Hacker News
new
past
comments
ask
show
jobs
points
by
js4ever
8 hours ago
|
comments
by
twobitshifter
8 hours ago
|
[-]
Right, just give the text llm access to a vision specific agent and that problem can be solved. Or if you really want let it even call Opus with an image - seems like you’d still save money
reply