learn-about・📖

[hollow depot]

Welcome to the #learn-about・📖 channel: Learn About is a 💬 conversational learning companion that adapts to your unique curiosity and learning goals. Think of it as your own guide who can help you learn and gain a better understanding of any subject, for school or for life.

We released Learn About last year to signed-in users located in the US; it is now open to all US users ages 13+!

To get started:

• Navigate to the Learn About homepage (you may be prompted to sign-in to your Google account if you’re not already).
• Type your question into the input box. You can also upload an image or a PDF file and ask a question about it, or choose from one of the suggested topics to explore.
• You should get a rich answer with images, videos, and interactive content like lists and diagrams. Try clicking on various parts of the UI to continue your conversation, or type your follow-up questions! ✨
• You can also explore related topics by using the card on the top-left.
• Your conversations will be saved (if you have Web & App Activity turned on for your Google account). You can find them by opening up the left-hand-side menu. Revisit past conversations any time!

We can’t wait to hear what you think – send us feedback in-product (via the 👍👎 icons on each response, or the popup surveys you may see) or in this space. Happy learning! ❤.

We are actively working on new capabilities and features, including a much better PDF experience, and new active-learning tools to help you deepen your understanding. Please let us know what you like, what you don’t, and what features you’d like to see!

[elder jolt]

Is this something new about it?

[zenith storm]

Hello

[elder jolt]

What's different from before?

[crude zinc]

Is it not allowed in the US? weird...

[zenith storm]

Is this another LLM?

[radiant hedge]

Hello

[elder jolt]

2.0 flash

[tribal patrol]

This sounds like something I've been looking for!

[stone slate]

when this gonna be available in outside US like india been desperately waiting from really long time like almost an year

i should ve tried US VPN 💀

[vapid cairn]

I got it in US

[elder jolt]

Yea it's a available here

Also

The new UI looks good

Except the home page

[crude tapir]

When is it coming to other countries and can I try advanced stuff like medicine?

[vapid cairn]

It's acting like 2.0 flash

[elder jolt]

Essential features like mind maps, audio overviews should be available

[elder jolt]

There's a learnML 2.0 flash model on AI Studio

[vapid cairn]

When j ask complicated questions the response is relevant but the little popup tiles arent

[vapid cairn]

Outdated

[vapid cairn]

Oh wow I was curious where they would use that

[vapid cairn]

Also how do you know

[elder jolt]

Yea they curated that model for here

[elfin reef]

That's sad

[elder jolt]

Cuz you can literally use it?

[unborn roost]

Hi everyone! I'm Matias and one of the eng leads on the Google Learn About team. Thank you for trying out and all the questions! I will try to quickly answer a few things here.

[boreal quartz]

This is really cool stuff. Love what I'm seeing so far

[elder jolt]

I would like to suggest a few things. Please bring mind maps and Audio overviews

Right now Notebook LM is a better one than Learn ML for studying

[unborn roost]

• US-only: yes, sadly as of now you need to be in the US to use Learn About, but we're hoping to expand to other countries soon!
• English-only: similarly, we've kept a narrow quality focus on English-only queries and responses for now, but we'd love to expand that as we are able to.
• Model: yes as some of you might have guessed, this is using Gemini 2.0 Flash at the moment.

[unborn roost]

Hi, it should work if you're in the US... Might you be using a VPN or some other special kind of network?

[elder jolt]

It's not the traditional 2.0 flash right? It's the one in the AI studio right? Is it just a promoted 2.0 flash?

I hope it gets 2.5 flash gets implemented because 2.5 flash with thinking maxed is like the best model for studying 🔥🔥🔥

[vapid cairn]

How do you know it's learnlm?

[unborn roost]

please send us feedback in-product by using the 👎 (or 👍 ) when you feel content is not relevant!

[vapid cairn]

Oh no jk not lol

[elder jolt]

Bruh because I can read? 💀

[vapid cairn]

Well, if you guys are taking questions why did you choose not to use learnlm? Isn't it a product specifically for teaching content?

[unborn roost]

We made a bunch of updates to Learn About at the beginning of April: new model (Gemini 2.0 Flash) with updated prompt, chat history, initial PDF upload support (better experience coming soon), 13-17yo access, dark mode, and a few other bug fixes!

[vapid cairn]

I was referring to this learn about product

[elder jolt]

Ohh wasn't they the same thing?

[elder jolt]

Ahh great choice

[vapid cairn]

No? This one uses 2.0 flash (not learnlm)

[unborn roost]

Yes, great question. LearnLM is actually built into the Gemini models. The model we're using, Gemini 2.0 Flash, has some of the initial LearnLM improvements. Newer models like Gemini 2.5 have more improvements; we're actively working on moving to that, it's not as easy as you might think, even (or because of) inside of Google!

[elder jolt]

LearnLM seems to be a promoted model of Flash

[elder jolt]

Ohh. Very excited for 2.5 👀 would you mind giving spoilers if it's gonna be 2.5 pro or flash?

[vapid cairn]

Built in? What would be the use case of just learnlm 2.0 then? It's not in vertex AI, so can't see pricing, is it significantly cheaper and somewhat similarly capable?

[vapid cairn]

Do you plan to support links?

[elder dragon]

I always had this idea of making something like learnabout. Cool stuff guys. Unfortunately I'm unable to try it out since I'm outside of the US

[unborn roost]

It's definitely confusing, but LearnLM is more like a set of training data that is getting incorporated into the main Gemini model releases; that way, every Gemini model incorporates LearnLM goodness. I think that the "LearmLM" models you see listed are perhaps from when Gemini 2.0 Flash was not out yet, or was in preview.

[dawn spire]

I'm going to uni for a degree in marketing and generative AI this year. This + Gemini Deep Research + Notebook LM + music making models is really going to amp up my learning. I love the interactive diagram!

This can also really help me level up my YouTube videos. Wow, this is incredible! This is what AI is for, not for writing everything for you. A tool for learning and expanding creativity and knowledge.

[lapis crow]

UX seem neet, can you stop making great ones please, you are too much fast ahah

[vapid cairn]

So no new learnLM models will release, right?

[lapis crow]

the 2.0 is already

[charred quarry]

This is so cool! This is exactly what I’ve been hoping for since LLMs released in general

[unborn roost]

right, as far as I know there won't be any more LearnLM standalone models, it's all built into Gemini models already. whenver there's a new Gemini flavor out it will have some new LearnLM goodness built-in; I just don't know how much the team advertises it... I can tell you there there's very active work on LearnLM improvements

[charred quarry]

Anyone here on X? Looking to connect with people who also care and use this stuff. Hopefully see what’s working from each other and what’s not

[vapid cairn]

I can give you a rlly good discord server for this stuff

DM me

[vapid cairn]

Nice, that's good to hear

[loud quartz]

unfortunately, not available in my area.

[toxic trout]

I am a vivid google geek but everytime I apply for one of the beta wait-list I keep getting not available in my area ..

Feels super bad honestly..

[unborn roost]

I'm sorry to hear that, and completely understand. If you don't mind sharing, what country are you located in?

[toxic trout]

İstanbul turkey lol

[fluid crow]

Does it already support Russian?

[unborn roost]

For context, a large part of the reason Learn About is only available in the US today is that sometimes countries have specific policies around usage of AI, and Google takes these very seriously, but it means that there's an impact on the engineering teams who have to sometimes change the product to account for country-specific policies; or, at least, verify that we in fact comply with country-specific policies. We're a small team so we've been focusing on the US for now just because, well, for Google the US is usually the default market and the first one were we test things. But I understand the frustration and we'll be looking to see if we could expand to a lot more countries (even if English-only) soon. (Incidentally, the way we determine "country" in this case is based on the browser's IP address).

[unborn roost]

No, we're US- and English-only for now

[toxic trout]

I totally get it, what blows in my case is that I not only studied in the us but spent 14 years and go back and forth between Istanbul a lot. But I do get that google doesn't cater specific to individual but rather has to account for some other stuff I understand

But it seems you guys have done a very cool thing here so good luck my best wishes

[fluid crow]

In principle, I have nothing against English. I don't understand people who hate Google for not adding other languages (especially Russian language) 🤷‍♂️. Fortunately, I'm not like that. I like any product you have released.

[heady nova]

Zen isn't about escaping from the world, but about engaging with it in a different way
Google Learn About

[heady nova]

VPN can solve some access. not everything.
Also a good practice generally

[fringe palm]

learnlm is so good

they should make it an actual API

that charges

[woven sequoia]

Is the current version of learn about the same as learnlm 2.0 in aistudio?

[surreal forge]

I got waitlisted. Is that normal?

The announcements here made it seem like it was GA, but only to US.

[unborn roost]

I don't know exaclty what revision of Gemini LearnLM 2.0 is using, but it's likely very close to what Learn About uses (which is Gemini 2.0 Flash)

[unborn roost]

That's correct. Are you in the US? If so, you should be able to use it. You may need to sign in to your Google account first?

[surreal forge]

I am in the US. I signed in with my google account.

Ahh, the complaint about using the wrong browser sent it into the "apply" mode. Now I'm in chrome and it shows up.

Oh, wow. I can see this coming in handy to learn things.

[late pulsar]

Hi, it seems the page for this tool is a bit broken in Firefox 😅

maybe that's not Google's first priority ofc xD but wanted to post here

it does seem Dark mode does work once you actually ask it something at least

there seems to be some contrast issues with it as well

[unborn roost]

Thanks for letting us know! Yes we have not tested very much on Firedox ☺️

[oak juniper]

i think if when learn about is finished and out of alpha/beta then it could be a cool thing to intergrate into google search

[native vine]

I think Canada, Australia and New Zealand would be safe options outside of the u.s to open up to more people.

[radiant hedge]

It is not only Firefox problem. I looked the same from my Arch Linux, but I used Google chrome

[silk herald]

it will be greate to have the option to use Learn About on my own collection of sources rather than the internet. Just like NotebookLM.

[noble bridge]

I learned about...this is cool 😎

[grave hatch]

Oh well...

[grizzled relic]

Does Google illuminate feature comes to Learn About?

[unborn roost]

Thanks, that is something we're considering!

[unborn roost]

No plans to integrate Illuminate yet, but thanks for the feedback!

[grizzled relic]

It would be a good feature for it just like NotebookLM that integrated it

[dense mauve]

I have a question can Google Gemini Advance Gems used image creation?

[robust turtle]

ask..add more details about the prompt details, images for,,

[unborn roost]

Hi, hmm, I'm not familiar enough with Gems to answer that, but maybe someone else here is!

[unborn roost]

Hi, can you clarify a bit more what kind of details you're looking for?

BTW we put out a few bugfixes and small changes to Learn About yesterday, including two prompt changes that I'm curious if anyone has feedback on:

• For questions that can be answered directly (e.g. a lot of homework questions), we now try to put the answer right up front, instead of making you read a whole explanation of how to get to the answer. This should work better for fill in the blank, multiple choice, math problems, etc.
• We now support answering queries in languages other than English, but unfortunately we still can only respond in English. I'm curious, for non-English speakers, if this is at all helpful, or maybe even worse than saying "Learn About is only available in English" 😢

As always you can use Learn About at g.co/learning/learn-about with a US IP address.

[dense mauve]

My reason for asking this question is I turn a Gemini AI gem into Gemstone anime model with text base designs by trying to create a enhanced prompt. I accidentally created a LoRA model that doesn't used images but google word and Gemini AI deep research. I wanted to know this function is a bug or a future Capabilities for Gemini AI models. I noticed the website allows my gemstone and Gemini AI Pro work together to create insane designs. But I might be wrong since I am still testing it out. I noticed Veo model on the internet can be supported by my custom Gemstone. But Veo tab on the internet was removed and I can't test that theory anymore. Well, I wanted verify this information before creating new Gemstone models before they get purge and become useless by the updates or upgrades. Here are examples of my anime gemstone model.

[unborn roost]

Gotcha, this is pretty specific and not related to Learn About, not sure if there's a better channel for this in this Discord, but maybe you could try in #general・🧺 ?

[dense mauve]

Okay because I wasn't sure where I can ask this question

[unborn roost]

Hi everyone, we have a few new features coming up soon on Google Learn About, your educational AI companion. We'd love to have some of you in our user research program, which will give you early access and a chance to send us direct feedback to help shape the product. If you're interested please sign up here! We'll likely be sending out the first opportunities next week.

Oh, also, we have an even shorter shortlink! g.co/learnabout. And you can use g.co/learnabout/<yourquery> to jump directly to a query, like g.co/learnabout/elephants

[lapis lodge]

@unborn roost Hey what kind feedback do you need this tool would be pretty good for law agent or cases to explain or find information related to a case..

[vivid aspen]

Will math be coming to learn about any time soon?

[unborn roost]

Thanks for asking, we're trying out some new features very soon so we'll likely be asking for feedback specific to those, but also any general feedback on Learn About is welcome any time, you can just put it here in this channel if you'd like.

[unborn roost]

Great question! Yes, we are working on adding math solver support, as well as proper Latex rendering (unfortunately right now we often just print out raw Latex which is really hard to read).

[sinful echo]

Is Whisk and ImageFx having an outage? Both are showing a message saying something went wrong, and when I sign out and try to sign back in they say "this is not available in your country"

I also made a new account thinking it had to do with a limit, but I kept getting the same problems.

[slim geyser]

Check if your country is supported https://support.google.com/websearch/answer/14184960?hl=en

[weary star]

Time to await another twenty years before Canada gets access to Search Labs 😔

[shut salmon]

How To Better utilize Flow To Its Full Potential Please How To Add Music And Sound And Everything

[lapis lodge]

@unborn roost , I would like dashboard for for like learning that you can connect knowledge based and provide your own image and context and have timeline with events and what relation it has to the subject and context window. Timemap for event and brainmap or clickable dashboard and under what circumstances and what timeline with events. Maybe for colab . @unborn roost @hollow depot @hard blaze

[unborn roost]

Thanks for the suggestion! Are you referring to bringing your content to LAX for grounding? We do have a much better PDF experience coming out soon, which may address some of your suggestions. If you want to get early access and a way to give feedback please sign up to try things out!

[lapis lodge]

@unborn roost maybe card based where you put in people and subject, and connect it different types databases and knowledge points. Cool with pdf experience, think more for law and sometimes when showing evidence and so it will be easy understand and see if someone's is innocent or when people have disability in talking so the evidence or facts show what really happened or so think that would be great this can also be good in medical trials and trying find a cure or showing up what is important and why it important. Maybe it's called grounding but it would work for Building webpages and show storyline and what should connect and why and why relevant. @unborn roost I sign in to the link but I would rather have something better that is available today chatgtp is pretty good when you wanna restrict the source's and subject. But think there many aspects where it call be usable..

[sonic sinew]

yo how tf do i make it so my veo 3 gens dont have subtitles....... like the videos are perfect where the hell are the subtitles coming from

[cyan sleet]

Yeah, mine were doing the same thing. I think there is a particular way you need to prompt the characters to speak properly. I’ve found if you just do like. The woman says “Oh look a butterfly” it will print the subtitles

[floral tide]

Great 👍

[unborn roost]

Hi everyone, a couple of exciting updates to Learn About! As always, you can try it out at g.co/learnabout (US-based IP address required)

1. We are now on Gemini 2.5 Flash, and with a larger context window. While you may not see an immediate quality change, the new model is quite fast, and will allow us to take advantage of more recent revisions of LearnLM to help improve quality.

2. We've rolled out our first big update for PDF uploads for desktop users, a built-in reader where you can browse the entire PDF. You can ask questions about your PDF on the side, and also highlight text in the PDF and get suggested queries. Check out the video 📹 for a little preview. And we're are going to be adding more AI-powered magic ✨ to the PDF experience over the next few weeks.

**Please let us know what you think in #learn-about・📖 ** Bug reports are very welcome too.

[silk herald]

@unborn roost can i upload more than one pdf and does it understand images inside pdfs?

[unborn roost]

@matias can i upload more than one pdf

[last star]

Hey guys is there a way to test this AI? I got restricted due to the location preference

[steady mango]

Hey you guys, did you know Microsoft's Learning platform for growing in tech and building projects that can make a different: https://learn.microsoft.com/?wt.mc_id=studentamb_467581

[unborn roost]

Hi Ahmad, unfortunately at the moment we're restricting to US-based IP addresses.

We're hoping to open it up to more countries soon

[hidden escarp]

When I try more complicated things, like "dotnet webapi with swagger and unit tests" - It worked for me just a few minutes ago. After the initial create, which works fine, I asked "add an endpoint" (in the proper way), and it's sitting on the build step. It says it was successful. And here's my question: It feels like Jules is "stuck" for lack of a better word. Is it true, due to fixed resources, that in this case, I'm just waiting for "my turn" so-to-speak?

[unborn roost]

Hi, you may want to ask this in #jules-general・🦑

[hidden escarp]

Okay will do thanks

[fallow wave]

Just curious, but how would this be different than using Notebook LM?

[unborn roost]

It's a great question. They are pretty different products at the moment. NotebookLM is purely and only grounded in what you upload, whereas LAX will always include and reference content from the web (search results, images, and videos), and works whether you bring your own content or not.

Would love to hear from people's experiences of using the same PDFs with NotebookLM and Learn About and see how the experiences differ, or when you might use one product over the other.

[unborn roost]

Hi all, we're trying out a new feature in Learn About, Comprehension Check. This is a new learning card that lets you test your comprehension by prompting you with a question, letting you type or record your answer, and then getting an evaluation and feedback from the model on your answer. There is learning science behind this principle as an effective way of going beyond just basic understanding and really cementing concepts in your brain! 🧠

The new card will usually appear after two or three turns, once there is enough to generate a useful question. It will appear just as a title near the bottom, which you can click on to open up the card. Below is a short screencast that shows it in action. We would love your feedback! Are the questions relevant? Does this feature generally seem useful, and if so, when would you use it (and when would you skip it)? Is the response feedback useful? Well explained? Too nitpicky? All feedback is welcome 🙂

[tall basalt]

Add history like Gemini. The lack of history is the only reason I don't even touch this. Ive tried it and it's super nice though.

[unborn roost]

We do have chat history! 🙂 You have to open the left panel to see it.
If it's not storing your chats, you may have Web & App Activity off, you can check here: https://myactivity.google.com/myactivity?pli=1

[ashen violet]

Hi. Does VEO 3 allow creating characters for videos and saving them in the AI's memory (in my account) so they can be reused in other videos?

[wispy patio]

Yes, Veo 3, particularly when used through Google's Flow interface, allows you to create characters and manage them for reuse across different videos.

[errant basin]

How does the student thing work?

[slate mural]

how can i use VEO3? i need help with it

[spare stag]

How can i use imagefx in africa ?? Need Help

[plush cosmos]

How can I get together in the picture, because I'm doing it with my troops with a lot of women?

[stray osprey]

What kind of help you need?

[slate mural]

I have paid google pro, gemini it's ok, unfortunatelly Flow upgrade doesn't work. why? how can i use the VEO3?

[turbid gazelle]

you can use VEO 3 by chatting with https://gemini.google.com/u/2/app and pressing the Video button under the chat. Check the settings button to make sure VEO3 is selected. Gemini videos have a watermark.

You can access flow and VEO3 here: https://labs.google/fx/tools/flow

[limpid summit]

hello any news about a new learn about update?

[limpid summit]

can an administrator answer this please?

[runic turtle]

how do we edit the output video please? the wordings i input is missing by an alphabet
thank you in advance

[plain eagle]

Unable to display images

[wispy rune]

how do i make yeti or bigfoot swear?

[limpid summit]

Hello, anyone here to answer?

[weak falcon]

Anything you need help with?

[limpid summit]

any recent updates or future news on learn about?

[weak falcon]

Well the most recent one was comprehension check that I know about, then before that they added 2.5 flash and some cool features for PDFs upload

Honestly, I feel like they should merge the two, NotebookLM and Learn About
As a student you do want something more than your textbook has and so having an instant switch from your textbook to the web to learn more about it could be great, I think the feature to add sources from simply the web is quite useless, I mean that deep research like thing, instead this is what they need

[limpid summit]

true, thank you

[primal stirrup]

https://www.wpbeginner.com/research/creator-economy-statistics-that-will-blow-you-away/

[rich reef]

Hi everyone,

I'm having trouble with the image-to-video feature. When I upload a starting image and add a text prompt, I get the notification: Switching you to a compatible model for this feature. Submit again to confirm (or check settings for details).

The problem is, even when I submit it again, it doesn't work. How can I get past this step and actually generate the video from my image?

When i use veo 3

[weak falcon]

VEO 3 fast is text to video only

That's the reason why you can't add images to make a video, you can try VEO 3 quality

[rare turret]

Hello guys, when i open the Flow page, i stuck on the screen that says "Flow A new state of creation", how do i "enter" the creation tool?

[dim lagoon]

Hello, I have subscribed to AI PRO to access Flow, but I get this window and when I click it, it says that you already have a subscription, how can I use Flow?

[hollow depot]

hello to some new faces 🙂 Flow is not currently available in all countries where Google AI Pro is offered.

Please note: VPNs will not enable access in unsupported regions. Check the list of Flow supported regions before subscribing.

If you’re in a supported region country and still experiencing this issue, please reach out to Google One support.

as a side note - we have a Flow section of the server dedicated to answering questions like these. for better visibility, please find the most appropriate channel for your question to ensure faster response times. here's an overview of the Flow channels:

#flow-announcements・📣 - a place for Flow-specific announcements. you can grab the Flow role in Onboarding if you'd like to be pinged!
#flow-general・🌊 - general space to discuss the tool
#flow-tips-n-tricks・✏️ - a place to share prompts, collab on them, or even ask for someone to generate your prompt
#flow-videos・📹 - a place to view and share Flow gens
#1319424187730755636 - a place to share things like this directly with the Flow team

[limpid summit]

any new features for learn about?

[elder jolt]

Yess hopefully they will update us with 2.5 pro (I wish)

[limpid summit]

i hope so, it would be amazing. do you work at google? i have some suggestions to make

[elder jolt]

Nope lol

I'm just someone like you haha

[limpid summit]

it would be amazing to be able to attach mind maps within learn about and when you click on a topic in the map generate a reply or description in learn about. All this without affecting the interactivity of the mind map in terms of opening other map topics.

[elder jolt]

Yea also latex should be fixed. It's one of the most important aspects of a studying site

[limpid summit]

I'm sure it would also be a great role.

[vernal mulch]

<@&1366831844502274180>

great job guys ❤️

[soft inlet]

<@&1366831844502274180>

[worthy mountain]

I am finally trying Learn About as well.

As it's experimental, I'm free.

[burnt wasp]

Very cool resource. Wikipedia killer

For philosophical figures, the tool needs improved links to source texts not just summaries of them

[marsh nest]

Are the features from Learn About being integrated into Guided Learning mode in the gemini app?

Would be cool to learn with the SoTA models

[green forge]

Hi! When is Learn About available in the Philippines? I really want to try it out like I'm so jealous of those who have access to it 😭

[still stratus]

Hi! I'd like learn deeper about AI

[worldly carbon]

Learn About is really cool I love it. I feel inspired to learn about new things and find out more from the suggested resources, although I have two small suggestions.

On mobile the buttons and colors don’t show properly and appear as plain text, maybe mobile isn’t the focus right now but it would be fun to be able to learn any time. For the desktop site, I love the colorful containers and interactive buttons.

Another suggestion, I’d love a read aloud feature for the definitions of Build Your Vocab words and maybe even all of the text in a response. The reading text feature that comes with my internet browser does not really work on the results of the page.

I think that would help me to learn more, because hearing the content and also seeing it visually would be helpful, and also allow more accessibility and suit people with different learning styles, maybe even allow for listening to the information like a podcast while focusing on other activites.

[split wraith]

Is there any free resources or course to learn ablut prompt engineering or Image prompting?

[slender nimbus]

hello, i need help, for over a month now, i havent been able to use any of Google labs models like whiskfx, imagefx and others. previously, i had been using the ai models, but recently, when i try to log in, it logs me out. i am in Malawi by the way.

[austere juniper]

Any news on Genie3 availability yet

How can I test it?

[vernal mulch]

@grand forge

[brittle atlas]

Hey! I’m new here and was wondering if someone could help me out. I have some doubts about whether we can generate images larger than 1024×1024. If it’s possible, what steps should I follow to achieve that?

[frank orbit]

Hi… I’m new here and really hoping that someone can help me out… I created an AI Influencer that I have been trying to use Flow to create video for but I keep getting a message saying that “Our policies prohibit the uploading of prominent people at this time.” The photos are based off of a description and are not of anyone but the AI Influencer that I developed. I have no idea who they think she looks like but she was AI developed. Also, I’ve made several AI generated photos of her with no problem. It’s only when I attempt to use a VEO3 application that this keeps occurring… Does anyone know how to get around this? Or who I can speak with in order to get past this?…. I have spent a lot of time and money developing this character. I don’t want for it all to become a waste because I can’t fully develop her the way that she needs to be!

[river saffron]

Howdy team! Learn About and Learn Your Way—psyched about both but there’s some overlap for use case, looking ahead, any thoughts on merging the 2 to avoid choice overload?

[hallow vortex]

Hey can anyone tell me What is different about Learn about

[near ridge]

Hello if you're encountering errors and /or bugs or would like to leave a general feedback about Flow please don't hesitate in utilizing the #1319424187730755636 channel with as much context as possible. It helps to keep the team organized and its also where all the technical issues are logged. Your post will get lost in the wrong channel. Thank you!

[worldly vessel]

https://services.google.com/fh/files/misc/learnlm_prompt_guide.pdf

[arctic cedar]

https://www.youtube.com/watch?v=sMtPIy-bkp0

Is there any video related to how this dude create so much consistent video using veo 3?
I want what prompt does this guy gave to veo 3

[hallow raft]

Would love to be able to use this system with my own internal set of pdfs and books

Unlike NotebookLM this does a great job of walking you through topics

[languid breach]

Hi, I get an 404 error on the website when clicking get started, any tips?

[hollow depot]

as an experiment, Learn About isn't available everywhere just yet. if you're not over 13, logged into your Google account, and in the U.S., you might experience this error. LMK if that's the case or not and I'd be happy to investigate further!

[limpid summit]

Do you have plans for other languages ​​and countries?

[stone brook]

Hi,
I used the app a while ago and wanted to express how wonderful the details and cards are. Feels like how I would imagine studying in future.
Although my recommendation is around the chat based view. It gives me a feeling "I clicked this cool card but wait it creates new content in the list. What if I click a few and I am lost in the chat scrolling all the time."
THIS DESERVES A DASHBOARD VIEW. A moodboard like carousel view with some different nested sidebar to track history.

I did a basic mockup and have attached the video on how the view could be. Would think a bit more on the nested sidebar for history

AGAIN! a wonderful product. Surely the next NotebookLM.

[sinful saddle]

Content blocked?
Looking for why this prompt/content got blocked. Seems pretty tame to me. Happened just now using Google AI Studio – Nano Banana gemini 2.5 flash image.

<@&1173646484311384114> -- Is this the correct channel to post inquiries like this? Will relocate this post if alternatives are suggested. TIA!

[hollow depot]

Content blocked?

[oblique geode]

Learning about AI

[spring prairie]

learn about in latam peru please

[weak falcon]

With Gemini, just casually chatting and not “reading it’s mind” (reading the thinking drop down) helps me at least, immerse myself in its predictions. It’s similar to learning a new language, overthinking it can impede progress. Just my experiences.

[last pelican]

I'd like to try this experiment. I'm using NLM so far to generate audio overview, some videos, and reports, but having a "teacher like" to guide me in every step would be amazing. I'm in the waitlist in the meantime.

[digital grove]

Positive relationship

[elder jolt]

Can we have improvements with the Latex rendering?

I think numerical solving experience is much better in gemini/notebooklm than in learn about because of the latex thing

it could make this app much better for students like me studying maths related subjects

[next sun]

hi

[meager nacelle]

Where can I report a bug?

[red lance]

Hello

[coarse wren]

Hello, When is this coming to India?

[molten badge]

india

[molten badge]

i am here.

[scenic flame]

Hi everyone, I have the Ultra account and I use Whisk for image generation, but it only produces images in 720p. Is there an upscale system available within Whisk, or in another app from the Google Lab AI suite?

[opaque birch]

Canada work?

[uncut ginkgo]

hi

[lunar flax]

Which is the best way to generate topology images or infographics? Without spelling errors..

[uncut ginkgo]

hi'

[graceful rock]

hi

[cosmic garnet]

hi, can i deploy my apps in google ai studio for free?

[abstract crane]

On Google Cloud I guess

[opaque hawk]

Hi

[modest echo]

Hello, I would like to try Learn Your Way.

I am looking to set up organizational charts for processes at the company where I work.

[surreal glacier]

I want to work in the creative field, what should I learn first?

[novel bough]

Rooted your phone and use.

[minor gale]

@vapid cairn@boreal jasperAs stated in #start-here・📍; currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.

[deft quest]

Hey ya'll, Vertex AI lets me host my own LLM models 4ight? Is it cheaper to use than Gemini's API?

[last tusk]

Hello

[opaque thorn]

does anybody know how to prompt that there is no camera zoom , i want a static shot but it alwasys ends up zooming

[fleet surge]

When is learn about coming to India? is it possible to get access early? I think it will change my life as a neurodivergent person - I will be greateful!

[elder quest]

When is leanring about coming to Switzerland?

[digital rover]

When is learn about coming to Pakistan? is it possible to get access early? I think it will change my life with many person I will be grateful. thanks

[elder jolt]

I wonder if Learn About will get a new LearnML model in the 3.0 series

It's one of the earliest labs experiment I remember

And the last update i could think of was 2.0 flash LearnML

[fleet surge]

hey, are you in india rn? if yes, can I DM you real quick?

[crude pewter]

Guys, does Veo have unlimited AI generations?

[agile stratus]

does anyone know what to do if your opal app reaches output limits? like how do i keep using it?

[elder jolt]

I'm sorry I'm not in India, but I can help you with what I know

[coarse wren]

Does this tool have access to India?

I have joined the waitlist

[coarse wren]

Hey, Can India above 18 join this program?

[elder jolt]

If it's not available in India, then not possible

you have to wait till it expands to India

[drifting field]

Hello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.

[wanton lava]

where can i generate videos

[final cairn]

google nano banana if you have a decent plan

[jade crane]

hey so I would really love to test and try the learn your way the lab, and submitted a form couple weeks ago, when can I expect a response or maybe access?

[silk swallow]

I am having trouble accessing Pomeli. It says to enter my website but there is no input box displaying. ANy help you can offer, thank you for very much

[weak falcon]

@silk swallow mee too

No one to hear

[coarse wren]

#pomelli・🍈 is where anything related pomelli is announced

Btw, Is Learn About supported in India? If No, When it will be expanded to India?

[worldly vessel]

No idea. It's been almost two years since learn LM was released as a experiment and it remains so..

Just use vpn and connect to US, to use it

[coarse wren]

I see

Thank you, Harsh

[worldly vessel]

Your bio is quite good. Where you mention you're a generalist, plus, you do personal projects... What are the ones have you worked on?

[ruby jewel]

@coarse wren His bio sparks intelligence and outcome

[coarse wren]

I have worked on making an Google Sheets Accounts Tracking System and a Opal App which saves some time in find which function is right function Google Sheets (Still Under Development I would say). I have also started on making a image pre-processing code in Colab for HTR Recognition

[worldly vessel]

Quite cool, and especially first one which seems unique idea

[coarse wren]

Thanks, What ya been doing here then?

[worldly vessel]

Long time learnLM user...
While doing any project, of which i would lack knowledge of, I use "things you should know" inbuilt feature of learnlm before proceeding with work

[coarse wren]

I see, Could be very useful. Can ya just share me a few screenies of how it is. Since, I couldnt use it at all since it hasnt expanded at all

Like, I dont know any proper VPNs to use (and they ae damn expensive)

[coarse wren]

Greetings, I am Revamp. It has been 6 months. Since the last announcement, When is this coming to other countries such as India? I would really like to know

[coarse wren]

Is this still the case? Like when is this will be available in India?

[mossy orbit]

Hi everyone! I'm new to this. Could you recommend the best platforms or tools for creating Al-generated videos?"

[coarse wren]

#flow-general・🌊 might be helpful i believe. Have a Nice Day!

[lime pagoda]

https://cdn.discordapp.com/emojis/1048062941964025997.png?size=48&quality=lossless&1

[foggy stag]

"Learn About" Feedback
@GoogleLabs | Proposal: The “Purpose Pivot” – Moving from Information to Activation

The Problem: Content personalization is excellent, but it hits a "retention ceiling" if it doesn't answer: “Why do I need to learn this for my life?” Whether it’s an elite student or a vocational learner, without perceived Utility Value, they "drift off."

The Solution: Focus on Directed Individualized Learning by solving the "Why" and the "How" together.

Key Feature Ideas:

The Vocational “Why” Engine: A toggle to reframe curriculum as a Mission Briefing (e.g., reframing Ratios through a Chef’s lens or a Nurse’s lens).

“Supervisor Briefing” Audio: Procedural, high-stakes audio between professionals on a job site rather than a standard lecture.

Digital Twin Avatars: A tool for teachers to create consistent digital presences, providing a "Safe Harbor" for students even when the teacher is absent (critical for trauma-informed environments).

Linking Storybook to Learn About: Allow "Narrative Skinning." push a lesson into the Storybook Gem so students can learn via Anime, Comic Book, or Cyberpunk stories.

Full detailed proposal with STEM examples and technical fixes (including Storybook PDF printing bugs) available!

[foggy stag]

The "Purpose-Driven UI"
Subject: Solving Text-Density and the Relevance Gap in Learn About

1. The Problem: The "Wall of Text" Barrier

Observations from the Classroom: Large paragraph chunks increase cognitive load and trigger task avoidance. In a world of short-form media, students need information that is "scannable."

The Attendance Link: Students stop attending (mentally or physically) when they cannot see a "Return on Investment" for their focus. Text-heavy summaries without clear utility feel like "busy work."

2. The Solution: Instructional Scaffolding (The Alchemist Framework) I propose that Learn About adopts a UI structure similar to my Project Alchemist science booklets.

Bullet Point Mandate: Every concept should be broken into "Three-Point Intel"—What it is, How it works, and Why you'll use it.

Visual Hierarchy: Use clear headers and "Sanctuary Margins" (white space) to let the concepts breathe.

3. Direct Interest Linking (The "Why" Engine)

The Logic: Instead of a generic summary, the UI should offer a "Lens Toggle."

Example (Topic: Volume of Cylinders):

General View: (Current text-heavy explanation).

The Tradie Lens (Bulleted): * Calculate concrete for a post-hole.

Avoid over-ordering and losing profit.

Result: Math becomes a tool for a $50k job.

The Medical Lens (Bulleted):

Calculate volume for IV fluid bags.

Ensure exact dosage delivery.

Result: Math becomes a life-saving protocol.

[foggy stag]

to the person that laughed, i hope you dont have to teach traumatised students, children who constantly check communication for whether someone is 'coming' for them, children physically, mentally, emotionally abused. when you work with these kids, you'll do anything and everything to help them

[coarse wren]

Greetings, Your work is Interesting. I am a Student Myself. I do learn Information & Communication Design on my spare time. I am also doing Desktop Publishing as an Vocational Student.

I do face the same issue of you have mentioned here in a much broader scale than students in specific. I really like the idea you are proposing.

One Question remains for me. How will these lens be accessed in the first place and is there any way to modify it and also how does it understand what the student is looking for (ie Information Scent).

Without Proper Prompting or UX Changes, Achieving this is quite difficult. In this Case, Prompting should be out of the window

[coarse wren]

What is necessary to create a lens? What information should the lens contain to do what it is meant to do. Also, I highly recommend to use Opal to have it as a Slide (if it is going to be used digitally) and would really love to implement some sort of a proper rubrics which can be customizable for various SRL Purposes

[foggy stag]

Hi. Thanks for the interest. Lens is Microsoft Lens an app you can get. It acts as a scanner for handwriting

If there is anything at all I can help with just ask

Sorry wrong lens. When I use the term "lens" in the work above it means "the perspective of". So if someone wants to be a doctor we need to present data to students through a "medical lens" information provided purely relating to medicine

There is no point explaining to someone who wants to be a doctor how the information helps builders. I hope this makes sense

[coarse wren]

Thats why i am asking? What Kind of Info someone needs to make their own perspective? It makes a lot of sense. Lets say, None of the Perspectives doesnt fit and we need a new one.

What do you think we should provide to the AI to fit the Perspective is what i am asking?

[foggy stag]

That is the problem. I teach in a special assistance trauma Informed school. We have maximum 15 kids in a class. For this to work I'd have to get storybook to make 15 books. So I focus on 4 main areas "Pillars" that most of my students will most likely get jobs in. Trades. Hospitality. Retail. Service and Life. Or you make one. Copy it into Gamma AI and have it individualize for each students.
BUT. do all your students need this sort of pure individual work. Probably not so you don't have to individualize for everyone
BUT who do we then individualize for the low ability the high ability? High ability and medium are already working so focus on your worse case students
BUT does it always work. No. Some kids just don't want to work. They put more effort into not working than the effort required to do the work

[coarse wren]

I see, Thanks for the Information

[foggy stag]

Do you need anything else. Do you teach kids? Feel free to ask anything else

[coarse wren]

No, I said itself in the first interaction of ours. I am a Student

[coarse wren]

How long have been using Gamma AI? What do you mean by individualize (you put relevant experiences? Anything else?)? Since, I am a Self-Regulated Learner as of now

[foggy stag]

As a student prompt your topic then what career or hobby you like. Then ask for the information to be displayed how you like it
I prefer bullet point
Doesn't have to be full sentences
Ask for lots of images
Try using Google Notebooklm. It has way more options for displaying info
Infographics
Presentations
Videos
Audio
Quizzes
Mindmaps

Get Google Vids to make videos for you on topics

Use Notebooklm to gather all your information

[coarse wren]

Okay, Got it. Btw, You can use Another one of Labs experiments to handle AI Slide Generation. Opal has had an Google Slide Output where you can do it

[foggy stag]

Cool cheers I'll try that

[coarse wren]

Cya Later

[fleet surge]

It was me. I only laughed because the proposal seems to be AI generated, or at least reads like it.

Language is a poor tool to communicate concepts and thus you can't rely on a LLM; it lacks the rich schemas we have in our head. Language is a compression tool with badly defined rules. If you don't understand what I am talking about, you can either read Wittgenstein or ask your fav AI assistant.
Sure, an LLM might synthesize research to do good brainstorming but still, it would be much more trustworthy if you wrote it yourself with your own subjective experience. People are growing tired of LLM generated content, at least the ones I talk to. It affects credibility.

I am sorry for mocking you, I should have been more constructive.

[potent shore]

Hello everyone

[foggy stag]

I use an LLM to make sure 8 get a precise and concise version of what I wish to say. I'm not an expert in technology terminology but as a teacher from England I'm pretty confident in my use of English. That was probably the tenth draft I had used in an LLM to make sure I use the language the correct language of the recipient not the sender.

[coarse wren]

Greetings, How ya doing?

[hot dune]

I think you are already getting to a core skill: refining prompts to a desired outcome rather than relying on the first output.

If you haven't already, you might want to try Canvas in Gemini. Fantastic tool for refining text, slides, whatever you need done.

[foggy stag]

Cheers for that. Ive never used Canvas in Gemini. What's it best for

[hot dune]

I've found a lot of use putting together documents. For instance, you can prompt Gemini on the subject of an article you would like to write, and it will create a rough draft of that document in markdown format. You can review it right there in the Gemini app, make instructions using voice or text, and it will make those edits for you in real time.

When you're satisfied, you can then export to Google Docs and receive a formatted copy for your use and further export

[foggy stag]

That will be great for my teaching. Thank you very much

[hot dune]

No worries!

[hot dune]

You can also execute deep research projects in Gemini on specific topics. It will also produce a research paper with sources for you on that topic. You can then export that to NotebookLM. You can also do most of this in NotebookLM itself now, but if you just want the article the Gemini app is probably the best bet

[foggy stag]

Thanks. I do a lot of research on notebooklm. Then specialize in individual notebooks then export in to Gemini. Also looking at other Gemini tools such as Learn your Way and Storybook Gem. Being 56 I'm very keen on AI to help my students in a Special Assistance school in Australia. Many of my students diagnosed with mental emotional and psychological issues. Trying hard to make learning fun

[hot dune]

NotebookLM has a lot of great education tools, and you can use those notebooks in Gemini as a knowledge source

[foggy stag]

I love the tools in notebooklm. Found making infographics in notebook then refining in Gemini works best. I'm making vids on Google vids. Thinking of making an avatar for the videos. I make booklets for my kids and used to individualize personalize them in GammaAI or Genspark. Any ideas of how I could do it in Gemini

[hot dune]

Nano Banana Pro is pretty sophisticated an image generator at this point, and has largely solved the text issue. Infographics in NotebookLM still have problems with it, I find. There is also Whisk, an image generation and editing tool in Labs that I find is good with text as well.

For videos, Flow is capable of generating 8 second videos, but that can be extended in a dedicated scene builder

I haven't tried exporting Flow videos to other generation tools yet, so you might want to ask around if you need longer videos

[foggy stag]

Google vids does over 3 mins. I can copy in a PDF and it makes the video and I can edit videos and images. It's Gold

[hot dune]

I'll have to check it out then. That might be the better tool

[foggy stag]

Yeah I keep forgetting how I do it. I think it's the last option of different suggestions but you copy in anything and edit

[hot dune]

I imagine there is a VEO generator built into the app

[foggy stag]

Go start a new video then go to Storyboard

[hot dune]

Yeah I see here. There is VEO, voiceover, and Nano Banana generators on the sidebar

That's super cool. Thanks for mentioning it

And it was right there in the Drive menu this whole time too 😄. Completely overlooked it.

[foggy stag]

Well you helped me with Canvas

[hot dune]

Fair enough XD

[near ridge]

Hello @bitter pike as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thank you!

[delicate flicker]

hello guys can any one help to use google labs how can i make an ai vdo

[hot dune]

You can try Flow, found here:

https://labs.google/fx/tools/flow

[ancient ingot]

Error comes multiple rimes

[hot dune]

I'm sorry?

[coarse wren]

Any Issues regarding Flow. You can check #flow-general・🌊 and #1319424187730755636

[ancient ingot]

Don't be sorry bro it's not your fault
Any trick to solve

[hot dune]

No I mean "I'm sorry" as in I don't understand what you're asking me

[coarse wren]

Irfan, What kind of error you got? Send it #1319424187730755636 Only if it is related to flow.

Post it here, If it is related to the Google AI Tool of Learn About

[ruby jewel]

Can anyone explain what Google Labs does to help your everyday life?

[coarse wren]

Idk really. We each have our own usecases and this question is really great in #general・🧺

[atomic pier]

Just missed your poll. I'm planning to start one soon. Just missing one thing, motivation lol.

[dark smelt]

Hi, What “learn about” doing better than NotebookLM ?

[hot dune]

I would say speed and flexibility are the big two

[rotund trellis]

I've been working on my Cisco Modern AI badges and practicing Python to build my own security shields! 🛡️ I wanted to share my first custom script to say hello to the team. Huge respect to Demis and the Google Labs crew! 🍄

boss_name = "Demis"
print("Hello " + boss_name + "!")
print("I am Laura, and I am learning to build AI!")

[hot dune]

Python is great because a lot of Gemini vibe coding defaults to Python as a language

[onyx spoke]

When is the launch in india ?

[coarse wren]

Probably This Year, I think

[lethal tide]

Jhvb

[stark garnet]

I already run one.

[real sleet]

hey guys how can i download png format image in flow nano banana

[formal quail]

Why not convert it

After saving

With another llm that can or a sinple website format converter

[formal quail]

^

[gentle breach]

I made a mistake investing money in this artificial intelligence research. Things are getting worse; image creation errors occur constantly, and image downloads are very poor. I feel stupid and disappointed. 😡

[muted nacelle]

Somebody off the device that's why you are getting an error

[gentle breach]

Could you please explain it more clearly for me?

[muted nacelle]

Press ctrl+f5 try again

[tribal bramble]

I have a YouTube channel, and I was really interested in making, AI generated content for it. We were working on a longer form classic brought to life . I've only uploaded a few shorts though. Well our content is ai generated it is definitely multiple humans in the loop... But it seems like the algorithm might have been recently weighted really against it? ...so I have just quit uploading anything.. trying to figure out what's going on... And if it will be a hostile platform or not.

[long nymph]

It might be an algorithm thing. What were your expectations?

[tribal bramble]

I'm sure it was an algorithm thing.
There's been a strong outcry against AI content and I understand that, but also know that I myself really like it, often even prefer some of it. So for now we only creating locally and not posting.

[outer folio]

I think you should post it. They'll come around. Most folks dont see success on youtube for months or years of consistent uploads. i would't give up so easy. I'm working on more vids now. Im setting up automations to help me. Samson The Farm Dog on youtube. He's an AI representaion of our dog Samson who live on our farm.

[twilit anchor]

how does one go about monetizing such videos ? or would it be impossible to do such a thing in youtube ?

[vernal terrace]

Hey recently I noticed that in gemini dynamic views in not showing so I think it is removed in your pc is it showing

?

[outer folio]

I dunno Im new

[crystal crescent]

Hit the requirement, it has like 500 subs, 3 Uploads in the last 90 days and 3M shorts views ig and that's it, ig no copyright issues most of the time

Am I right coz I'm new

[vital goblet]

How do i use this platform my internet is peaked

[crystal crescent]

Just use it

[iron tapir]

i am new here

[drifting field]

welcome!

[young nova]

Hi, i want to build a website with responsive animations, what are the best ways to create and deploy it? I am now looking at STITCH, but i wonder what other tools are available there and what are the best options to deploy a vibe-coded website. Thanks in advance ❤️

[stone spire]

Orchids.app

[deep fable]

Hi, have you taken a look at Google Antigravity?

[young nova]

Yes, ofc, as well as claude code, but wonder what else is there

[tranquil heath]

It wasnt working for me last week. I'll have to see if the servers are hitting me this weekend.

[tranquil heath]

He'll yes, python.

boss_name = "Demis"
print(f"Hello {boss_name}! I'm GasMan - Cisco badge earner, Python builder! ")
print()

Meet the crew

deepmind_models = ["Gemini", "AlphaFold", "AlphaCode", "Genie", "Lyria"]

print(" DeepMind Hall of Fame:")
for model in deepmind_models:
print(f" {model} - legendary.")

print()

[vale wind]

Hey im new here and I want to use AI to help with research for work and help organize some stuff in my life, where should I start?

[halcyon sentinel]

Hi Community,

I am currently working on a bot detection project using Causal AI (specifically the DoWhy/GCM library). My architectural goal is to build a causal model (DAG) based on Google Analytics 4 parameters, such as ASN, User Agent, mouse behavior, path consistency, etc.

The Challenge:
At this stage, I do not have direct access to a "live" data stream from BigQuery/GA4. To train and validate the model, I am generating synthetic datasets. However, initial tests show that the model tends to follow the path of least resistance—overfitting on infrastructure-level features like ASN_Type while largely ignoring complex behavioral patterns.

Questions for the experts:

Feature Nuances: Based on your experience with GA4, which "hidden" dimensions or parameters are the most difficult for bots to spoof? (e.g., specific combinations of screen resolution vs. touch support, or certain custom event sequences).

Adversarial Data Structuring: What is the best way to structure "Adversarial" synthetic data to force a causal model to prioritize behavioral signals over technical/infrastructure ones?

Realistic Noise: Are there any open-source references or "noise profiles" for GA4 traffic that could help make synthetic data more "messy" and representative of real-world edge cases?

Looking forward to your insights!

[tranquil heath]

Why dont you just build out the dataset...its going too take acouple of extra steps, but once you get the flow down for the columns and etc, you'd have a better actual dataset.. If you do this please record your sources and document as you go, is my recommendation.

[halcyon sentinel]

@tranquil heath Thanks for ansver)

[tranquil heath]

He'll yeah, go knock it out the park @halcyon sentinel

[halcyon sentinel]

@tranquil heath Thanks for the tip)

I've actually started using the GA4 BigQuery sample data as my base. Now I'm going to 'inject' some specific bot patterns (like robotic timing and rigid paths) into it to see how the model reacts. I'm also looking for more bot-specific datasets to make it even more robust.

Thanks again and have a great day)

[tranquil heath]

I would love to see more of your work if thats okay of course. See if i can help out in maybe other datapoints your at as well.

Thx!

[tranquil heath]

Depends on your work, and all.
You should conduct some R&D in your trade, and see if you notice things that are obivous to you and start digging there.

Thats what i would do if i was new and wanted to impress the boss.

[craggy flume]

Hey guys,

I'm a little new to AI besides claude and GPT so I'm finding it very easy to get lost.

Currently researching the AI website development business model for local companies and businesses and was wondering if anyone here is doing something similar to that.

If not I would still love to hear from anyone with knowledge on this AI as I'm currently a college student in my second year doing civil engineering and still haven't truly set my eyes on a stable future plan.

[tranquil heath]

Here's some advice: https://docs.google.com/document/d/16IrJKSZHCgI3lRxrApsHSIMbDvi4iTLiTGWXpjvLnow/edit?usp=sharing

& A quick recap of it as well.

@craggy flume

[craggy flume]

tysm

[tranquil heath]

https://tenor.com/view/batman-cape-darkness-solo-thswessel-gif-14510993171327393706

[shadow rain]

https://tenor.com/view/cat-bird-bird-cat-pointer-finger-point-finger-cat-bird-point-finger-cat-gif-2676038606799380578

[west flint]

i need somones big brain

[jade pilot]

hello everyone.

[broken fractal]

curious about tools

[wild spear]

i actually do something similar -

1. i do custom development using vibe coding tools
2. i'm just deploying an MVP of a multitenant vibe coding platform, similar to lovable and replit. Great for small business owners if you can inspire them with few easy to do ideas - from customer facing apps, through marketing automation to backoffice apps helping them run their business more efficiently. Its all about providing value for those 20 bucks a month

[crystal crescent]

thats so cool

u r changing their lives for 20 bucks a month

[sacred trail]

It's worth asking, as the world gets swamped with AI content that grows exponentially (socials, web, blogs, movies, podcasts and more), what does that mean for this kind of work. Every individual piece of content becomes worth less, a drop in a giant ocean, signal gets lost in noise. So what's the role of a website in all that? Maybe content is free, but setting up a company a presence, matching socials is still an effort, and a brand and reputation finds a location there. But what website development is, will change. Vanilla content and code is worth pennies, design instincts still matter (for now), client relationships are good to help people understand what tech exists (maybe that is a role now, but it's temporary), but navigating the core problem of getting seen, how to rise your signal above the noise, those are actual problems becoming more difficult for a business over time. That's different than building websites. This isn't really advice, but it's more just sharing how I try to look at things, not fitting myself into a past that's kinda becoming irrelevant, but trying to understand the future so I know where to go and how to think about it.

[topaz trout]

Hii there, actually, I am currently learning how to use Google Flow. Can somebody help me with adding music to its background? #learn-about・📖

[dire mist]

if the noise is increasing, the ability to discover things within greater noise is also improving and I would argue its improving at a faster rate than the noise

[steady axle]

When I discovered this tool, I almost cried with emotion.

[austere falcon]

Hey again! I am trying to learn how to make a mixboard with Google Gems so that I can coordinate with other people on a curriculum basis, can anyone help?

Along with this, I made a new type of AI where you can find responses quickly if somebody asks you a question!

Any tips on how I can update this AI model?

[warm meadow]

hi there

[solar oracle]

How can i join

[frozen jackal]

Hi All,
I am new in Al. I am doing one project for a song, I have generated images and then I generated video of that images. Now the problem is of lipsing and action part in the song, I am not able to do lipsing part perfectly so can anyone help me in sharing the tutorials and tools of lipsing so that I will watch and try to do lipsing part.
Thank you

[warm meadow]

hello

[gray shard]

Hey can any one tell how to use the Antigravity with agent skills before starting and project adding stitch , claude, ui/ux and different skills integrating in the Antigravity

[rough rose]

hey，brother，
Here’s the simple version: Antigravity loads skills from a local folder called .gemini/skills. Each skill is just a small folder with a config file and one entry file. You drop your skill folder inside .gemini/skills, reload the agent, and it shows up automatically. The official docs explain the exact structure and a few examples, so you can follow them directly: https://antigravity.dev/docs/skills.

[rough rose]

If you already have skills from Claude Code, Stitch, UI/UX tools, or other agent systems: just wrap them with a tiny adapter. Keep your original logic, expose one clean function as the entrypoint, and describe its inputs/outputs in the skill config. For API‑based tools, call their API inside the skill and return the result in Antigravity’s format. For local code, move it into a folder under .gemini/skills and add a minimal config. No heavy migration — just a thin wrapper so Antigravity can call it normally.

[rough rose]

https://antigravity.google/docs/skills

[gray shard]

Ok, thanks your ur help!! 😁

I have also found some repos on GitHub having 1400+ skills so m thinking to clone it globally and use it in workspace .
Is there any good repo if u can share which has skills?

[rough rose]

I usually avoid sharing commercial aggregator sites here — they’re not very reliable and not really aligned with the engineering focus of this channel.

If you want high‑quality skill or agent patterns, Google’s own open‑source repos are much better:
• google-gemini-samples (official Gemini + MCP examples)
• google-deepmind/agent-framework (Google’s agent architecture patterns)
• google-deepmind/alphafold or graphcast (for high‑quality structured pipelines)

Curated > massive dumps.

[gray shard]

Ok, will look into it 🙂

[minor gale]

@quiet archHey! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!

[quiet arch]

@minor gale Hej!Of course, I understand. I don’t speak English very well, so I wrote in my own language. But I will follow the server rules and translate my messages into English, just like I translate every post for myself. Thank you for understanding.

[orchid finch]

Yo

I wanna learn about ML ai and biomedical ai

@quiet arch @minor gale

Where should I start

Where to learn from

And I wanna do projects can you guide which softwares to use and which projects would be good to make

[vapid ivy]

learn about is a literal web app, it's not a channel to learn about stuff, sorry

[mighty saffron]

guys anyone teach me how to make ai videos

[hushed ridge]

same