Artificial Intelligence Tools

secretagentdad said:
It blatantly refused to help me recode a one page website I gave it the code for.

I wonder what benchmark that's optimized for.
Click to expand...
What did you do to it lol? I've found it able to understand a SaaS I'm working on for a client that Claude REALLY couldn't even help with at all...
 
I think I just gave it to much stuff and guidelines to not try to break it into microchunks.

It appears to have new ways of being assertive with this version.

On resubmission's its working fine just forcing small chunks and lots of please continue.
 
The general two agreements about GPT5 seems to be:

1. It's fine, people just had parasocial relationships with 4o

and

2. It's a step back and a cost saving exercise
 
chubes said:
Wan - Open Source AI Video Generator

https://github.com/Wan-Video/Wan2.2
Click to expand...
Has anyone tried an open-source, local video (not image) generator? How fast do they tend to run on something akin to a Mac Studio or even a Mac Mini? Heck they may not even work for Mac, I didn't check. Just wondering if paying for access to another company's GPU's is worth it, especially since so many renders "fail" in that they aren't what you need. Thanks.
 
Ryuzaki said:
Has anyone tried an open-source, local video (not image) generator? How fast do they tend to run on something akin to a Mac Studio or even a Mac Mini? Heck they may not even work for Mac, I didn't check. Just wondering if paying for access to another company's GPU's is worth it, especially since so many renders "fail" in that they aren't what you need. Thanks.
Click to expand...
I understand the wanting to keep it local/open-source to pay less for fails.

However, as a heavy user of Ai for video ads, I can tell you the paid models are far more advanced.

You will spend a hella lot more time and wasted stress on regenerating fails on the open source, then any cost you will have on the paid model fails. I maybe spend $100 on credits and 2 hours on pro quality videos on Veo 3. Everyday Im getting that slimmed down and faster

I would tell you to learn up and really study/practice on Google Veo3 and their studio to keep frame by frame references tight and the same. And then look into Ai voice systems ( outside of Google Veo ) for the audio portion instead of open source models atm
 
You must log in or register to reply here.
Back