Hacker News Reader

Comment by dantillberg | Hacker News Reader

> deciding that this is the most relevant activity to keep doing at the moment, instead of doing math or going out for a walk.

How many bits of actual decision is going on here, as compared to the period of time that decision applies to?

For example, if a person decided once per second whether or not to go for a walk, that could be 1 bit per second. But if that person is constantly transitioning back and forth between walking and not-walking, we could consider their behavior pathological. Instead, for most people, the information density of these decisions is quite low, i.e. the per-second decision bits are very compressible.

Personally, I only decide whether to go for a walk (or not) _at most_ once every few minutes. Even if we add in bits for "where" and "precisely when" and "how long", I think we're still at just a small fraction of 1 bit per second.

View on HN

Your conscious planner may be making 1 decision/second, but your senses, proprioception, balancing system, etc. are handing you about a gigabit/second of data, most of which never rises to your conscious attention.

When I'm reading, that's roughly 2000 bits/second, but I am engaging it with model-making systems that can completely enrapture me.

I/O is not the same as computation; conscious computation is not the same as all computation.

Yeah, that's what the paper is talking about. A whole lot of input vs. a small amount of conscious focus.

There is a whole "OS" in the background that is way more complex than all of our programmed systems. I may be hyped focusing on a complex program and really not think about anything else, but my body is still constantly senses, processes, and reacts to signals. E.g. I'm not ragdoll falling to the ground, which requires holding balance and tone, which is far from trivial. I also constantly listening to every possible danger, which would immediately override my current activity. I also react to internal signals like hunger or pain.

A bit dumb, but maybe relevant comparison might be asking why can an Apple Watch stay on for a single day only on a charge, while Garmin can do 2 weeks/a month? Because one is a general purpose computer, while the other is an embedded software that can only ever do that few things it is preprogrammed to do.

I don't see your comparison as relevant. I'm a metre away from a nine-year-old Android 6 phone that can easily go a month on a single charge (while tracking its location!). Your Apple Watch is (a) constantly using its radio, (b) running a fairly high-end graphics chip, and (c) actively probing your body at intervals. Nothing to do with software.

I think you lost the actual comparison. Is not about Android and iOS.

Is about having a ton of mini specific "computers" vs. the very generic conscience.

For example all the balancing, background hearing, hungry, thirsty , and so on are very specific zones in the brain. While our conscient mind is doing...well, whatever we like in a very generic way, basically keeping the joy of living on.

I mentioned two watches with similarly sized batteries.

The phone battery is only about twice the volume of a smartwatch battery.

"Personally, I only decide whether to go for a walk (or not) _at most_ once every few minutes. "

Consciously. Subconsciously much more is going on.

  Personally, I only decide whether to go for a walk (or not) _at most_ once every few minutes.

Just because usually the decision is 'keep the current course' that doesn't mean no decision has been made.

The vast majority of the brain's processing power, including conceptual processing, is not conscious. Conscious thought is the framebuffer of the mind, subconscious is the RAM.

> Personally, I only decide whether to go for a walk (or not) _at most_ once every few minutes.

And yet, if I walked into your office and shouted your name, you would have less than a second of reaction time, indicating that you are processing all sound around you and deciding whether to get up and talk to someone pretty much continuously.

Reply on news.ycombinator.com