Wow, I haven’t used Gemini for a while, and it looks like there are a lot of awesome new features here. I really like conversation branching, this line, “No more choosing between following a new idea or protecting existing progress,” really describes how I feel when I’m deep in a long conversation and need to try multiple approaches.
By the way, with these features, it seems like Google is focusing on developing their AI in the direction of learning and research. I think this is a pretty good approach to compete with other tools.
This is a great reminder that real innovation often happens quietly. While the spotlight tends to follow the flashiest announcements, Gemini has been steadily adding features that truly enhance the way we work. These are the kinds of updates that not only impress but also make a real difference in everyday productivity. Sometimes the most meaningful progress is the kind that builds in the background until it becomes impossible to ignore.
Google’s comeback is wild to watch. We always knew they had the infrastructure, but now it feels like they’re really activating it, and the systems they’ve built lately are just next level.
What I love about what they did is that it’s always about practical, deeply integrated tools that actually reflect how people work.
This post is such a solid breakdown, Wyndo. We all chase the next shiny thing, but there’s so much unexplored power in mastering what’s already at our fingertips.
YouTube video summary is one of my favorite features of Gemini. The best part is that it's not just using the transcript, it knows what's happening on screen. (Test this out by asking, "What's on screen at 6:32 in this video."), so you can get much deeper information if you need it.
You can also automate this by piping the URL into n8n, then save the information in a spreadsheet, database, or just have it reply to the email with the info.
Didn't realize the audio overviews had made the leap over from NotebookLM. Thanks for the tip.
yeah, esp if you use Gemini flash it's so fast to get summary on the video!
Oh I didn't know it was that detailed but I can see why because for every summary I ask, it generates specific timestamp that explains specific lines in the video.
Yeah it can be integrated with many AI automation, I mainly use Gumloop at the moment. Massive game-changer for sure, thanks for sharing :)
It appears I'm unable to attach a screenshot to a comment, so further to my last one, here is an example of a webpage Gemini made for me in a *single click* from a large study I completed in 2024, 'Pathways to Prosperity', it's an analysis of the economy we're moving toward. There are some amazing tools it automatically built like a drop down economic model comparator and a summarizer.
This makes me super excited. Ive found Gemini far superior for writing collaborations. ChatGPT, no matter how meticulous the prompts will almost certainly AI-ify my voice into banality. GPT also seems obsessed with recommending edits for their own sake and forcing my writing into max-efficiency LinkedIn mode.
Gemini alternatively leaves my voice totally intact if I ask for edits and rewrites (‘show me a different way to say this’; ‘show me what this draft could look like if I made XYZ changes’). It’s also far more capable of remixing and editing without rewriting — I can be a bit recursive in my first few drafts and it will take the draft, shave 30% and reduce redundancy without altering my voice or expression. It’s precise and surgical relative to GPTs tilt toward slop.
I would love to hear you speak to why that is? Regardless I love it.
ALSO, something I accidentally discovered I didn’t see here, directly in Gemini it can create webpages, infographics, quizzes or audio over views just from a single click while in chat 😮🤯👏🏼 it was a total accidental discovery. Look for the ‘create’ button on the web chat interface. It was just mind blowing. I’ll see if I can add an image to a comment to share what it did - but took a deep research report it had just generated and it gave me the 4 options from a drop down menu. I picked webpage and got a fully functional interactive site from a 17 page report in a single click. No prompting, no debugging. One click. I’m still picking my jaw off the floor 2 weeks later.
Thanks so much for your posts, Wyndo! I'm learning a lot. Glad the algo put you on my radar.
Good to know you are having better exp writing on Gemini. My writing mostly happening on Claude but haven’t done any in-depth comparison with Gemini but I have given up with ChatGPT 😅 as I didn’t get as I expected.
Yes, those 4 buttons on right top navigations are magic as it can turn your document into infographic, etc.
I created the financial graph from there. But u can also do it by typing it “create infographic/webpage/“ and it will turn it for you. It works like a magic even if you give it using vague prompt.
I love the “your next 30 minutes” callout. There is so much great content, but we do need to pause and actually try out what we’re learning, apply it, move it from short to long term memory. Good work!
yes. after all, the real learning comes from the new behavior we do from the information we consume. and the best way to do it is just to apply even if it's small!
Great post. Just to clarify, when it comes to analysing YouTube videos, how is Gemini better than notebook LM, which can also take a video link along with other inputs? Also, I recently wanted to get a comprehensive analysis of an entire youtube channel, not just a single video, and was able to just input the name of the channel into Perplexity labs and didn't have to specify the url of every video on the channel, whereas Notebook LM seems to need the specific url of the video.
Yes, NLM is much superior for this, but NLM is more suitable for research with multiple sources, but if you just want more speed of the output, then Gemini is the better choice.
It really depends on your goal. I'd go to Gemini if I just want to learn a podcast content fast. But if I want to write an essay and in-depth research, NLM is the answer.
Awesome post, thanks! I love this different approach and how you highlight their ecosystem vs pure power angle. I'm intrigued by how you're using voice + screen share to really collaborate with Gemini. I really want to try that now!
Right now my main stack is ChatGPT and Manus, but I'll have to consider adding Gemini to this...also looking at Perplexity. But I'm impressed with the features and mindset of these Gemini updates. Thanks again for highlighting this!
Glad to know this was helpful for you. Yeah, I guess the only way to learn is try bunch of things until you find which suits best with your style. I love Gemini because it's integrated with their ecosystem, multimodals and long context window.
Thank you Wyndo, I appreciate your time and effort. I need to check the conversation branching you mentioned here. I also agree, we shouldn't eliminate any chatbots or tools, but learn to integrate them all to the best possible way to get our projects done. I also still use ChatGPT, and sometimes Grok for my work, but as you mentioned here, Google AI Studio has really outdone itself with its new approach.
That’s a ton of awesome experience and insights you’re sharing, thank you!
I really appreciate the all-in-one ecosystem Google is offering. It’s kind of wild to realize how seamlessly they’ve been able to support this, largely because of the foundation they’ve already built over the years. Honestly, only Google could pull something like this off at this scale.
And that project branching strategy is exactly what I’ve always wished Cursor would do. Sure, it technically can with Git, but having it in a user-friendly interface like that makes a world of difference.
Wow, I haven’t used Gemini for a while, and it looks like there are a lot of awesome new features here. I really like conversation branching, this line, “No more choosing between following a new idea or protecting existing progress,” really describes how I feel when I’m deep in a long conversation and need to try multiple approaches.
By the way, with these features, it seems like Google is focusing on developing their AI in the direction of learning and research. I think this is a pretty good approach to compete with other tools.
Gemini, esp AI studio still underrated tools tbh.
I’m sure there will be cooler features soon!
Yes agree with u. They go to the direction of learning, research and ecosystem for work.
I found this extremely insightful, thanks for sharing!
This is a great reminder that real innovation often happens quietly. While the spotlight tends to follow the flashiest announcements, Gemini has been steadily adding features that truly enhance the way we work. These are the kinds of updates that not only impress but also make a real difference in everyday productivity. Sometimes the most meaningful progress is the kind that builds in the background until it becomes impossible to ignore.
Google’s comeback is wild to watch. We always knew they had the infrastructure, but now it feels like they’re really activating it, and the systems they’ve built lately are just next level.
What I love about what they did is that it’s always about practical, deeply integrated tools that actually reflect how people work.
This post is such a solid breakdown, Wyndo. We all chase the next shiny thing, but there’s so much unexplored power in mastering what’s already at our fingertips.
Indeed.
Given google products are all around us, there's so much opportunity to make it work. Hope they are doing it right!
Yes, sometimes we dont need another shiny tools, we just need something that really works and impact our work directly.
If you don’t want to pay for custom GPTs, the gems feature is great to set up for different usecases to store your system prompts
yes, great alternative compare to Custom GPTs, thanks for highlighting this too :)
I like reading snippets, so I use Gemini often to get the tldr of long YouTube vidéos often. Great pointers, Wyndo!
this feature is truly time saver :)
YouTube video summary is one of my favorite features of Gemini. The best part is that it's not just using the transcript, it knows what's happening on screen. (Test this out by asking, "What's on screen at 6:32 in this video."), so you can get much deeper information if you need it.
You can also automate this by piping the URL into n8n, then save the information in a spreadsheet, database, or just have it reply to the email with the info.
Didn't realize the audio overviews had made the leap over from NotebookLM. Thanks for the tip.
yeah, esp if you use Gemini flash it's so fast to get summary on the video!
Oh I didn't know it was that detailed but I can see why because for every summary I ask, it generates specific timestamp that explains specific lines in the video.
Yeah it can be integrated with many AI automation, I mainly use Gumloop at the moment. Massive game-changer for sure, thanks for sharing :)
It appears I'm unable to attach a screenshot to a comment, so further to my last one, here is an example of a webpage Gemini made for me in a *single click* from a large study I completed in 2024, 'Pathways to Prosperity', it's an analysis of the economy we're moving toward. There are some amazing tools it automatically built like a drop down economic model comparator and a summarizer.
https://gemini.google.com/share/63a5ba47437e
(this link claims to be publicly shareable but I've been told it doesn't always work. Fingers crossed.)
This looks great Jenn!
Thanks for sharing.
U know whats cooler?
You can ask it to create a “download button PDF” so u can download the infographic to pdf entirely :)
Wait whhhhaaat?! haha now I have to try that. Just a text prompt? I'll mess around and figure it out. Thanks! Appreciate learning with you :)
Indeed, just a text prompt :)
This makes me super excited. Ive found Gemini far superior for writing collaborations. ChatGPT, no matter how meticulous the prompts will almost certainly AI-ify my voice into banality. GPT also seems obsessed with recommending edits for their own sake and forcing my writing into max-efficiency LinkedIn mode.
Gemini alternatively leaves my voice totally intact if I ask for edits and rewrites (‘show me a different way to say this’; ‘show me what this draft could look like if I made XYZ changes’). It’s also far more capable of remixing and editing without rewriting — I can be a bit recursive in my first few drafts and it will take the draft, shave 30% and reduce redundancy without altering my voice or expression. It’s precise and surgical relative to GPTs tilt toward slop.
I would love to hear you speak to why that is? Regardless I love it.
ALSO, something I accidentally discovered I didn’t see here, directly in Gemini it can create webpages, infographics, quizzes or audio over views just from a single click while in chat 😮🤯👏🏼 it was a total accidental discovery. Look for the ‘create’ button on the web chat interface. It was just mind blowing. I’ll see if I can add an image to a comment to share what it did - but took a deep research report it had just generated and it gave me the 4 options from a drop down menu. I picked webpage and got a fully functional interactive site from a 17 page report in a single click. No prompting, no debugging. One click. I’m still picking my jaw off the floor 2 weeks later.
Thanks so much for your posts, Wyndo! I'm learning a lot. Glad the algo put you on my radar.
Good to know you are having better exp writing on Gemini. My writing mostly happening on Claude but haven’t done any in-depth comparison with Gemini but I have given up with ChatGPT 😅 as I didn’t get as I expected.
Yes, those 4 buttons on right top navigations are magic as it can turn your document into infographic, etc.
I created the financial graph from there. But u can also do it by typing it “create infographic/webpage/“ and it will turn it for you. It works like a magic even if you give it using vague prompt.
Glad u find this useful! :)
I love the “your next 30 minutes” callout. There is so much great content, but we do need to pause and actually try out what we’re learning, apply it, move it from short to long term memory. Good work!
yes. after all, the real learning comes from the new behavior we do from the information we consume. and the best way to do it is just to apply even if it's small!
Great post. Just to clarify, when it comes to analysing YouTube videos, how is Gemini better than notebook LM, which can also take a video link along with other inputs? Also, I recently wanted to get a comprehensive analysis of an entire youtube channel, not just a single video, and was able to just input the name of the channel into Perplexity labs and didn't have to specify the url of every video on the channel, whereas Notebook LM seems to need the specific url of the video.
Great question!
Yes, NLM is much superior for this, but NLM is more suitable for research with multiple sources, but if you just want more speed of the output, then Gemini is the better choice.
It really depends on your goal. I'd go to Gemini if I just want to learn a podcast content fast. But if I want to write an essay and in-depth research, NLM is the answer.
Awesome post, thanks! I love this different approach and how you highlight their ecosystem vs pure power angle. I'm intrigued by how you're using voice + screen share to really collaborate with Gemini. I really want to try that now!
Right now my main stack is ChatGPT and Manus, but I'll have to consider adding Gemini to this...also looking at Perplexity. But I'm impressed with the features and mindset of these Gemini updates. Thanks again for highlighting this!
Awesome!
Glad to know this was helpful for you. Yeah, I guess the only way to learn is try bunch of things until you find which suits best with your style. I love Gemini because it's integrated with their ecosystem, multimodals and long context window.
Dont forget to have fun trying all these :)
Great post, Wyndo, thanks! Gemini 2.5 Pro is my favorite coding model now 🙂
my fav too and using it with cursor! incredibly fast!
Thank you Wyndo, I appreciate your time and effort. I need to check the conversation branching you mentioned here. I also agree, we shouldn't eliminate any chatbots or tools, but learn to integrate them all to the best possible way to get our projects done. I also still use ChatGPT, and sometimes Grok for my work, but as you mentioned here, Google AI Studio has really outdone itself with its new approach.
appreciated for DM-ing me and make me also realized i'm not alone in this despite every eyes are on ChatGPT :)
That’s a ton of awesome experience and insights you’re sharing, thank you!
I really appreciate the all-in-one ecosystem Google is offering. It’s kind of wild to realize how seamlessly they’ve been able to support this, largely because of the foundation they’ve already built over the years. Honestly, only Google could pull something like this off at this scale.
And that project branching strategy is exactly what I’ve always wished Cursor would do. Sure, it technically can with Git, but having it in a user-friendly interface like that makes a world of difference.
Now I have to explore Gemini a whole lot more!
Thanks Jenny! :)
I still wish they could have done this sooner haha took a long time but finally took a major leap.
I think it’s hard for Cursor to pursue that feature given we could do it on git, but I agree that would be helpful for sure!
Yeah, that’s true.
Also, LOVE your personal website! It’s so well made!
Dang, I didn’t know Gemini had all these features!
I hope we can get local models doing all this stuff soon…
Gemini is my new fav :)
I really hope so. Gonna be more impactful to the workflow!