3 ms·
This sounds like something you could use LLMs for as LLMs are quite good at removing bias in content. You might run into context length issues feeding the whole
by michaelkrem 3y ago
This sounds like something you could use LLMs for as LLMs are quite good at removing bias in content. You might run into context length issues feeding the whole transcript in though .
- langsoul-com 3y agoI think that would be insanely expensive. Perhaps could be possible for a few channels though. Unless there was something like sampling, similar to how shazam recognising music from a few seconds works. Could someone only get a portion of the content and, semi accurately generate a summary?