The Computer Between the Ears

[boulderjohn]

boulderjohn: I find the concept of "it is possible for a computer to make an error ... so you can never trust one" interesting in that we generally accept e.g. that people lie, but we're still willing to live with them. I.e. we are normally operate in an imperfect world guided by at best incomplete information -- yet the computer must be perfect or else it's useless.

I believe the psychological effect appears in other areas as well.

• You mention lying, and I have seen interesting references to that in the current political scene. (Note--please don't take this down a road that gets this moved to the pub!) I saw someone argue that because a certain well established news reporter published a story decades ago which was not well researched and may or may not have been true, that person is a liar who can never be trusted, even though there is no other known transgression in his past, but a politician who makes outright lies regularly can be trusted on everything because some of it might be true.
• The fallacy known as the appeal to ignorance contains that same contrast in thought. It has two primary forms, and individuals who use this fallacy will frequently use both forms on the same argument on the same topic.
  • You cannot prove your point to a 100% level, so you are wrong.
  • You cannot disprove my point with 100% certainty, so I am right.

What they all have in common, IMO, is a grasping at any argument that will allow you to believe what you most want to believe, regardless of the weight of contrary evidence.

 

[PfcAJ]

Frankly, and I know I will get flamed, I would never dive without one. Actually, two. It helps keep me safe, allows me to use multiple gas mixes during a dive and not have to think about it. Does that mean I am totally reliant on them, no. In my pocket is a set of tables I have run that should get me out of the water safely in the unlikely event that both computers die, which I have built using computer software running the same algorithm as my computers. So, really just another computer.

Sounds like a great plan

 

[NatureLVR]

Excellent article!

 

[dmaziuk]

I saw someone argue that because a certain well established news reporter published a story decades ago which was not well researched and may or may not have been true, that person is a liar who can never be trusted, even though there is no other known transgression in his past, but a politician who makes outright lies regularly can be trusted on everything because some of it might be true.

Technically, the statement "the journalist can always be trusted" can be trivially proven false by just one counterexample. So that one's easy. The case of a politician whose lips are moving is a bit trickier: I suspect there's liar paradox in there somewhere.

 

[Blasto]

The case of a politician whose lips are moving is a bit trickier: I suspect there's liar paradox in there somewhere.

Normalization of deviance plus power dynamics. The public lets a popular politician get away with a lie once - they've done a lot of good; repeat until it's normal and accepted.
Journalists and doctors have been held up to some codes of conduct, and it's much easier to change them than to change who's in power.

Back to the subject, it's pretty hard to change your brain too. So people exempt the errors they make from the judgement applied to everyone/everything else.

 

[wetb4igetinthewater]

Excellent article. I’d like to add that the human brain is excellent at things like object recognition, speech recognition, etc.. Computers have always been superior in raw calculations. The probability of error by a computer is so much lower than the brain. I cannot understand the Luddite arrogance of thinking that the brain can outperform a computer for computations. To say that, one must either not have any understanding of technology or have drunk massive volumes of Kool Aid.

I’d rather have both my Shearwaters and a various plans with slight deviances from the dive plan in my wetnotes.

Any deviations from computers are covered in debrief.

 

[barth]

Terrific article, John... I totally agree.

The human brain is an incredible computer, but it's pretty silly that we compare it to a dive computer and argue about which is "better". They are completely different mechanisms that excel in some tasks, and fail at others.

There is no way that my brain (or the brains of the best divers in the world) can accurately track a complex dive profile curve over time and depth and continuously generate an ascent schedule with anywhere near the accuracy of the cheapest Mares Puck on eBay. On the other hand, I wouldn't expect my Petrel to write a sonnet, or even one of these posts.

They are different tools for different tasks, and I think that a dive computer is the best tool for managing decompression on the fly. Sure, I back that up with a written plan for tech diving, but that's with the understanding that if I have to do that, I'm probably going to be in the water a lot longer than necessary.

How accurate should a dive planning be?

I think it is hard to tell how your body Will react If you compare it to a theoretical model. How fit is your body ? Did you use electric heating during the dive an when did you use it? Did you drink enough? How accurate was the gas analyse? How accurate is the sensor which is measuring depth ? Etc

You can do a 10 h dive, you computer will tell you exactly what you have to do. But should it be that accurate? And how accurate is it really, will the diver get bent? Or could he exit the water earlier ?

For me there were 2 reasons to stop using my gekko fo REC diving.

- when I was at my car, I did set the percentage of 32 after analyzing. It was a long walk. When I was in the water the percentage was reset to 21 by the computer. It was not possible to change it anymore.

- The safety stop was at 5 meter. The computer accepted everything between 4,0 and 6,0 meter. I did my stop at 6,5 meter because there was no current. Next stop I did at 3 meter there was also no current. This was the diveplan. The result was computer in error mode.

Because my gekko didn’t have a gauge mode I sold it. And bought a uwatec TEC 2g which I did use in gauge mode for ten years.

Maybe When dives are more complicated I want to use a computer. But I have to make a gas and diveplanning. For now my tables, Iphone and gauge mode is enough.

 

[Tassi Devil Diver]

For the 12 years I have been doing technical divers I have always primarily relied on using a computer with tables as back up, and for the past 6 years have used two computers and don't bother with taking a written tables. I review my dive plan using a desktop planner and then use my brain to make decisions about the length of bottom time/ time to surface based on gas consumption, how much bailout I am willing to carry, sorb capacity, and will change my plan on the fly as planned depth and conditions vary and follow the deco schedule spat out by the computer. Personally I think those who rely solely on ratio deco are foolish.

 

[dmaziuk]

Excellent article. I’d like to add that the human brain is excellent at things like object recognition, speech recognition, etc..

Pattern recognition.

It is quite possible to train your subconscious to recognize ascent patterns for various dive profiles, you just have to dive a lot. Then you could express them as, say, ratios. But while you're training it, ascents have to be calculated by some other computer.

And then you've trained your brain to follow whichever model that dive planner was using. So if you used the one that generates deeper stops, your brain now wants to make deeper stops. If that is no longer advisable -- re-training the brain takes forever and the older we get, the harder it gets.

There might be a human out there capable of keeping track of 16 tissue compartments simultaneously in his head, but I'm sure most of us are not him. Brain is not running the same model and all those comparisons to computers are bit of apples to Pink Floyd.

 

[100days-a-year]

There might be a human out there capable of keeping track of 16 tissue compartments simultaneously in his head, but I'm sure most of us are not him. Brain is not running the same model and all those comparisons to computers are bit of apples to Pink Floyd.

Not necessary to track 16 theoretical compartments,in any given dive only a couple will be the ones that determine stop depth and length.My EDGE and several dive planning software give a graphical representation of the theoretical compartments.One can review past dives and see some patterns emerge easily.


RD could be modified to fit empirical data as well.

Apples to Pink Floyd is a brilliant analogy.Surprised no one has mentioned the processing power needed to monitor autonomous functions along with sensory data.Not to mention the background subconscious that does on.