#general・🧺
1 messages · Page 2 of 1
vscode for code, nano for editing configs
vim?
im more of a cat + echo person
🧌
you are just wow
I used vim and emacs during my undergrad, I used vim a bunch on AIX machines too
if is cursed, i use
yay vamp person is vim
for JS code editing i use the Sources panel of browser devtools and write it in the Watch expression
Which one you like better?
I think I preferred emacs. It seems like both of them can be really configured to do anything you want though
lol
i lauv using discrd for kode edater and i cpy past into Dolphin
i laaauuvvv ed!!
emily is og now
she used all vi emacs nano vscode and gmail
Bedros approved editor?
yis!!!
i lov
src kode simpl as well!
is goosd assignment for lerning C
recreting ed
de ting
dis one corse in priceton
dey make de rite new
is your name grandell?
looks like someone here has seen my content 🤩
Excuse yourself
grandell 2.0
Mark, I do not have access, no one currently has access it has not even been rolled out to production servers.
you sound very sure
Love Mr Bean
day 12 of asking to get imagen 3 / videofx access to regular / original server members
day 12 of asking to get imagen 3 / videofx access to regular / original server members
day 12 of asking to get imagen 3 / videofx access to regular / original server members
same
Are some already have access?
Not to my knowledge.
i cannot believe the ad i just got
these people used gen dino to create a jumping avocado
jumping over pizza
day 13 of asking to get imagen 3 / videofx access to regular / original server members
day 13 of asking to get imagen 3 / videofx access to regular / original server members
🅱️edros’ influence is undeniable, unrelinquishable, unrelenting
that may be due to me being chronically online
don't you have a job or something
Bedros with a capital B
I'm 16 i cant get one where i live even if i wanted to 😔😔
day 13 of asking to get imagen 3 / videofx access to regular / original server members
day 14 of asking to get imagen 3 / videofx access to regular / original server members
day 14 of asking to get imagen 3 / videofx access to regular / original server members
2 weeks in
is anyone in here a real AI content creator, businessman, and not a memer or just do this for funsies? dm me, like please.
ive made some non meme content with ai
what sup
sup bro
day 15 of asking to get imagen 3 / videofx access to regular / original server members
day 15 of asking to get imagen 3 / videofx access to regular / original server members
day 1 of asking @whole granite and @surreal saffron to stop spamming
its about the grind
day 15 of asking to get imagen 3 / videofx access to regular / original server members
We're all hyped for the new tools but this is getting repetitive and spammish imo - maybe make a dedicated thread for these daily messages 
its like asking protesters please do this in a private room
not that we're protesters or anything though
day 16 of asking to get imagen 3 / videofx access to regular / original server members

@surreal saffron https://discuss.ai.google.dev/u/bedros/summary
Yes that is me
wait until you find out about my whole online presence
day 16 of asking to get imagen 3 / videofx access to regular / original server members

wait until you find out about mine
waiting
finally, things get interesting ;)
so, moderator of forum or developer of ML @ openai?
mhh
is that all though? 🧌
no
https://www.pcmag.com/news/google-code-tips-launch-of-bard-advanced-as-paid-ai-experience interesting...
you seem to have a few news articles on you
including one from The Verge but they entirely misunderstood what i was saying 😔
Let me make a website that has all of my profiles to make it easy to share with you
do you have a service you would recommend?
for posting a bunch of links?
ye
probably pastebin
i mean if its just for this purpose / convo, a pastebin would be fine no? 😭
Here are the ones straight off of my head @surreal saffron I will finish it later: https://grandell.carrd.co/
whatever happened to quality over quantity? 🧌
the blog link no workie 😔
that's in beta rn to trusted people I know
you can save my email to contact me outside of this forum
i dont have many socials, but the socials that i do have, i'm rather happy with
even if i dont have that many followers, i'm happy to say i've never done follow for follow and this is just my account existing
the platforms I am most active on don't have followers
also, if you send me an email I can add you to my contacts
I have one that involves me (not by name) :P
I will not share it due to the misunderstanding nature of the article
Sure, that's fair. I have a couple on me too, but they aren't very interesting.
day 17 of asking to get imagen 3 / videofx access to regular / original server members

we want video generation, chop chop google
day 17 of asking to get imagen 3 / videofx access to regular / original server members

show me the most inetersting songs you've made with musicfx :)
check out some of the submissions in #music-🍭 🔥 so good
day ONE MILLION 18 of asking to get imagen 3 / videofx access to regular / original server members

day 18 of asking to get imagen 3 / videofx access to regular / original server members
- tis a nice day today
i'm probably gonna leave this server
tbh i don't support the creation of music with ai
just top tier lazy
what about as an idea for people to get their ideas out when they dont have enough talent to do it?
oh hes gone
some people forget that music is self expression and the instruments are just a tool
day 18 of asking to get imagen 3 / videofx access to regular / original server members
day 18 of asking to get imagen 3 / videofx access to regular / original server members
Day 1 of asking for Imagen 4 / Video double fx / I’m just messing around 🥴
messaging is top tier lazy, you should walk over to his location and say that to him (since driving is soo lazy/is a tool that anti ai people seem to HATE)
leaving is top tier lazy. i don't support leaving servers
using phones are top tier lazy, sending emails? bruh soo lazy, go train a carrier pidgin to fly real letters.. oop thats to lazy, you might have to walk over to the location/boat over (oh sorry boating is too lazy) ur just gonna have to walk. don't sleep because thats top tier lazy just stay up forever and never sleep
blinking is soo lazy, it shows that you can be blinded for a few milliseconds which means anyone can attack you
loll
hi @still latch
👋
just walk on water, ez
thats top tier lazy, drowning is better
oh wait
dying is top tier lazy
pls check dms 🙏
When will Google Notes in search become available on the desktop and/or mobile Web interface
A future Google product I request is to port arcvm to Windows, macOS, and desktop Linux, although arcvm has already been ported to Windows via Google Play Games desktop, it would be great to expand it to the Google Play Games for Windows to non-game apps, and add integrations like window management, shortcuts, file type handling, etc (basically encompassing (almost) everything arcvm on ChromeOS supports to Windows, macOS, and Linux)
Porting arcvm to Linux would probably be the easiest, although Windows has the highest userbase, with Windows Subsystem for Android either shutting down or being exclusive to China (People's Republic of). WSA failed due to the lackluster content of Amazon App Store, and APK installation requiring third-party solutions (Which still then needs microG)
But macOS would be the last one to support due to being harder than Linux to port, and not having as big as an audience Windows has (and the fact that part of the App Store is already supported, and can run iOS apps completely natively)
QEMU can be used as an alternative to virtualization if it is not supported on some computers, although it would be slower
Another thing highly suggested is to make the Pixel Data Transfer tool part of the AOSP, so app developers can program in support for their own apps to be transferred, and users can add support for bringing in more settings and app data, even if they don't have access to the app's codebases or documentation via reverse engineering the filesystems (or referring to digital forensics reasearch on the app).
Guys as we know Imagen is in beta, so will there be credits/subscription in the future when it releases?
Any else see
yeah I saw it a while ago
Looks so cool, makes me want to listen and learn. Can't wait to get access, hope it's quick
yeah
Signed up awhile ago
Just got access to Imagen 3 by Google and oh my...
It's "first time DALL-E 2" all over again.
Absolutely in disbelief with the results.
I will share some outputs soon.
I really, really hope they open source this. Would be the best thing after the SD3 fiasco.
Bedros?
beef...
Image 3?
I've been waiting for this for soooo long, the reason I came here originally was to see how people were using it in preparation
Yeah it sounds promising
But then it's just a combination of rag tools
Kinda useful, hope they make it free
yeah this seems really neat.
it looks pretty cool
i gotta check it out soon
Yesss
any eta on rollout of VideoFX for Trusted Testers?
will be sharing as soon as we can! stay tuned!!
Oh? What’s a trusted tester?
Kind of an old reply, but what app is that? 🤔
google photos
Lmaoo ok, I thought it was some ai demo app or smth 😭
Mb then
perhaps the event link should be posted in #labs-announcements・📣
https://canary.discord.com/events/1172568727942860810/1248690689840124064
@outer wagon what OS and editor does labs.google peeps use?
probably project idx 
the featureset really is super compelling, I hope a way to keep costs afloat is found equitably!
Find out in the fireside chat? 😉 perhaps Framework Chromebooks!
What’s the fireside?
Look at the #events link at the top of the channel list :D
Who tried the notebook lm beta version
I have, it's interesting.
Nice
Nooo!!
This guy isn’t my least favorite AI “influencer” but he certainly isn’t good
Curious how so? Been seeking more industry folk to balance out my AI academics list
I just don’t really think he knows what he’s talking about - I think he uses a lot of buzzwords to grab attention around AI, but he doesn’t actually understand AI
https://x.com/minchoi?s=21&t=tcE_EmsNX4Y4iXyD_8-HpQ is much worse imo
AI Educator. 𝕏 about AI, solutions and interesting things. Showing how to leverage AI in practical ways for you and your business.
Bedros
datpep
Huh
Hi active
I have tried! Was learning to do camera input for interaction with Processing and I used all the web pages and my notes as input and generate some interesting summary.
Any word on videofx? @outer wagon I have image3 just not videofx.
I popped in about 300 pages of cyber security information to help me study for my tests. Seem solid
Oh sick
That’s a good idea actually
I might do that with some cybersecurity material
I’m working on fine tuning llama3 w/hacking stuff to make it a good CTF aid lol
hacking stuff as in just a conversational CTF aid?
Love this sticker
Cool do you need help?
Did y'all see the deepmind video2audio tech? 🔥
Awww this is cute
Nah
I want it to understand code a bit more
And craft exploits better
Well, yeah, I guess yes then
Lmao you got some spare A100s?
i got h100s
hey what ever happened to the dino gen game?
looks like it's extinct again
Which meteor killed it?
I think nvidia has program that get you some
They have a start-up program, but all it does is give you the best deals when buying compute.
For uni students
Oh?
anyone got out of the waiting list and now can access ImageFX and VideoFX ? --> https://aitestkitchen.withgoogle.com/tools/image-fx
it's been over a month or 2 now since the event...
@outer wagon i just saw the access post you did...
Why is it "Us-based" only... ?
Ah, misread the latest form 😅
Gl to all that get accepted!
Might be a legal thing :)
just wanted to shoutout the team for the awesome event the other week, was fun using gentype
my match my phreak bag got great reception
@outer wagon on the sign up form it mentions Twitter... Not X. 🦆😜
why us based everytime 
whats the program
yoo imagen3!!
super excited for y'all to start playing around with it 🙂
sammme
but no one would see it :( my friends wont see it
oh wait
i can share in the labs server
so that works i guess
hey y'all 🙂 as a reminder for everyone: per the trusted tester agreement, you may not share creations publicly (in this server or elsewhere) without written consent. be sure to check out the description on the G Form for some more info!
This form is for the labs.google Discord community to apply to be a Trusted Tester for Imagen 3 on ImageFX. You'll be one of the first people to try out the latest update to our tool. Select outputs may be featured on Twitter as inspiration for future users when the tool launches publicly.
Criteria: Must be 18+ years, U.S. based.
Please note: ...
i signed up!
Hi, was just wondering if its possible to get a tour of google offices in Toronto.
Does anyone know how I can organize something like that?
OK guys only 5h left!
if you don't have a team for Google Capture the Flag here is an opportunity, since it seens there is no limit of participants in a team, you can try to participate in: "The Biggest" team!
The access key is: -O-usz8N9mSIXY9q6ykf
https://capturetheflag.withgoogle.com
Let's hack!🔥 :flag:
talk to me if you enter
just saw you in the ctf discord 😆
hey y'all, as a reminder for everyone: per the trusted tester agreement, you may not share creations publicly (in this server or elsewhere) without written consent.
I'll pretend I saw nothing 👀
May i dm you about it then?
damn i think im a bit late
Damn i didn't know i couldn't even share an error screenshot in here... Sorry @outer wagon :/
💀
So excited to try Imagen 3 when I get access 🙂
I’m just getting errors now after it was working for a few hours.
Did u get yet, I did
Am I allowed to share more specific feedback
Mostly positive here
No pics
@outer wagon
And now its working for me again ? lol. Maybe we get x amount usage a day ? 🤷♂️
I wouldn't belive so
Nope, nothing yet!
Check email
And spam in email
You filled out form before today right
anyone having issues generating images with imagen3?
Yeah sometimes it won’t generate and I’m not sure if it’s an internal error or it’s not allowed a prompt from me
Yesterday it wouldn’t generate at all for a few hours
the "im feeling lucky" prompts always seem to work but when it comes to my own i have to retry about 20 times to get one image
its whatever tho, its still super cool
I think the guardrails at the moment are super tight.
prolly
Yeah same
Yeah filled out the form a few days back, I'm not seeing anything in my email
Maybe you weren’t selected
can i dm you the art it made?
its really good. after a subtle Magnific upscale on some images fooled me... i had to pixel peep to find wrong stuff
What’s the best way to prompt for realistic images?
Hey guys
Heyya
hey Trusted Testers
we’d love to see your creations and hear what you think of Imagen 3! please refer to the onboarding email we’ve sent for the correct contact information where y'all can share creations, and instructions on how to communicate within the program. appreciate y'all and really looking forward to hearing about your experience through those communication channels. 
for those who are still awaiting access to become a Trusted Tester, be sure that you signed up for the Discord-only G Form that I [shared last week](#labs-announcements・📣 message) -- this is separate from the general waitlist that's available at labs.google. invites are rolling, so if you're waiting, hang tight 🙂 reminder, you must be at least 18 years old and located in the U.S. at this time to sign up for early access.
Thank you Georgia, have completed the form!
the email it gave me does not work
would you mind elaborating? does it error out?
Hello everyone!
welcome, doc!
Thank you! I'm excited to be here.
i can dm you it
sent
I would love to gain access to some of the programs
awesome! check out my message from this morning for the requirements & steps to sign up for imagen 3 access! 
Image3 is def a game changer.
curious, y'all -- ***how do you like to use ImageFX in your daily lives? ***
for me, personally, i like to use it to try to execute on art styles that i enjoy and can't necessarily replicate myself (i am not artistic, like, at all). would love to see any examples y'all have too 
idk if i really use ImageFX or any image generator in my day-to-day life
maybe for infographics for school (uni) + work?
it's certainly fun to play around w/
for work we use ai a lot, but not imagefx.. and for personal work i like doing portraits and landscapes simulating film, but imagefx is really tight on barriers when asking for people ... ive been getting amazing results in imagen3 and people, but it censors a lot of generations so it makes the process kinda boring.
imagen3s editing capabilities are meh as well, so you cant edit out AI 'slop'
it uses imagen 2
for editing
it said
ohh ok thanks
neat! do you have any examples? totally understand if not as i recognise it's not a frequent thing 
😔
?
hi
Hello.
wth, openai forum moderator
speaking of the OpenAI forum, they never let me in 😔
I moderate the public forum community.openai.com
The private forum is mostly for students at big colleges or CEOs.
Ohhhhh
Thought you were talking about that, oh well lol
Still neat
There is like one message sent a day on the private one, as it's mostly only video. There wouldn't really be a point to moderate it.
I notice @kindred crescent that you are friends with Kesku. How do you know them?
Wonder what the point is then 🤔
The videos.
From the OpenAI discord lol
Are you active on there a lot?
Yep
Puzzeler?
Used to be, embarrassing how I lost it but trying to get it back
Actually, not really. I can't say I use image gen all that much, but I really like to play around with MusicFX a lot lol
close! GenType 
https://labs.google/gentype
Make an alphabet out of anything
Ah
Lemme try it though
The DJ mode for MusicFX is cool
Though the waiting music that was at Google I/O that was being live generated by AI was a disappointment
And it threw an error :/
It is in Beta and the moderation endpoint they use is very strict.
No it's not the moderation
Agreed! Love messing around with it
It's just that the model is busy
Wow, I have never gotten that before.
Anyone else is getting prompt errors when using imageFX?
Like i write a prompt and immediately after clicking create it shows an error and say please try again
And it gets stuck like that.
yup ... the other day was for half of the day ...yesterday for about 3 hours.
And whats the fix - keep trying or wait a few hours and retry?
If i make the prompt shorter. Less sentences. It works but it totally lose traction of whats needed.
it doesnt work for me with shorter prompts ... i just go to do other things and comeback later to test if its working
o dont think there is a fix in the users side
Your best bet is to wait a few hours and retry.
+1 to this! sorry it's being a bit slow -- recommend waiting for a bit and retrying. also recommend trying a different browser!
Hi, I'm from the ImageFX team. I wouldn't expect to see that behaviour, and wouldn't expect you need to try again or try a different browser. Any chance someone can capture a screen recording, or anything else that would help with debugging?
yea i’ve noticed the same too. short prompts or if u give any advanced prompt it will usually give an error
but i expect that since it’s under beta
Here is what I get, for like every other prompt I run
Thanks. What's the URL you're using? Are you in the Trusted Tester program?
Yes, I am using https://aitestkitchen.withgoogle.com/tools/image-fx This happens with Imagen 2 and 3
Is it possible you've exceeded your quota of image generations? i.e. have you generated many images today/this month? (edit: probably not, apparently we show a useful error message for that)
If you're having an error could you please DM me the email address you're using?
Aha, I think I see a potential issue. When we are unable to generate images due to our safety filters, we should show you a meaningful error message, but that may not be happening right now and you're seeing "Something went wrong" instead. Will look into it. Thanks for reporting the issue!
edit: yep, that's the issue. So if you're seeing "Something went wrong. Please try your prompt again." then it's our safety filters. If you think that your prompt is something we should generate images for, given the policies in our FAQ, then please submit feedback in the app and we can look at the individual cases and try to improve. That's always extremely valuable information for us. We'll fix the error message, probably today.
I can do that for you
Where do i send this?
Last time i reported a screenshot here about this issue i got deleted and warned...
And i didnt share any images... I was sharing a bug.
" then please submit feedback in the app and we can look at the individual cases and try to improve. "
here
I see. Policies making prompt to stop.
Is there any guidelines or list of keywords it will automatically avoid from receving the prompt?
Cause most of my images are cats lol
"Black cat holding the excalibur sword. He is defending his castle, his people and his honor. Standing on a rock and his cape moves with the wind. Staring at the sunset and world shine upon him"
I dont see anything incompliant there 😜
I did. But it didnt give me options to specify what i did to get the error. So i tried here afterwards. And Georgia warned me and all about sharing screenshots
There's defiently a few issues to iron out but yeah just keep dumping them as how we are asked
Overly strict filters being 1 issue etc
oh wow ..hahaha .. try this prompt in imagen3 ...it gets the filament direction and shininess really really good ...some images even came with filament errors at the borders... impressive ..... "multiple 3d printed vase set in natural pastel colors , minimalistic shape"
ideogram gets it too, but the outputs have multiple direction of filament in the same object... midjourney struggles and firefly apparently doesnt know what 3d printing is ..
Hahaha, 'cat' is one we just fixed. I think that should be fixed now. Turns out some minor musician's name is 'cat' and our systems got a bit confused. Don't worry, we love cats on this team too.
To the general questions, the policy our system attempts to implement is the one in our FAQ. There's always room for improvement there so please do keep sending any prompts that fail for you via our in-app feedback, we read every single one, I promise.
Thanks, I have already seen some prompts that shouldn't violate the guidelines be fixed. I will report any more "false-flags" I receive to you.
Seriously.... Oh my cat...
Are we working with an intentionaly small dataset or ? I find im getting very similar images multiple times when I try certain prompts
Have you checked you seed?
Yes it’s unlocked
Try creating some in a new session. Also, are you using Imagen 2 or 3?
Imagen 3
Waiting for Imagen 3 access because I joined late :/
I finally got my Rabbit R1 today
Can we get a breakup Google ai lab
Nah man it ain’t worth it
I am having fun with it
Fair enough
If this is disappointing please file in-app feedback (and note you're using Imagen 3). There are various things we're tweaking there to try and see what works best for users, so your feedback on things like variety of returned images is extremely useful and important right now. As ever, thanks for testing!
The session shouldn't make any difference, but yes, the seed might well make some difference in variety. Just FYI @karmic scaffold .
Yes have already submitted some feedback, will keep submitting more as things pop up
Thank you so much!
What if we need a consent to use certain AI generated elements as part of our google slides presentations?
Like icons, emojis, illustrations to back up the content.
I would love to generate items i can then photoshop and help me explain things better.
Let me know.
I know google slides + gemini have a generative tool for that but it isnt as great as imagenFX
I have workspace standard + gemini.
We need permission to share things w/Imagen3 - I think Imagen2 is fine? But tbh I'd just use a diff image generator for your slides (Imagen2 is fine, but there are better)
Imagen3 on the other hand is very capable, just, yk, private
Imagen2 is fine.
Thought so, just wanted to make sure
k back to sorting thru google sheets (I wish i had gemini in GSheets)
If you have a Google Workspace subscription, from Google Slides you can generate images using Imagen @ornate tendon
It's $6 a month
Compare flexible pricing options for Google Workspace with plans for businesses of all sizes.
I never use it, but I also don't use Google services that often
I got a Google Certification in the Gemini aspect of Google Workspace so I know all about it
Here are some screenshots I just took of Gemini in slides.
I don't pay for it.
hey @kindred crescent glad you found the new "Multimodal Canvas" demo. We just posted it this morning, here's a bit more context on it on X (https://x.com/alexanderchen/status/1806691815333958023) but basically it's a demo built around the Gemini 1.5 Flash API that lets you quickly test out different multimodal prompts with drawing, images, emoji, etc.
Want to play with Gemini + multimodality? ✏️🧠 Here's a new "Multimodal Canvas" demo by @dcmotz built w/ Gemini 1.5 Flash that lets you quickly test prompts using drawing, camera, images, emojis: https://t.co/b6QHKt2EBk One of my favorite things to try is pictionary:
Is it using real time video or just images for the camera input?
Just images. It takes a snapshot of your current canvas state and sends it with the text prompt when you submit the text prompt.
🥲
And you guys just HAD to reuse the Project Astra UI lol
Does screen share even work?
Yes screenshare works, but it's image based (takes a snapshot). Haha. And yes, this is NOT Astra as it's not using realtime video. This is a demo that's an example of what any dev can make with the existing Gemini 1.5 API (it uses Flash)
here is an example with screenshare on. I share a tab to the canvas, then i said "describe my screen" and it treats your screenshare just like pixels. you can also draw on top of the screenshot.
you can draw on top of the screen share too, like this
Is screen share desktop only? :(
that's what I said - but no, it's not as good as ImagenFX 😛
where did you get this? I want to learn more about it. I'm a google fan
oh haha this is fun. added my API - what else can i do with this? 😛
looks like image to text.. mainly
They have a bunch of certifications https://www.coursera.org/instructor/google-career-certificates
But are these official google certification i can use to brag and attract credibility to my clients?
Coursera is full of fluffy certificates from unknown teachers :/
Update: oh Google hosted their own training there.
Which one did you take related to gemini + workspace? It has many subjects and didnt see that..
Woaaahhh
GC
Yes it's just a frontend for making simple testing of image + text prompting faster. You can drag in images, draw, screenshare.
GC?
Check your Google Account privacy settin...
I want to try Google DeepMind's V2A
Whats that?
Im excited to try videoFX instead. That has to be sooo good.
omg ikr
its gonna be lit
ImageFX lags with this "edit" image where you use brush, and give a new prompt.
Wonder whats a good guideline to make it work.
Until now it just blur-push-scramble things under the brush 😜
ImageFX lags with this "edit" image
https://deepmind.google/discover/blog/generating-audio-for-video/
litrally Google it lol
But that is to generate audio... I was referring to.videos.
I have in mind illustration animations for education purposes. To teach practical stuff about the brain psychology, human behavior and learning development
Anyone have a general time frame for when they signed up for trusted tester through the discord link and when they were accepted into the program? Thanks in advance 
A few hours, but a bunch of Staff are on vacation right now so I would wait a week or two.
Awesome, thanks for the response
I was just accepted into the program about 10 minutes ago. So to answer my own question, the process is very quick 😂
LUCKY
Oh, sorry if that wasn't your experience, I wasn't trying to be rude about it. I'm sure if you're in this discord you'll be accepted in no time as well!
bro.
I signed up the day I joined this server :(
That's not so long ago though. Like I said I'm sure you'll be accepted soon 👌🏻
Thanks
Good luck 😊
But yk 🥲
Imagen 3 is awesome don't get me wrong, but I am much more excited for Illuminate and NotebookLM. I've used NotebookLM quite extensively and I think an it's absolutely incredible resource
I'm hoping I get access to Illuminate so I can create my own projects. So I too am on a waitlist. I feel your pain my friend 🥲
Remember not everyone is accepted.
It will eventually be released to the public.
I have a pretty extensive background in the IT and Cyber Security fields and work in the military sector so that could potentially be why I was accepted fairly quickly.
I also want this
More useful utility for me vs image generation
Exactly. The demos on the experiment are already so impressive so I'm super stoked for that to get rolled out more widely.
Especially once that's integrated into notebookLM like it was shown at I/O 2024
Game changer
which one is Illuminate? what does it do?
Illuminate is the one where you give the tool source material and then it's almost like a podcast where 2 AI voices chat back and forth about the subject matter on the source material. Eventually they will make it so you can interject into the conversation to add to it or ask questions and the AI will recognize that you are speaking automatically and such so you can have a proper conversation.
Look at the I/O 2024 demo for it, it looks awesome.
Sorry forgot to tag you in the messages above
ah. got a link for that? I was present in the I/O event and dont remember this one.
2 AI having a conversation like a podcast? that's crazy - now add 3d realistic models moving as people talking and we can create the next-generation news / tv shows / movies 100% with AI lol
BUT... that is giving me ideas...
AI assistant.
Remember the CLIP from microsoft words?
would be fun to create an actual AI assistant that moves. just like the guy in the time machine movie
https://illuminate.withgoogle.com/home this? @chrome flax
Yep
You can listen already to the pre made library. Highly suggest checking it out and giving feedback 👌🏻
damn the voices are AI?
with all the interaccial accents mixing now at days globally... no one would care or distinguish if it's AI or not.
I am listening now.
Yeah it's fully made with AI as far as I know
But of course that's pre-made by google to show where they want to take the project, not sure exactly how far along it is as of right now
joined the list 🙂
Awesome!
I don’t think they look into you that far tbh
They ask about you when you apply though and that was the info I provided them
Ah ok
I don’t think I put much into my form lmao
I also have a cybersecurity background, getting into AI/ML + cybersecurity rn via a project
Well if you were accepted that's all that matters
Arming llama3 with my version of naptime (https://googleprojectzero.blogspot.com/2024/06/project-naptime.html)
Tbh I honestly don’t think AI image generators are as cool as other applications of AI
I think AI video is awesome
And AI images certainly have their use cases
I just don’t see them as useful
With that being said, Imagen3 is certainly capable (with a only few categories it performs poor at)
Came with my Pixel, sadly it's only from the Google 25th anniversary sale lol
It is $20 with free shipping and returns, and you get to pick your color.
Oh nice they have it on eBay lol
It's the only reason I go to google.com, is to click on it's image search feature as it can get you the exact image, or one's around it, or shopping results, and it has always worked great for me.
Lucky you! Let me know what you think of it 👌🏻
Let me know if you can upload your own material to it pls
You can paste PDF urls but not sure if you can upload your own
You get like 5 gens a day
Damn that's pretty limited but it's also a lot of processing power to generate all that so I understand that part. How long ago did you sign up if you remember?
@kindred crescent sorry keep forgetting to tag you in my responses
Check dms
Day the waitlist launched
Aka May 14
Nvm 😭
All good, I hope I get access soon
I'd love to experiment with that for academic and career purposes. Such a great tool
If I'm not supposed to upload this, then the mods will delete it but for now it's basically free to share as I can't see anything that tells me not to lol
I joined the waiting list
Great stuff. I want to experiment with it too - so i can create digital podcasts..
I wonder how to design the voice of who talks what.
Hey guys btw is there an official discord server for these 2?
• Google ads (performance max of search)
• Google workspace
Anyone used multimodal canvas yet? It's pretty awesome
I don't think those exist as official discords but I think there are a few unofficial ones if you search around
Anyone else get an Google user research study invite?
Nah unofficial are usually scam or someone in the leader group trying to sell something
GDC has a workspace channel and SEO channel
channels might look dead but there are people there waiting for a question to answer
GDC?
Yeah, google developers community
imagen3 when you ask it to generate space cats makes them just look photoshopped lmao
i wont share the picture here (duh)
gonna include that in a feedback email lmao
imagen3 has the tendency to seemingly photoshop in stuff
and doesnt color match
^pls send in feedback email so we can collect accordingly. thank you!! 🙏
let me test
it wont even work
somehow "space cats" and "cats in space" is bad?
Kk
The sticker book prompt is so much better when you use Imagen 3 vs Imagen 2. It's night and day 🫢
What’s the sticker book prompt?
In the today's special section it prompted us to create a stickerbook style photo.
lmao mb i missed that
daily quotas for imagen3 are activated now?
I noticed Google Vids is out!
waut what!?!?!??!?
videofx is out???
oh its not
still waitlist
lol no Google Vids is something different
oh ok
The cat ending was hilarious!
How did you do that? Edit function?
I can see Memes being generated with that type of joke. So funny
@ornate tendon Have you had any luck generating people directly?
I honestly never liked "recreating" real stuff.
So i use anime / illustration / 3d realistic - but never real humans.
And i have this cat-love... so its cathumans haha
But I'll play with humans and see what occurs - especially babies
Lmao yeah cats are awesome
Here’s mine
Thats imagen 3 generated 🌟😜
@raw bison you know da Vinci's work and how he used to hide secret messages in his paintings... Right?
Can imagenFX 3 encode some secret code so Google knows when images are being published FROM imagenFX?
Random idea sharing here.
@ornate tendon have you heard of synth id? https://deepmind.google/technologies/synthid/
mind blowing. I can't be more in love with Google and everything they do lol
I assume this works only from Google's ecosystem, not images/videos/audio made with other AI out there.
anyways, yep. that's what i was referring to. Amazing.
#Question guys...
Im telling imageFX to use white background so i can remove the background. but using photoshop to do so takes too much time and doesnt always end up being a good product.
And I could use Google Slides to remove the background nice & fast - BUT i can't find a way to export the image without background from Slides as .png or something.
Any ideas?
Prompt-wise or google slide option im not aware of ?
@raw bison could you pls check dms 🙏
Nah
I’d rather not get removed from the program LMAO
Is this a thing? lol... guess I can't test this tool enough
By the way, images are still being created despite this 'warning'
can someone remind me where do we report bugs?
Hi! Is this in ImageFX?
Yeah
i asked some days ago if daily limits where "activated" somehow, because a couple of week ago i could do several tests... is maybe around 20 daily now??
I wonder if googles affected by this crowdstrike thing
Crowdstrike?
Oof
Well
Uhh
Several millions of computers are down
Because someone pushed a faulty update
And it’s causing havoc in the IT space
Jsyk, it's just for Windows, so me and everyone else who uses Mac's are fine.
Mac’s are really stupid for business and really good at the same time
Just the almost always mandatory $99 for Parallels 🥲
Why is it mandatory?
I don't use a Windows machine for my business, or for my work.
Some software
Such as?
Is just for windows. Lol.
What is something you actually use though that there is not a Mac replacement for.
well
Probably a really bad example but like
Gaming…
Yeah, really bad when you said business in your inital post.
yeah forgot about that lol
I don't game, and if I have a need to for work I use NVIDIA's servers.
But I guess you can pull some compatibility layer stuff?
But it’s usually easier on Windows at least
Idk, I actually never used a Mac so I take back my original statement
👍
I had only used Linux until I tried out Mac and now I just use Mac as my primary system but still have a linux on hand.
the prompt " man swimming " doesnt work / isnt allowed ? It was working a week ago ...
Hi Larry! Thanks for the heads up. Did you happen to report that in the system via bug reports?
I was going to but as of today its working again .. I think there's a cat and mouse game of prompt fixing
Thanks.
I mean
macOS is a worse OS for professional use imo
Besides video editing maybe
Windows or bust
Don’t get me wrong, macOS has some great features
But I think windows is better for most people
Some Mac replacements are just bad knockoffs, no?
No, companies make there stuff for MacOS these days.
Explain why please.
You can even get the full Microsoft suite of stuff such as word on Mac.
I mean, sure, but Windows has more support for basically everything
Windows is just more customizable & compatable with everything
Windows computers are also less expensive than Macs
Making them the better choice imo
Not necessarily.
Can you name something that you can customize on Windows that you can't on Mac?
Hardware
and, from my understanding, there are some mac tweaks that are paid
Most of Macs are laptops, the hardware being hard to customize is the same on most Windows laptops.
Expensive is all relative. I can easily get a Windows machine that closes just as much.
Also, I was talking about the OS not hardware.
I mean its no secret that MacOS/iOS and just apple products in general are much more closed off. Their operating systems are less open to sideloading, much less customizable and yes they are more expensive for the hardware that you get.
Ans if you want an example, the dock is a perfect one. On windows with the Taskbar you can add widgets and theme it within the system settings and change more than you can with the dock in mac where your only option is it's location.
This is not true at all. There are lots of things you can customize on it. Also, Windows task bar seems to have a different use.
The task bar has the little windows launcher where you can add stuff to it. As well as the dock
Also, there is way more software support on windows for random tools than on mac
Cain and able is a perfect example
Extremely valuable tool for people in the IT field but doesn't work for mac
And there are many many more tools like that
Idk why you are even arguing this, it's a fact that MacOS is more closed off and it is in fact costlier for the specs you get versus the price you pay. Period.
You're arguing things that have been know for years
There is a reason almost all enterprise systems run windows. It has better support for encryption and management tools and is generally such cheaper and more customizable and upgradeable then macs
I literally work in IT specifically the Cyber Security field lol you aren't going to create an argument that I won't have a rebuttal to 😅
The lightning cable is all I need to say and the conversarion is over. Creating a proprietary cable just to sell your product that has less compatability for a higher cost is just a shady business practice in general.
That's why the EU literally forced apple to out USB c in all devices going forward and its a good thing they did. Additionally they have been forced to allow mode sideloading and make their devices more repairable. Apple did not want to do any if that until they were legally FORCED to do so.
They want to overcharge for their repair services going as far as adding geolocation to repairs so that if you aren't getting a repair done in an authorized location, wuth authorized parts, they will void your warranties just for trying to avoid their exorbitant costs to repair devices.
Shall I go on?
First off this is an OS discussion. Second off, I speak a LOT with people who work at the top cybersecurity companies. There are other password crackling tools that work with Mac's the launcher and sidebar of whatever you are talking about on Mac has the ability to add custom widgets and whatever. I won't take broad claims. People say it's well known so it should be easy to give me examples.
He mentioned a password-cracking application
But it’s just a personal preference in that case
I saw
I am fine hacking on a Mac
Yeah
Web apps or…?
What’s your speciality
(Just asking, separate from the discussion)
I mostly hack web apps
Unless you are asking outside of hacking
as hacking is not my carrier
Career*
ah ok
when are getting videofx
Never, they just made the UI, but they forgot to actually create the model.
yep
What?
Gold eggs in an easter egg hunt
This conversation makes no since.
@tender ice why did you ping @spice pulsar
Why do you think OAI got Golden eggs at an easter egg hunt
the easter egg hunt of humanity
what does that even mean
and what's with all the Sam Altman emojis
@spice pulsar
.......is this thing on
What thing?
the discord
Yes, why?
just checking since i've been gone for a month 🙂
Sure, no one is doing #todays-special・🥣
I am pretty sure it's a boycott for VideoFX
You’re back??? 🤯
Same, just got back yesterday
When is the next NYC Meet up?
lmao no
😭 i dont think we're boycotting for videofx
And I thought it would be videofx 😭
It is very expensive to get a live production setting going, especially for video models.
I would much prefer if they had a paid beta instead of free so that there could be some kind of price deferentation that could help with release.
Also, they probably don't have much training data.

They have. The entirety of YouTube.
I actually never thought about the fact that they own Youtube, I just hope they aren't training it on kids content 💀
😭 if I see Skibidi toilet represented in VideoFX I’m nuking Google
That isn't even scratching the surface of the worst kids content
Don't go onto YouTube kids you will be scared for life
question about the new people that got imagen 3 access, can they post the images or no?
No
ok because some people leaked images
i told them to delete and they did
one is offline and has a bunch
Tell them not too
Are you a imagen3 user?
yes.
Nice
it was on a public server.
oh weird
Yep that's how you get your access revoked
ohh
Ask who they asked
they emailed the google acc that posted the access thing
Ah
he said that we can't post the images on websites but are alowed on chatting platforms like discord small servers
with other trusted users
like posting on twitter they would strike you but over discord they can't do that, so if its in a smallish server you can post the images
i may be wrong
imagen 2?
Because staff specifically said to not post on this server you are 100% getting access revoked.
yeah maybe remove that please
ye im gonna email the mod to ask just in case, i think i got it right
yes
cool
"cocomelon!!!"
"sing with me! "a, b, c, d"
i cant lmao
Bruh I never got in at ALL
I guess you aren't who they are looking for.
LMAO rip
FINALLY
After 7 applications, Jaicraft has been accepted.
wait...
bro.
But
@outer wagon any reason why I got a fake invite? 😭
OH my bad
@violet trench @ @kindred crescent — you mention that you both received emails — who are these from?
I'm so sorry, didn't realize I HAVE to click the link in the invite
just want to confirm
It works now
ah yes, you do
nice!!
Oh and one more question
I remember someone saying that we can share them to people that we trust not to share outside, is that right?
I'm not fully certain about it, so I'm checking first
I’ve already had imagen3 for a while, they were from labs.google-fx-tt-admin@google.com
definitely check out the trusted testers agreement for specific details on that 🙂
I might have skipped it by accident... 😭
Is there another place where I can access it?
@outer wagon I just lost access again
Bro
EVERY TIME I PING
IT FIXED ITSELF
oh crap
THIS IS GOOD
aw man
I really need a copy of that agreement
@raw bison ? Can I have a copy of the trusted testers agreement please?
hang tight so we can resolve!
Thanks
Wow
@kindred crescent the agreement appears in-product, you have to read & accept to start using the tool. as for sharing guidelines: you may not share anything publicly (outputs, experience, etc) without Google's approval. to submit content for approval, follow the instructions in the email.
(also, welcome to the tester program! 🥳 )
Thanks, hope I don't get sniped for just saying that it's a good model 😂
You can say that just ifne.
The agreement is per-product yeah?
Is there a trusted tester channel where we can post images generated by Imagen 3?
Not at the moment, but a nice feature request @outer wagon
a user claimed to ask one of the mods and they said you could post imagen 3 images in smaller servers with other trusted members and they can't post on like twitter or instagram or they will be striked by google
we do not have a space like this just yet, but we're workin' on finding ways for tt's to connect 🙂
thanks for flagging this -- this is incorrect information. as Bri mentioned above, trusted testers may not share anything publicly (outputs, experience, etc) without Google's approval -- this includes sharing to other servers and small friend groups.
yep i contacted the user and they didnt know, the email was sketchy and didnt double check before telling people to share, when he told us we could post the images i pointed out the TOS but i guess they didnt listen, they now deleted the imagen images
can we share images made with imagen 3 in here?
No
okay, thanks for the wuick reply
any word on veo?
no
I think the Google team will tell us - they already have a plan for release I’m guessing, no point in asking for it
We don’t have it yet for a good reason
but it is exciting
can’t wait
Been away for a while with work stuff... We only get 3 images per day now 
WAIT WHAT
Usually takes at least 10 before I get the desired results
Well that odd, I got a notification on imageFX saying I had hit my daily limit but can still download old ones if i need to
yeah man I've been hard struggling with 3 today
it's been such a pain
Im good now, its let me create more
I don't think there's a limit on my account, i've probably generated like 30 images today
why caps?
gl
I thought the daily limit was 3 according to them
@summer slate Does it specifically say you ran out of limit?
Yeah
I cant remember exact words but something along the lines of "You've hit your daily limit, you can still download old images if needed"
Our limits are much higher, must have been a temporary glitch. Can you try again please?
Just start a new session if you get this error and almost one hundred percent of the time it gets reset.
I have access to a billion image generation services but I never use them. I am interested in knowing what you make with them.
google study rn
What images do you create to study it, controversial ones?
Oh oen of the things with Google.
Indeed this seems to be the case/
can we share images tho??? since everyone (im the usa) have access??
@outer wagon we can now share images right?
its true!!! nice detective skills 🔍
we'll more formally announce on socials soon. yes share your work!! we'd love to see what you all are creating.
you guys are so on top of it LOL i love it
plz share with us in #todays-special・🥣 too 
will Imagen 3 be integrated into gentype?
ok good now that imagen3 chat is allowed here

it is pretty bad at generating planes + missiles (but I can see why missiles aren't really in the dataset)
it also latches on to weird details and makes some of my realistic images anime
or partially 2d
appreciate the feedback! please keep sharing your thoughts with us 
we don't know just yet, but we hope to! 
I dunno, this looks pretty good to me
seems fine for me
let me get the images
pros - amazing at photorealistc stuff cons - not so good at art / limited data set
Also amazing at small details and prompt accuracy a lot better.
Damn, what's the prompt?
man surfing holding a flare at night
I'm actually gonna beta test another frontier image gen tonight/tomorrow, so would love to test it on the prompt for this
Those missiles are completely messed up
Front landing gear is open
The others aren’t
Missiles aren’t on proper pylons
Planes get worse w/distance
And when there are multiple planes it messes it up
Generally fine but it has a French flag and a messed up USAF roundel
Cockpits a bit messed up too
These are generally fine alas disappointing
Particularly the super hornet
finally ..i have a google photos album full of images i couldt show xd...
its also great with pixel art
are ppl getting access to videofx yet?
on the settings of ImageFX I don't see the Trusted Tester anymore but it's Imagen3
It's general availability now
Everybody has it
No
**What Gemini model is used to power Notebook LM? **
Hi guys, I'm interested in understanding what model is being used to power Notebook LM. There are a variaty of Gemini models out there with different context windows and perfomance metrics, and it is clear that Gemini-1.5-Pro-Exp-0801 is by far Google's best model.https://chat.lmsys.org/?leaderboard
As such I'm trying to understand if I should use Notebook LM or Google Ai studio, as Ai studio provides access to Gemini-1.5-Pro-Exp-0801.
I prefer the UX for Notebook LM way more, as it is targetted specifically for my use case (document Q&A). However I don't know if NotebookLM is using Gemini-1.5-Pro-Exp-0801 under the hood .
Could you could provide some infromation about the exact model that is used to power Notebook LM? If the model is not Gemini-1.5-Pro-Exp-0801 , do you have any benchmarks as to performance compared to Gemini-1.5-Pro-Exp-0801 on document retrival?
Thank you for your help!
NotebookLM uses Gemini 1.5 Pro
do you know which exact release? as this makes a big difference on performance: https://chat.lmsys.org/?leaderboard
e.g.
Latest: gemini-1.5-pro-latest
Latest stable: gemini-1.5-pro
Stable: gemini-1.5-pro-001
Experimental: gemini-1.5-pro-exp-0801
From here: https://ai.google.dev/gemini-api/docs/models/gemini#gemini-1.5-pro
i want that chair
WHAT
I haven't touched it incase I'm not supposed to have it
Lucky…
Already emailed one of you guys letting you know - but if someone sees this here first
I don’t have it :(
Let me know if I'm actually allowed to use it
sad
You should fr test out at least one video
No
No need to send it ofc
Nope, I'm ok
Interesting.
WHAT WHAT WHAT WHAT
REAL???
AAAA
BRUHHHH
hey devs if you need any testers, don't be scared to ask me 👉🏼 👈🏼 thanks
refreshing my email like crazy
watching the google stream afterparty!
Me too 😭🙏
looks like they got it resolved
i still have the trusted tester role but i cant gen videos anymore
(they replied to the email)
Were you able to gen videos?
yep
i ended up trying one as a test
yeah
i emailed them and it's removed so we good
¯_(ツ)_/¯
still have the role somehow
You could have been very popular on Youtube 😆, being the only one with vid access, but I don't really care about it too much.,
I would happily accept it, but I don't really use image / video gen too much.
i don't use image generation all that much day-to-day but video gen would honestly just be kinda fun

