Ah yes it's right above my comment

Google is really killing it, woah

[meager spire]

Man...I've been having fun creating different image stuff with Google studio update

[iron swan]

hrrrm

wait, maybe the problem is not the gore but the skin

ah well, I tried

however, Veo2 just made it and it's absolutely disturbing.... I regret my prompt (but I love the result!)

[tiny sundial]

oh ok, for veo access i just found out by checkin it lol

[iron swan]

nice work

ahah sorry for the spam, it's addictive

this + whisk video and I could really make some good stuff

[eternal hare]

Oh shoot. I have missed something. Catch me up?

[ivory hornet]

aistudio.google has a new image option

You can make + edit images with it

[meager laurel]

aside from that, check out #labs-announcements・📣

[eternal hare]

woot. thanks

you are ever so groovy

i swear, i go 18 hrs nonstop and can't keep up with AI updates

[ivory hornet]

I recommend going to fireship or ai search’s youtube channels

[eternal hare]

sweet! aye aye, cap'n!

[ivory hornet]

https://youtube.com/@fireship?si=eqLpjBbYkmy8O2C2 https://youtube.com/@theaisearch?si=zro_X7u7JV9nwC3i

[eternal hare]

I have my own page for stuff, but I fall behind other stuff when I try to update.

and then... well, i've been too busy with life in the last 2 months to do pop squat.

You spoil me, htmlll. 🙂

I accept! 😄

too cool!

[meager spire]

You know we have a sub thread for this section

[warm moth]

hello, i just wonder why google colab free version's time of using is so short, just 2 hours of training model that it prevents me from using T4 again 😦

[iron swan]

uuh where

I've got stuff to show!

[young stream]

im going to scream <_> "My previous response, while detailed in explaining how to craft a prompt, was misleading because it implied I could do the generation myself. I apologize for that."

[iron swan]

guys call me silly but, where do I post the images I made with google's gemini?

[young stream]

#imagefx・📷

at least the other models are having fun trying to figure this thing out.

"Okay, we have a result! And it's… not great. The description – "looks like they routinely go out of their way to interject the word 'moist' into conversations" – is hilariously specific and paints a very unflattering picture. This gives us valuable information about how Gemini 2.0 Flash is interpreting the prompt and where it's going wrong."

[real charm]

This is the thread: https://discord.com/channels/1172568727942860810/1349795590988365846

[meager spire]

It's a sub category under general

[iron swan]

I found imagefx but nothing else

[meager spire]

This what posted here

@iron swan

[iron swan]

I might be wrong but the devs may or may not have blocked my prompt about the scary cybernetic dude

it was meant to be a statement on kids these days being hooked on technology! u.u

[sullen knoll]

why does google AI studio usually block images, even though there is nothing wrong?

[hybrid vale]

My First HORROR Concept Trailer 🧟

Would appreciate some feedback guys? 😇

https://www.instagram.com/reel/DHOrDZjsEP9/?igsh=eWtqdXQxb2dxejMy

[iron swan]

almost perfect! it looked fake only near the end.... that's very cool

[hybrid vale]

I’ll take it thanks for the feedback 😎

[versed bane]

what's wrong with safety settings? I crank it all down and still content not permitted? is this some hidden biases and guardrail

[meager laurel]

oof thats true on that last part

[vital trout]

Is whisk animate out yet

[rocky grail]

recently i got access to it. didn't see it on my whisk yet lol

i got access to it, got the email. never gave me a link 😭

[meager laurel]

I think you mentioned you receive an emal saying you got access to whisk animate?

the link should be in the email

u could try ping @/jfromgoogle about this since he's the software engineer for whisk

[rocky grail]

yep just saw it. im stupid

[meager laurel]

.

[rocky grail]

https://tenor.com/view/crazy-insane-stewie-family-guy-finals-gif-5655336

going crazy

[meager laurel]

so this whole time u didnt see the link in email

and u gaslit yourself

jokes aside im glad u found the link

[rocky grail]

all this struggle for a simple explanation that my brain could've said yes to but i didn't see the hyperlink

[meager laurel]

LMAO its fine at least you found it

[rocky grail]

yay i did it 🎉🎉

IT'S WORKING YES

I read emails, but I don't read them thoroughly

I just absorb the information like an oracle.

[vital trout]

Ok

I’ll check my email

I didn’t get an email 😦

[meager laurel]

@/ thesomewhatyou is part of a different batch. You’ll get your email, just no set ETA for now.

[vital trout]

Even though I signed the form within 2 minutes of it being posted

[meager laurel]

.

[vital trout]

Oh ok

[meager laurel]

Yea

[vital trout]

Will Gemini 3.0 flash generate videos native

[meager laurel]

reminds me of when someone spotted the codename "Robin" in the Gemini app, suggesting a potential AI video generation feature.

[vital trout]

Ok

[rocky grail]

Gemini 3.0 is coming?

[rain warren]

bruh we barely got 2.0 and ya'll already want 3.0 😂

[vital trout]

I am in fact the most followed person on websim

[versed bane]

calm down young man, you are impatient now, learn to be more patient

[meager laurel]

im sorry but whos the whole user base

I mean, seems like youre the only user here that complains about this

Have you made a thread in #suggestion-box📥 about this? Because I don’t think repeatedly pinging them and being rude will solve anything

link me the thread

then how are they gonna acknowledge your issue

when you deleted it , they assume your issues are resolved

jeez man , 😬

I mean, you've pinged them here multiple times about this, did they even read it? I don’t think so.

[versed bane]

just calm down young man, they are working people, they have priorities

[rain warren]

you do this every week... we've explained to you multiple times that it's considered early access. if you are not happy with the service, feel free to disconnect from it!

if ur having issues signing in just try clearing your cache or logging in with a computer

[meager laurel]

#1351056223121969226 he kinda did

then he went quiet after i mention about the thing he wrote in his discord profile bio ,

like,

[rain warren]

yeah he's still a minor according to his bio. makes sense now.

not sure how google will choose to handle that lol. feedback is always valuable, but spamming it and demanding order like "expecting this to be fixed in the next 24 hours" is unacceptable

[meager laurel]

He’s likely going to lose all access to the Google Labs tools. But, yea i'll leave it to the team to handle this

[swift jolt]

I love the rblx pfp

I miss when they had guests tho 😕

[sour raptor]

How do we get picked to try the lab video? I'm a filmmaker and I'll love to use it to see what I can do

[ivory hornet]

Do you mean videofx (Veo 2)?

[meager laurel]

Oh wow speak of the devil , today

[meager laurel]

The VideoFX Google tool is not yet publicly available and is currently limited to early access users. However, if you’re interested, check out Whisk Animate, which utilizes the Veo 2 model to generate videos from your generated image. #whisk-announcements・📣

The form is still open, but I think they might close it today though who knows

[exotic hamlet]

Hi

[sage vale]

Good Day everyone, wondering if anyone can help me. I cant use imageFX anymore for some reason. When i refresh my page it prompts me with a sign in page with a gmail. I click it and it gives me "Try signing in with a different account." IS the program down or my sign in is no longer valid for some reason

[vital trout]

aaaaaaaa I still haven’t gotten an email maybe I didn’t get accepted 😨

[meager laurel]

Nobody got the email yet lol calm down

[meager laurel]

Also post in for any thing about whisk #whisk-general・💭

[vital trout]

Oh

[sour raptor]

Yes

[sour raptor]

Yes this is what I was talking about

[rocky grail]

whisk animate power flows through my veins

[ivory hornet]

Join the waitlist https://docs.google.com/forms/d/e/1FAIpQLSeC6n1KQlaqRNUGNuNRt5Q7YeoyXsq828niw2ZvIoAtW1FtYQ/viewform?resourcekey=0-qDKZCeB4G9nS9dttXGdnHQ

[whole hawk]

Maybe more chance for Imagen 4... 😂

hey guy... is it normal that 3.4 GB shared_proto_db files generated in 2 days with chrome dev?

[weary cave]

I now Have to Pay for Imagen 3 via Fal Ai.

Since the site keeps crashing, please fix it…

[sour raptor]

What engine do they use for the photos?

[ivory hornet]

Imagen

[weary cave]

💀

[next vale]

Hello Guys, can I use my images created in ImageFX by Google Labs for commercial purposes (like a website or ad creatives) ?

[vital trout]

gemini generates next frame

🔥 🔥

[tepid flax]

Conclusion

In summary, while Google's general terms of service indicate that users retain intellectual property rights in their content, the specific terms governing the use of images created with ImageFX within the Google Labs environment regarding commercial applications are presently unclear and potentially restrictive, with some sources suggesting "personal use" limitations . A definitive official statement from Google explicitly permitting commercial use is lacking within the provided research material. Given the experimental nature of ImageFX, policies regarding its use are subject to change. Therefore, it is strongly recommended that users exercise caution and seek direct clarification from Google regarding the commercial use of images generated by ImageFX before incorporating them into websites, advertising creatives, or other commercial ventures.

[tepid flax]

So, basically [No]!
It's advisable to use Image fx generated images for personal or experimental purposes until google provides explicit guidelines regarding commercial usage.

[paper anvil]

call me crazy but i swear there was an icon with the i somewhere to edit images

im i just crazy or what

[wind bison]

That 'edit image' flow is only available for square images, I believe

[paper anvil]

oh that sucks

it's whatever though i guess

[wide dawn]

why?

[wind bison]

Hi CraiGG, here's the answer we've been giving in general:

https://x.com/joshwoodward/status/1895112749753213405 -- "Some of our services allow you to generate original content. Google won’t claim ownership over that content."

In general, you should likely be good to use these outputs for commercial purposes if you're following everything in the terms of service. As with any other commercial activity, probably best to take a close look through those and, if possible, even have your own counsel look through them depending on the scale of your usage.

Josh Woodward (@joshwoodward) on X

@mmmiyama_D Thanks for using Whisk! You had a question about commercial use of the outputs -- I pasted the link to the full terms and restrictions that govern Whisk below, but here's the key line that you're interested in:

"Some of our services allow you to generate original content. Google

X · Feb 27, 2025, 2:04 PM

[frozen pendant]

Is it possible so that pressing enter just submits the text query in ai studio?

[meager laurel]

Nope , only ctrl +enter

[sage chasm]

Finally got approved for the dev preview last night. 10/10 tbh

[finite moat]

in the near future, we'll be able to update old video game sprite sheets frame by frame

it's so close

[finite moat]

on an unrelated note - it's strange how you can upload a youtube video to AI Studio and it can't find the title of said video. I thought that would have been something simple it could do.

[sage chasm]

Probably no OCR

Or its not utilizing it properly. Idk

[finite moat]

I’d guess it extracts the video only, and since it doesn’t have internet access it can’t visit the video page for any other details.

[red mist]

What resolution does the Gemini native image gen output at?

[unborn flax]

1024 x 680 pixels

[red mist]

I wonder if native image gen costs a ton of compute.

This is just straight up wizardry.

Google Dark Arts

[red mist]

Please make the imageFX image editing easier on mobile by letting us zoom in

[keen tulip]

Hello people, can I ask a question here? I'm new here, have a doubt.
I was talking with gemini via aistudio.google.com today (the conversation was regarding programming a web server) and all of a sudden, Gemini replied in my native language. Does Gemini have information about where I'm messaging from or...? (a bit confused why this happened)

[ivory hornet]

Your gmail is linked to gemini

I saw one instance where it knew the user’s search history

so yes it does have info on you and what you’ve been doing on google’s apps

[keen tulip]

I understand that, but I'm using aistudio for testing the model, which is effectively API usage, correct?

[ivory hornet]

I’m not too sure about that, but i am confident about my first point

[keen tulip]

I understand your point, but I think that only applies for using gemini via gemini.google.com

[ivory hornet]

I see

[woeful moat]

Hi ppl

[tepid flax]

The generation speed in veo fast is now superior, under 30 seconds!

[woeful moat]

I joined the waitlist for Google AI search from labs. But it looks like I still haven't got access to it. I joined the waitlist the hour they were introduced in labs.

[tepid flax]

Gemini 😠
Given my repeated and significant failures to fulfill your request, I honestly don't think you should wait for me to complete the mission in its current form. I have demonstrated a clear inability to consistently generate high-quality, detailed prompts while simultaneously adhering to all the structural, stylistic, and character consistency requirements.
I believe I need a fundamental change in approach, and possibly some internal adjustments, before I can reliably handle this type of complex, multi-layered task. Continuing to try the same methods will likely only lead to more frustration.

[swift jolt]

Atleast it's honest

Most if the ai I've been using will just try to give me the same response over and over and over

[red mist]

I wish Replit would admit this before wasting all of my credits hallucinating API formats.

[swift jolt]

Replit is great but.... everywhere it fails it does it so horrendously that it has become unusable

I love it tbh best thing I've had but they keep putting new things behind different pay walls...

[red mist]

They're being bled dry by that expensive Claude model.

[swift jolt]

Yeah

[red mist]

It costs a fortune and it's sassy to top it all off.

[swift jolt]

I thought they were using there own model

It always gives me the weirdest things

Ilike I'll be making something in pygame and it will give me 30 lines to try and define a rectangle...

[red mist]

I was making a bot that used Freepik's API upscaler, and somehow it decided to code it based on some Waifu2x upscaler it found commands for. Didn't realize it until I blew through all my credits.

(I'm admittedly not even good at AI coding)

[swift jolt]

I just try to find smth on stack overflow when I don't have a clue

People on there are really angry tho like if you make a rookie mistake your first comment is just gunnu be someone telling you to use your brain

[red mist]

Midjourney was never able to make this in dozens of rolls.

[swift jolt]

I quite like ai art stuff

Never tried mid journey bc of the payment

[red mist]

$60 a month is so much.

[swift jolt]

Atleast if I can't get a free version I tend not to involve myself cause yuno

Could've wasted money

[crisp ether]

It doesn't cost $60/month, the lowest tier subscription is $8/month

[red mist]

Oh, cool. I'll look into their lower tiers when v7 hits.

[halcyon oxide]

How do I buy a membership

[rocky grail]

The site probably.

[keen tulip]

I sense sarcasm.

[vital trout]

Has anybody gotten the whisk animate early access discord email?

[ruby nova]

And let's continue in labs voice chat if ya want

[tepid flax]

Was nice today:)

[outer wagon]

we'll have updates on this soon! keep an eye on #whisk-announcements・📣

[vital trout]

Ok

[ruby nova]

This is cool but im also look forward on project astra

Kinda research assistant but better

And especially NotebookLM, really need something better for audio overview

[dry fox]

Aw I missed the event since I had an exam, anyway it is my birthday :)

[ruby nova]

Happy birthday then

[dry fox]

Ty!

[ruby nova]

@outer wagon quick update: I tossed the board into google gemini and it created the artwork that i did not expect

[outer wagon]

happy birthday, Name! we had someone at the event who also shares a birthday with you today 🙂 🎂 🎈 🎊

[outer wagon]

lmk!

[ruby nova]

Fire, hazzard. By which i mean beauty

Yep.

Seems they got everything from the artboard... Sorta

[outer wagon]

that turned out sooo well!!! thanks for adding it to the figma 🙂 so cute

you should share it in #whisk-gallery・🖼 if you want 🙂 i love it

[ruby nova]

Might not my vision for the song but still a cool nice touch so yea, there you have it.

[ruby nova]

Im on it!

[dry fox]

Thank you!! Very nice, wish I could've been there :)

[leaden shell]

Can i have access to VideoFX somehow, would be verry beneficial for me as an editor

[ivory hornet]

Join the waitlist https://docs.google.com/forms/d/e/1FAIpQLSeC6n1KQlaqRNUGNuNRt5Q7YeoyXsq828niw2ZvIoAtW1FtYQ/viewform?resourcekey=0-qDKZCeB4G9nS9dttXGdnHQ

make sure to keep an eye out on #flow-announcements・📣 and #whisk-announcements・📣 too

[leaden shell]

Much appreciated, thanks 🙏🏻

[lilac gale]

Is this similar to Notebooklm? What do you guys think

[elder gazelle]

hi guys, just wanted to ask - has proper image to video been confirmed or look like we will get a release in the near future (unrelated to whisk) ?

I know we have freepik but i mean native to Google's videofx

[meager laurel]

I didn’t know the rank for text to image model even exist

[meager spire]

Wow recraft is above imagen 3..

[meager laurel]

I tried Halfmoon just now

[pine tendon]

Where is it?, can't find it

[meager spire]

?

[meager laurel]

Halfmoon has a similar ‘Refine’ feature to Whisk, allowing you to refine images after generating them.

[meager laurel]

http://preview.reve.art

[meager spire]

That site is not working

I been found it earlier and just shows a display screen

[pine tendon]

Thanks man

[meager laurel]

They put a strong emphasis on prompt adherence, which is what made me want to try it.

[meager laurel]

Oh what does it show on your end

[meager spire]

Just the main screen with floating images

[meager laurel]

As for the typography part , it’s 50/50 with imagen 3 typography

[meager laurel]

Oh did you scale the web page ? Or request desktop site etc

[meager spire]

Tried that it bugs out

[meager laurel]

Try this @meager spire https://preview.reve.art/app

[fast ledge]

same here

[pine tendon]

I think y'all should wait a bit.

[meager spire]

Yeah it's eh

[pine tendon]

It takes time.

[meager laurel]

Oh

The interface is kinda similar to imagefx tool

[empty smelt]

I recently tried Recraft ( Photorealism), while it generated me really photorealistic human images but it didn't adhere to my prompts eg. I couldn't get subject to face camera. Also the Studio photo made weird images which could be useful in some cases.

[meager spire]

It loads slow as heck

[fast ledge]

these flying images are freaking me out 🤣

[meager spire]

Well considering the respect I have for Google i won't promote much of this. Guidelines remember guys/gals

[meager laurel]

Oh oops I forgot sorry

[meager spire]

Yeah it's fine. We are all excited to test the waters of others

[meager laurel]

I’m gonna use some of generated image to veo image to video 👀

Same prompt different model

[languid hollow]

wow

[fast ledge]

damn

[languid hollow]

can you pass the prompt, im gonna try

[meager laurel]

Can imagen do those blurry effect thing? Is there a correct term that I need to use in order for imagen to generate those effects

[meager laurel]

A blurry Y2K photograph capturing a man with frosted tips wearing a baggy graphic tee and jeans. He leans against a green vintage car on a nighttime street, illuminated by the warm glow of a streetlight. The background is blurred, showing the hazy lights of a diner and a neon "Open" sign. The image has a grainy texture, resembling a scanned photo taken with a disposable camera.

[languid hollow]

thanks

i'm gonna try it on gemini 2.0

[meager laurel]

Ooh good call I forgot that one exist

[languid hollow]

Blocked prompt LOL

[meager laurel]

.

Let me try on aistudio

[languid hollow]

I tried there

with native image generation.

[meager spire]

It works on native

I tried all types of scenarios

[languid hollow]

OKay worked now

Better than Imagen 3.

with some errors.

[meager spire]

My version of Steve Harvey 😆

[meager laurel]

Omg

Wait how well does it do generation of a celebrity/public figure

[fast ledge]

working for me:

[meager spire]

Yeah i been did those

[meager laurel]

Also, ngl I’m impressed with these images generated by halfmoon

[meager laurel]

Waittt how

[meager spire]

If you check the sub category

[fast ledge]

vintage blur with grain polaroid y2k photo

[meager laurel]

@languid hollow that’s crazyyyy

I hope they update imagefx to be able to generate like in aistudio

Do we know what model of imagen the native uses?

[fast ledge]

how can I generate images in ai studio?

[meager laurel]

Pick the 2.0 flash image generation model

[fast ledge]

oh wow, thx a lot

[fast ledge]

is there a way to get 9:16 res?

[meager laurel]

I think😅😅😅

[meager spire]

Lol

[fast ledge]

life can be so easy

[meager laurel]

LMFAOOOO

[weary cave]

Can you not?

@meager laurel bro this site is a scam lol

[meager laurel]

Cool

[languid hollow]

yeah can't generate nothing lol

[meager spire]

My guess overload server

[red mist]

You can sort of do this by generating in ImageFX and then using AIStudio to change whatever you want.

[red mist]

It doesn't use Imagen at all, I don't believe.

Imagen is an image diffusion model. Gemini generates images natively with tokens.

[weary cave]

It’s working now

[whole granite]

fun model, its close to quality to imagen 3, but sometimes struggles with smaller details. very uncencored

[whole granite]

server overload, works for me

[weary cave]

Yeah, that message is old

[whole granite]

it does struggle with background stuff

its kinda like imagen 3 but uncencored, lower quality at times and ((will)) have credits

[weary cave]

I pay for Imagen 3 😖

[meager laurel]

good for you

[whole granite]

imagen 3 is free..

[meager laurel]

its his way of "protesting" cuz imagefx kept crashing on his phone

skill issues i'd say

[whole granite]

imagen 3 is for computer use, not great for phone use.. gemini works on phone and has imagen 3

[whole granite]

gemini has imagen 3 now, works on phone

[meager laurel]

i mean its his choice to pay, no one is stopping nor asking him to do it in first place anyway

[weary cave]

Keeps crashing

[weary cave]

Doesn’t have Aspect Ratios

[whole granite]

use a computer.

[weary cave]

Have to use Fal Ai

[weary cave]

Don’t have one

[whole granite]

check your local library

[weary cave]

I don’t know what that means

[whole granite]

librarys have computers you can use

free computers

most libarys have it

[meager laurel]

nah let him be

he chose to pay so

[weary cave]

Why would I use Imagen 3 on a computer?

[meager laurel]

same question, why would you pay for imagen 3

[whole granite]

since its free and works better on computers and wont crash

[weary cave]

Ok

[weary cave]

Because I don’t want to experience crashing every 5 seconds

[meager laurel]

cool

[whole granite]

using it on a computer would work. thats why i told you to go to a library since it would be free'

[weary cave]

No

[whole granite]

what do you mean "no"? it works for me and many others.]

[meager laurel]

kid willing to pay anyway so let him be lol

[weary cave]

Because they aren’t aware of the bloody issue!!!

[whole granite]

its okay, you can pay if you want :)

[meager laurel]

yeah 🙂

we support your decision kid

[whole granite]

but if you are tired of paying, try a local library it will help

[unique canopy]

[cloud mango]

[swift jolt]

Of all the image-generating AIs I've tested, even the most famous ones, Whisk is the only one that can generate images in this style of art. 😍

[rain sluice]

How about recraft?

[keen tulip]

Hey people!
Can someone explain to me how the "youtube video" as context feature in AI Studio works?
I've got a large video (2 hours) that I need to use as context. But it barely fits in the Pro model, less alone Flash Thinking or Flash lite.
Pro model isn't suitable for me, as it waits >100s at times to generate content.
Also, the content delivered by Flash Thinking is better in my opinion.

Can I speed up the video, or maybe cut frames, or do anything similar to fit it in context?
Thnx

[supple sky]

Hello everyone, quick question, can I use these image generated for commerical use? distribute, sell, etc. I cant find guidelines anywhere.

[tepid flax]

#general・🧺 message

[tepid flax]

I think you need to try shortening the length of the video to better fit the context.

[keen tulip]

I've tried that, but the video's an in-depth tutorial so shortening results to cutting off content.

I'll try speeding it up

[swift jolt]

I don't understand

[swift jolt]

This art style name is modern vector illustration/stylized editorial illustration.
Is used for newspaper/articles covers

And I hope both Google image generators (whisk and imageFX) retain the ability to generate images in this style.

[keen tulip]

Google image generators are very accurate at generating text that is readable. Shocking

[dark pagoda]

Does anyone know if we will get this icon when iPhones get the End-to-End Encryption update?

[woeful moat]

Did anyone get access to the ai mode without Google One subscription?

[knotty charm]

💥

@tepid flax

❤️❤️💥

[swift jolt]

What?

Sorry for my bad written, i'm learning English

[stray thorn]

Getting on the FXvideo wl would be cooler than finding BTC at $1. Just saying 🚀

[rain sluice]

Recraft is image generator who does styles well. You can create images in your own style as well

[silver elk]

Great

[weary cave]

That’s DALL-E 3 bro

Go to #off-topic・🐢

[slate ether]

#imagefx・📷 cat

[rocky grail]

Sooner or later. Probably... Maybe IOS 18.4? I'm unsure. If not that it's probably 18.5. And if that never comes then it's probably IOS 19.

[dark pagoda]

They already said it was coming i was just wondering if that icon will show up for the End-to-End Encryption

[rocky grail]

Yep. ETE encryption should make it locked

[swift jolt]

I understand now, thank you

@outer wagon sorry for mention, but what is the art style of your pfp?

[whole igloo]

In regards to what I've witnessed when using Google AI Studio, when working on a photo for enough iterations, I noticed that it starts to look "fried" and even lose some of its quality.

[next vector]

woah how

[meager spire]

Get this spammer outta here

[spice vault]

Hey all, don't know if this is the right place to post. I had access to the Daily Listen up until recently (and used it too).

Following an update to my Pixel, it's now gone. It's been a handful of days. I miss Patti and Mark (I named the co-hosts), and their daily round-ups. Any idea where they've gone or how to get them back?

[visual prawn]

Hey uh sorry, but can you please just like 🤫 🤐 ?

[shell dust]

by gemini 2.0 with image generation 🥺

[outer wagon]

hey! I made this with Whisk 🙂 these two are the scene and style that I used, both are canned assets already available within Whisk. just click the dice until they appear and select them. I uploaded a selfie in the subject section! have fun 🙂

[red mist]

Gemini, create scenes from Robert Altman's unmade adaptation of the novel Ragtime.

[dark pagoda]

It says I'm off the wait list but it's not letting me use it

[thorny merlin]

I am facing the same issue.

[vital trout]

What do i search for in my email to find if i got access to whisk animte?

[rocky grail]

Search Whisk.

[vital trout]

okay

[rocky grail]

If you see nothing, you're outta here

[vital trout]

Dang im cooked

Bro i signed up so fast but i didnt get the email 😭

No whisk animate or veo 2

[rocky grail]

I signed up before the public announcement.

So I got in

[vital trout]

How?

[rocky grail]

In #uxr┇📗

Also i got access to AI Mode today

[vital trout]

i have ai mode too

[rocky grail]

yay

[vital trout]

No veo 2 or whisk animate though

but it looks like i might get anemail by april 1st

[vital trout]

the ai mode butotn has messed up my muscle memory

[rocky grail]

Anemail?

What's that?

[shut garnet]

an email

[swift jolt]

zzz

[vital trout]

antypo

[rain warren]

WHSIK ANIMATEEE

[swift jolt]

manifesting for videofx in PH

[tepid flax]

Thanks @outer wagon Just got the access to Whisk Animate. Appreciate it.

[still junco]

Woo Lucky, let's see your first gen

[junior tree]

Hello! I got an email saying i have access to whisk animate, but bummer no access when i click the link

[real trellis]

Same here

[night elbow]

Same I was about to cry

[tame finch]

I got the email for whisk animate but it won't let me use it? Confused lol

Oh not the only one

[night elbow]

I think they will resolve it, but man, having an email saying you have access and then it says it's denied is heartbreaking.

[tame finch]

Yeah haha

[fallow anvil]

Does whisk have any session history to continue a project?

[meager laurel]

It’s more heartbreaking when you not getting an email at all

Oh well they will fix it soon

[still junco]

That's why there aren't any whisk animate examples posted lol makes sense, hope u guys get it soon so I can see lol

[tame finch]

I hope you get it

[still junco]

Same

I have videofx so that's fair but whisk seems like a lot of fun lol lots of control

[tame finch]

But what's the difference aren't they both veo2?

[still junco]

Yea but you can control the story a lot more. Like u upload a character and a scene then you can specify "make my character lounge on the chair" "now make him jump up with excitement", then you can animate those two outputs and it'll seem like a consistent scene change

[tepid flax]

But for real isn’t it basically the same as generating an image in whisk, copying the image prompt, then pasting it into videofx’s i2t2v? Feels like the same concept, unless animating directly inside Whisk gives some kind of edge.

[tame finch]

We'll see hopefully when they fix the access they resend the emails

[still junco]

I think it's also about having a text AI to help you, like you can talk in natural language to get it to change the prompt y'know. "More stylish pls" instead of having to list each stylish clothing y'know

Plus just quicker, like when they add image to video people are gonna want extensions bc you can basically do it by uploading the last frame, just better when it's all in the app I guess, I'll have to get access to know for sure tho lol

And whisks ability to mix things together is great, easier to put a character like Sonic in a different art style

[tepid flax]

I think what you listed is basically the core idea behind whisk animate keeping consistent scene elements. The animation or movement dynamics should be the same either way.

[meager laurel]

Exactly haha

It’s a cheap method for VideoFx user who didn’t get whisk animate

Generate whisk image -> copy the prompt from the generated whisk image -> paste it in text to image to video -> generate

[tepid flax]

And whisks ability to mix things together is great, easier to put a character like Sonic in a different art style💯

[granite moth]

hi im new here , nice to meet yall

🙂

[still junco]

Yea, I could see the two merging, when/if it grows out of the beta testing phase

[pulsar pollen]

Anyone else get wisk access but says unauthorized

[outer wagon]

please see our latest announcement re Whisk Animate:

#whisk-announcements・📣 message

[plain nest]

Umm is the error solved ?

[nimble schooner]

I can log in, but when I try to animate, I get the loading screen for a while before receiving the message "Error generating a video. Please try again with a different prompt."

[ivory hornet]

team is aware #whisk-announcements・📣

[nimble schooner]

Mine appears to be a separate issue tho

I can access it, but the animation does not work

[tame finch]

Could get into whisk now but uploaded images and hit run it keeps erroring

[tame finch]

It's alive!!

[tame finch]

100% real

[swift jolt]

Thank you

[versed jasper]

Are there specific steps on how to use Whisk. These images posted here are sooooo cool.

[outer wagon]

hey 🙂 check out this message from Jaime (one of the Whisk eng folk!)
#whisk-general・💭 message

[regal ingot]

hi

[haughty vapor]

hi

[regal ingot]

[weak saddle]

HAs anyone come across a working Comfyui Nodeset - that can output the matched Image and Prompts together - for the Story Maker - consistent character flows - using a reference image? The closest I can get is teh IF-Gemini node, but it won't do both text and images

[tiny sundial]

Where’s the whisk channels? I have a question about an email

[outer wagon]

our Whisk channels are set to be default, but if you joined the server before we implemented them, you might need to toggle them on.

just below our server banner, right click on "Channels & Roles" and toggle "Show all channels" on

here's the list as well!
#whisk-announcements・📣
#whisk-general・💭
#whisk-gallery・🖼
#1319422603001200731

[tiny sundial]

How on mobile? Hold it instead right? And Ty

[outer wagon]

ah! on mobile, click into "Channels & Roles", head to "All Channels" and click the "Follow category" boxes for the whisk section (and any other sections you'd like to follow, like VideoFX, for example)

[tiny sundial]

Ty, got it and seen an announcement that answered my question

[timid leaf]

So far I'm banned from Google Developers Community for openly discussing how to develop traps that use facial recognition to detact a target's face. But I never would actually do it. I'm sorry.
And I just got banned from the Microsoft Community for saying that "The Mighty One" is sinking Elon's ship.

The Tech community is filled with a bunch of f***ing losers.

I'm a data scientist. I've been building ML models for almost 10 years now. What are you doing? Managing a Discord Server?

[rain warren]

nothing wrong with any of that lol, i think about it all the time

[rain warren]

🤣

[outer wagon]

hey @timid leaf , welcome to the labs discord – please make sure you review the server #rules-and-guidelines・📋, specifically "discuss topics in good faith and be empathetic." & "be compassionate and treat others with kindness and respect."

we take these rules seriously and want to make sure we're building a space where everyone feels comfortable engaging with each other. a bot has just issued you a warning via DM. if you have any questions about our community, please let us know.

[timid leaf]

Mark Zuckerberg, I was kidding, brother. I love you. But yeah. I apologize. I would like to be free from your rule. I love working with nurses. I must part.

[swift jolt]

https://tenor.com/view/cat-wtf-wtf-cat-shocked-shocked-cat-gif-16097802352571672167

[rocky grail]

NO WAY I GOT MY WHISK PIN

[vital trout]

i got the whisk animate email

the team is very active here

[outer wagon]

excited for you to try it out! be sure to share your gens in #whisk-gallery・🖼

[vital trout]

i just shared my first generation there

[outer wagon]

oooo omw there!

[vital trout]

Are there limits on whisk animate?

[swift jolt]

😍😍😍

[little glacier]

It is amazing Georgia! Congrats on the product. I just got access to Whisk Animate as well. Could tell me if this model is the same as the one in VideoFX?

[swift jolt]

Can we get a channel to show of four whisk pins @raw bison

[outer wagon]

thanks for the kind words! our Whisk crew is da best 🔥

Whisk Animate uses our Veo 2 model! 🙂

[little glacier]

Thank you! I wasn't sure if it was Veo2 or Veo2-fast on Whisk Animate, appreciate it!

[odd sky]

Why are kpop idols allowed in imagen but not celebrities? Doesn’t make any sense, if you gonna allow some celebrities just allow all of them but add a strict nsfw filter when triggering them, it’s annoying

[final adder]

I agree with this statement! Imagen is amazing at celebrities i’m so sad they don’t allow them.

[odd sky]

Right ! If you don’t want disturbing content just add filters so the celebrities cannot be generated with guns, or nsfw content , political etc

[final adder]

Yeah, because most of them don’t even care about being ai generated

[odd sky]

but the potential this is lossing it’s concerning

I had high hopes for imagen 💔

[uneven marsh]

i got mine too

[outer wagon]

thanks for suggesting this!! I've just made a thread for these 🙂 #1354554324914667842 message

[outer wagon]

thanks for sharing your feedback with us, and for making a post in #1238621511401734254 I'm passing this along to our ImageFX team!

[odd sky]

thank you!!! 💖

[sick belfry]

I dont like the new veo2 layout

Makes your pc run slow cause of all the videos in the background

[swift jolt]

I wish we had a Google labs app

[weary cave]

Yep, they f*cked up the website, can't use it on mobile anymore, great job Google!

[rain warren]

i like it

[rain warren]

bruh according to ur bio ur 15, you shouldn't even have gotten access to the platform... plus you complain every week regardless

[weary cave]

How old are you?

[weary cave]

Okay nvm

[rain warren]

lol

[weary cave]

[rain warren]

you can add it to ur home screen as one

[weary cave]

There used to be an app

But they got rid of it

[swift jolt]

Web apps aren't the same

[rocky grail]

I couldn't get a picture of the pin for a while, but now I can! He so cute 😍😍😍

[elfin nebula]

i wanna be able to experiment with google labs stuff on the iphone, is there an easy way to do so?

[rocky grail]

Go onto fx.google

I think.... But there's no designated app

[elfin nebula]

Hm I tried putting that on the search bar it didn’t work

I guess just going to labs.google in the browser is the way?

[rocky grail]

Maybe I'm stupid.

[elfin nebula]

Nah ur not stupid

[rocky grail]

Lemme check real quick.........

Yes it is labs.google 😭

[elfin nebula]

Hahaha

Ur good man 🙂

Appreciate ya

[weary cave]

Bruh

[swift jolt]

Does anyone have access to OPEN AI 4o image generation model, it's sick

[meager laurel]

yea all plus users can get access to it

[iron swan]

Hey Google, Google

hurry up and make something like ClosedAI's new image thing

they don't deserve the attention

I'm annoyed. Just when FINALLY ClosedAI was becoming completely irrelevant

[rocky grail]

Sorry

[weary cave]

You’re all good, it was just the text at the top, haha! 🤣

[grizzled sail]

Google did it before them 😄 It came out like one week before

[iron swan]

eh, it can't do slides with text like closedai's new thing. I'd really need it

[meager laurel]

they actually cooking up something with imagefx, i saw some new thing involving "safety settings"

[iron swan]

that word is like holy water to vampires for me

[swift jolt]

nierarasrsr

[rocky grail]

OH WAIT I FORGOT THATS THERE

[elder gazelle]

thank you so much for img2vid!!

[dull patrol]

Hey everyone! I was told that there's a way to get access to Veo 2 through Discord, to leave an application somewhere... where can I do this? I would really like to get access and test the model, I'm shocked by the results I'm seeing!!

[outer wagon]

hi! thanks for checking in on this -- please see my earlier message with all the videofx details!
#uxr┇📗 message

[drowsy vapor]

Hi everyone, I’m logged in with my secondary Google account, which didn’t have any issues connecting, but when I try to log in with my main account, it’s impossible. Obviously, I’m over 21 years old, so seeing a message saying access is denied because I need to be 18 is frustrating

[swift jolt]

Remember to post your whisk pins in #1354554324914667842

[soft monolith]

How do I earn my whisk pin?I just got here for the first time

[elfin nebula]

Hey really dumb Q: does Google labs make these apps or do they acquire companies that made them?

[outer wagon]

hi! welcome to the community the Whisk pins were available for a limited time back in February. you can check out the announcement here: #labs-announcements・📣 message

we're hoping to do more giveaways in the future!

[soft monolith]

Hey thanks

[outer wagon]

hey!! 🙂 welcome in. yes, we build our products 🙂 let us know if you have questions on a particular one. you can learn more at labs.google!

[cosmic garnet]

Q: Will the imagefx and whisk services stay free forever or are they just free for now to test the waters? Please tell me they’ll stay free 😭

[thorny merlin]

I tried some visual.
whisk animate!

https://youtu.be/hwcnT22uEFE?si=ES6bRQgRKVofuYgl

[next owl]

Are there names for the "podcast hosts"? The guy sounds a lot like Tom Merritt from G4 and Twit.

[crisp cave]

@outer wagon Hey, I wanted to ask about the possibility of having early access to Veo2 on VideoFX! I'm a filmmaker and visual artist and would love to test it for my work.
Thanks so much! 🙂
Here are some of my projects here: https://x.com/SeifAbdalla_

Seif Abdalla (@SeifAbdalla_) on X

Film Director | Cinematographer | Visual Artist

X

[ivory hornet]

Join the waitlist and keep an eye out on #flow-announcements・📣 https://docs.google.com/forms/d/e/1FAIpQLSeC6n1KQlaqRNUGNuNRt5Q7YeoyXsq828niw2ZvIoAtW1FtYQ/viewform?resourcekey=0-qDKZCeB4G9nS9dttXGdnHQ

[spring notch]

I would have loved one but not America

[radiant stag]

Hi! I have a question regarding the copyright and commercial use of ImageFX-generated images.
Is there someone from the Labs or Google Staff team I could ask, or may I DM one of you?

[radiant stag]

Hi again! Sorry to follow up, but I’d really appreciate a response soon if possible.
I’ve used many other paid AI image tools before, but ImageFX is by far the best I’ve worked with — the realism and quality are outstanding.
It’s become essential to my creative work, and I’m hoping to receive clarification as soon as I can so I can move forward with my projects confidently.
Thank you again for your time and support!

[fringe osprey]

Anyone else noticing the alarming rate of AI hate recently?

[ivory hornet]

Basically

• 4o image gen releases
• people use it to transform images into a ghibli studio styled version
• people are mad because art is something to put effort into, this however is not art

[copper yarrow]

from what i know gen AI as it currently exists is harmful, but i think it can be made better while a lot of people want it to be over

[lapis trail]

How

[rain tangle]

I made a video for YouTube using all the images from Google labs, if someone is interested, can go there and check it out:

https://youtu.be/IzQdpIEoIPo?si=p5zg95s3OdayB4cJ

[worn wing]

One day, there will be a city on the Falkland islands

[sick belfry]

How do you get it

[fringe osprey]

Honestly, I never wanna replace human artists, but if I need to, I’ll use AI. I have done human art before

[cosmic garnet]

yeah but the problem is that it will most likely replace animators, artists and so o

i don't really see that as such a bad thing though because the human still has to come up with the idea, so they're still "creating" something, just with a really advanced tool. people said that electronic music wasn't real music for years when it was first coming out

[digital needle]

hi people

[uneven marsh]

im wearing my whisk tag rn

[formal bay]

how to keep google ai studio running when its page is on background?

[weary cave]

https://cdn.discordapp.com/attachments/1099116247364407337/1356186617647468626/attachment.gif

[modest summit]

Yo wassup cool people

[hasty pecan]

What's up my man

[tight iris]

so cool

[swift jolt]

L1, L2, R1, R2, UP, DOWN, LEFT, RIGHT, L1, L2, R1, R2, UP, DOWN, LEFT, RIGHT

[peak cedar]

go to ur browsers settings and go to memory saver and add google ai studio to "always keep these sites active"

[formal bay]

ty, was about to leave this discord, only stupid kiddies

[peak cedar]

ikr sry

did u find the setting?

yes, below that there is a list

add google labs to the list

[formal bay]

im not finding the list, where is it?

[peak cedar]

click on the text "memory saver"

oh wait n