Gemini: Difference between revisions
| Line 21: | Line 21: | ||
=== Conversations === | === Conversations === | ||
Deep Research<ref>[https://drive.google.com/drive/folders/1YmAOOlLxmlnHXcpPt9ZQujqjM8TDRfS8 Deep Research Folder]</ref> generates a lot of information. Pages tend to be around 20 pages. I haven't generated many of these recently, but I think I'm going to start doing so again. I need to have the mindset to use these prompts before diving into my more generous ones. | |||
These are my saved conversations with the Thinking 3.0 Model<ref>[https://drive.google.com/drive/folders/1MSKRD-cyCql5O5SrqMjj64q7tqgcEzb1?usp=sharing Gemini Conversations]</ref>. These might turn into web pages on the front page or articles on my wiki, but take a look if you want a preview. | |||
=== Gems === | === Gems === | ||
Gems<ref>[https://drive.google.com/drive/folders/1xv7WLZTEmpVKRobdOjVb2DIL7eG9nWql?usp=drive_link Gemini Gems Responses Folder]</ref> are great when you have an idea for a repeated query. It allows me to keep a consistent response format, like if I want a deconstruction of an image organized in a certain way. It also allows me to have a set of instructions that I can reuse without having to re-define and set the stage, this includes. | |||
=== Images === | === Images === | ||
| Line 33: | Line 35: | ||
=== Veo === | === Veo === | ||
My Veo<ref>[https://drive.google.com/drive/folders/1JnUyB79x9OCKAawIn5tquwbmji_wDfD5?usp=drive_link Veo Folder]</ref> contains my select favorite in Veo generation. Veo is still very unpredictable, but what is generated can be funny or cute. Rarely do I get my vision when doing a Veo prompt, but I do tend to be intrigue by what is generated. My Veo generations with Megumi<ref>[https://drive.google.com/drive/folders/1710ypPP4SznJXjOU3qVxKQwjqXwKv9_j?usp=drive_link Megumi (Veo) Folder]</ref> is the most recent. So, it's done with the Thinking 3.0 Model. | My Veo<ref>[https://drive.google.com/drive/folders/1JnUyB79x9OCKAawIn5tquwbmji_wDfD5?usp=drive_link Veo Folder]</ref> contains my select favorite in Veo generation. Veo is still very unpredictable, but what is generated can be funny or cute. Rarely do I get my vision when doing a Veo prompt, but I do tend to be intrigue by what is generated. My Veo generations with Megumi<ref>[https://drive.google.com/drive/folders/1710ypPP4SznJXjOU3qVxKQwjqXwKv9_j?usp=drive_link Megumi (Veo) Folder]</ref> is the most recent. So, it's done with the Thinking 3.0 Model. | ||
== References == | == References == | ||
<references /> | <references /> | ||
Revision as of 03:34, 26 November 2025
Sharing
With each of your saved conversation, you can click on More Options, and select "Share" to get a link that you can send the conversation to. These links are available for as long as the conversation exists, which makes it not a great long term solution if you're planning on having a lot of saved conversations. To manage that, I recommend saving the conversation to Google Drive as a Google Doc. That option does appear to be a desktop only feature as the mobile option only offers validation and a copy to clipboard where save to Google Drive would be.
Clean up
With each of your saved conversation, you can click on More Options, and select "Delete" to remove the conversation. I found that it is more efficient to delete from a mobile device. There, you can long press and click "Delete".
Imagen
Imagen is Gemini's image generation feature. You can switch back and forth between the fast model and thinking model depending on what your goals on. Since, the thinking model has a limited number of requests, it's good to get an idea of when you can get away with fast.
Subscription
On the best free model of Gemini, you get 5 requests on the "Thinking" model (3.0), 5 Deep Research requests per month, 0 Veo requests, and you're limited to 50 sources in NotebookLM. If you want more, you can subscribe to Gemini Advanced for 100 requests on the "Thinking" model per day, 20 Deep Research requests per day, 3 Veo requests per day, and you're NotebookLM limit increases to 300 sources.
Saved Generations
For a bit I had this on my front page, but now I just want to tuck it back here. Even though, I enjoy what I generate, I want to try to focus on having content that I have at least had my hands on more than just providing a prompt.
Conversations
Deep Research[1] generates a lot of information. Pages tend to be around 20 pages. I haven't generated many of these recently, but I think I'm going to start doing so again. I need to have the mindset to use these prompts before diving into my more generous ones.
These are my saved conversations with the Thinking 3.0 Model[2]. These might turn into web pages on the front page or articles on my wiki, but take a look if you want a preview.
Gems
Gems[3] are great when you have an idea for a repeated query. It allows me to keep a consistent response format, like if I want a deconstruction of an image organized in a certain way. It also allows me to have a set of instructions that I can reuse without having to re-define and set the stage, this includes.
Images
Megumi[4] is a character of mine that I generated around the time that Nano Banana was advertised with the Flash 2.5 model. It allowed for very reliable AI augmentation to images. Now longer was I asking Gemini to generate prompts for two different images and then asking Gemini to splice the two prompts together into one. I just had to start with an image and ask it to wear something else or be in another pose. There's still so much to explore and process, so I haven't gotten back to my older characters yet. If you want to dive deeper, my outfits[5] folder contains her in different outfits.
Veo
My Veo[6] contains my select favorite in Veo generation. Veo is still very unpredictable, but what is generated can be funny or cute. Rarely do I get my vision when doing a Veo prompt, but I do tend to be intrigue by what is generated. My Veo generations with Megumi[7] is the most recent. So, it's done with the Thinking 3.0 Model.