Hacker News new | ask | show | jobs
by what-the-grump 355 days ago
I cannot reproduce this in any version of copilot?

Copilot with outlook.com

Copilot base one that comes with M365,

And the add-on one for 30$/mo.

Copilot in VS code

All produce: ffmpeg -i movie.mov -vcodec libx264 -crf 23 -preset medium -acodec aac -b:a 128k output.mp4

Which is not surprising because its just an Open AI 4o call... so how are you getting this?

3 comments

> I cannot reproduce this in any version of copilot?

Because LLM-based service outputs are fundamentally not-reproduceable. We have no insight into any of the model settings, the context, what model is being run, etc.

The fact they have so many different version is hilarious to me; whern they're in the context of getting every day entreprise customers.

Already gemini giving me flash or pro like i'm supposed to know and decide which i need is missing the point, but 4 of them is crazy.

Wait whats so confusing?

GPT3.5, 4.0, 4o, 4o-mini, 4.1 o3, 4,5?

I want to slap every person at OpenAI.

call freaking tuned models what they are, basic, advanced, coding, research, etc. etc. Slow, Fast. Like cmon.

Copilot is not OpenAI
The pipeline Microsoft is using for Copilot products hides actually what models they are using, and you have no influence over it. Sometimes they use smaller models, but I have no clear source from Microsoft saying this ...
It is OpenAI weights under the hood, unless something changed recently?
At least Bing Chat was GPT-4-base with Microsoft's own fine-tuning.
And there's been a number of iterations of 4o. So it could be a really old, pruned one if upgraded from gpt-4.
Copilot is OpenAI in branding since day 1.

https://en.wikipedia.org/wiki/GitHub_Copilot#Implementation