Which generative AI tool have you used the most and how?

I recently got access to Runway model 4.0 for video generation and all I can say is holy moly it is good. Type in a prompt and it can get you beautiful almost realistic clips in few minutes. Much better than anything I've used. Though kling is pretty good too and cheaper.
 
Chat gpt and Gemini
Chat gpt for creating notes, creating notes by detecting text from images, doing all the documentation for the institution. We are saving a lot of time by using AI for these documentation purposes without any errors. Posters and notices are created within minutes. Even created syllabus for some add on courses and we just need to fine tune a little.
Gemini is mainly used for finding information about what's on the screen instantaneously.

In my experience, the pdf,word files etc created directly from chat gpt are trash. We need to manually create the files by copy pasting.also some times gpt skips some part of notes without any reason and can go unnoticed easily.
 
Last edited:
Open-WebUI with Gemini Flash 2.0 has been a game-changer for me. Haven't felt the need to fully subscribe to anything yet, but I'm keeping a close eye on Open WebUI development. They have been adding a lot of good features recently.
 
  • Like
Reactions: vishalrao
Ive mostly used chatgpt for the general and coding related queries. Ive also used github copilot combined with sonnet 3.5 for coding assistance. Recently Gemini 2.5 Pro is making a lot of buzz, will switch to it once copilot officially supports it especially for its context tokens size. For now sonnet 3.5 and 3.7 for the win.
 
Renegade said:
Is copilot better than chatgpt? How are you using it and which all languages does it excel in?
Click to expand...
Copilot is a coding assistance whereas ChatGPT is a LLM. With Chatgpt, you can use for asking any generic query such as weather info, grammar check, etc including coding questions. Copilot is purely for coding related queries. It provides an interface to use any of the supported LLM under the hood to provide coding assistance via plugins in an IDE. Copilot can use ChatGPT models as well.

Coming to the question. When Chapgpt is used within copilot, its context tokens are limited by Microsoft behind the scenes to reduce costs so that you can use it for multiple queries without hitting the rate limit. Now this is not the case when using ChatGPT directly from the website or using it through API key. Though this will consume a lot of tokens (higher cost) but will give you the full potential of the LLM hence you will get better results.
 
Kloud said:
Do you mean plain OCR? or something complicated?

You should give Gemini a try. I think the free tier is much more generous than ChatGPT.
Click to expand...
Just want to extract texts from a bunch of continuous photos of text books and need to manipulate the extracted text to get summarised note, ppt etc
 
  • Like
Reactions: Husky
MrdK said:
Just want to extract texts from a bunch of continuous photos of text books and need to manipulate the extracted text to get summarised note, ppt etc
Click to expand...
AI is a poor tool to extract text. Use a high quality OCR. You can use AI to summarize the extracted text.
 
  • Like
Reactions: MrdK
dopplenoob said:
I recently got access to Runway model 4.0 for video generation and all I can say is holy moly it is good. Type in a prompt and it can get you beautiful almost realistic clips in few minutes. Much better than anything I've used. Though kling is pretty good too and cheaper.
Click to expand...
can this be used with lm studio?
 
Top Bottom