Still, I get the feeling that the apparent bit rate of the conscious output is not the right metric to be measuring. There are so many ways in which the bit rate you're measuring could have been bottle-necked by factors outside of the mind, such as the speed at which you can move your tongue to articulate language, or a cultural expectation that prefers compact answers over chain-of-thought babbles. The mind also leaks a lot of information, both consciously and unconsciously, through side channels such as facial expression, gestures, and even variations in the rate of output itself.
My theory is that consciousness is super complicated and brain has barely enough juice to crank out any of it.
Conscious field of vision is about 2% of total field of vision and we observe the world by constantly swiping it with this 2%. This way we reuse both neural circuitry of our brains and also training data that would otherwise necessarily be spread across larger size of neural network if our conscious field of vision was larger.
So it short, conscious output is so small because we are dumb and short lived.
I also wouldn't say that there's a lot of information in the side channels. Properly compressed it's probably less than engaged conscious output.