Would a 9 lb sapient spider producing 1/10 lb of silk a day suspend your disbelief?

[Anthony]

Originally Posted by sir_pudding

You don't have to invoke supernatural forces merely biochemical processes.

There aren't any known physical processes that can't be emulated with a Turing machine plus a noise source, though some of them require impractically large quantities of computer power.

[Alden Loveshade]

Originally Posted by Ze'Manel Cunha

1 Terabyte and 100 Megaflops is the estimate for an insect, if I recall, the standard basic estimate of the functioning of the human brain are usually conjectured at around 1 Exabyte and 10 Petaflops.

If I recall (and I admit this is only from memory), GURPS Ultra Tech in 3rd edition estimated it as 100 Gigabytes for the human mind, which was upped to 1 Terabyte in 4th edition. But again, that's from memory, and I could be wrong.

Originally Posted by sir_pudding

...The computational model of intelligence doesn't really seem to work. Neuroscience currently holds more mysteries than answers....

I agree. And I don't think we're going to solve any real world problem in neuroscience in this discussion. But I didn't believe this discussion was created to come up with a working model of a sapient arachnidoid; more as a "if this were possible how might it work" discussion.

Recent research suggests that even crows, which have very small brains, may learn to identify faces by being taught this from their parents. Some crows have shown tool-making abilities. A couple or so decades ago, tool-making was considered a measure of what in GURPs terms would be called "sapience." The relative size of the brain relative to the body is important in addition to absolute brain size, as is the percentage of "groves and crannies." But that's if we're assuming a brain that works like a mammalian brain does. It might work quite differently for an arachnid, or a bird.

"The first thing that Dr Kramer came up with was that the penguin has a much smaller brain than the man....If we increase the size of the penguin until it is the same height as the man and then compare the relative brain size, we now find that the penguin's brain is still smaller. But, and this is the point, it is larger than it was." -- Monty Python Scientist

[teviet]

Originally Posted by Ze'Manel Cunha

1 Terabyte and 100 Megaflops is the estimate for an insect, if I recall, the standard basic estimate of the functioning of the human brain are usually conjectured at around 1 Exabyte and 10 Petaflops.

Do you have a reference for this? What I recall is that the brain has maybe 1e11 neurons and 1e14 synapses; each synapse would require a few bytes to encode its connectivity, action potential, and timing; maybe a similar amount of information again to define more coarse-grained properties (chemical environment etc.) All in all I get about around a petabyte, not an exabyte. But I am genuinely curious so if you do have a reference I'd like to look it up.

Of course this is just information content in the sense of "what it would take to store a braintape", not processing power; I haven't given much thought to this, and it's where the particular computational model starts to get important.

As for the spiders, well, if they're engineered they could theoretically be intelligent with much smaller brains. (In principle you could get up to 12 orders of magnitude by encoding information at the molecular rather than the cellular level.) But for naturally-evolved critters I would think that each order of magnitude reduction in brain mass will cost you at least a couple of points of IQ.

TeV

[sir_pudding]

Originally Posted by Anthony

There aren't any known physical processes that can't be emulated with a Turing machine plus a noise source, though some of them require impractically large quantities of computer power.

Yes, which was my point. I don't think it's likely that you model a brain with only a terabyte of storage space.

[Alden Loveshade]

Originally Posted by sir_pudding

This I think is a common trap. If you make claims of plausibility then you invite nitpicking and weaken suspension of disbelief. It is far better just to not mention it. If a player brings it up, just tell them that's the way it is.

Rule Number One: The GM is always right.

Rule Number Two: If the GM is wrong, see Rule Number One.

[sir_pudding]

Originally Posted by Alden_Loveshade

Rule Number One: The GM is always right.

Rule Number Two: If the GM is wrong, see Rule Number One.

It doesn't just apply to games, but to all speculative fiction. People who can accept magic spells should be able to accept anything. The problem is when you say, "This is my awesome realistic setting. It makes perfect sense in every way. I dare you to find something even remotely implausible in it!" Then they tear you apart.

[Anthony]

Originally Posted by sir_pudding

Yes, which was my point. I don't think it's likely that you model a brain with only a terabyte of storage space.

Oh, that's probably true; there are on the order of 100 trillion synapses in the brain, and it's unlikely you can model them on less than a byte each (actually, just marking each synapse with a unique ID for each associated neuron would be about 10 bytes).

[DanHoward]

Originally Posted by Flyndaran

Too bad Low Tech armor only gives a CF rather than breaking it down into materials and labor costs.

Because the number was pulled from thin air. Use the Wired article if you want more detail
"To produce this unique golden cloth, 70 people spent four years collecting golden orb spiders from telephone poles in Madagascar, while another dozen workers carefully extracted about 80 feet of silk filament from each of the arachnids. The resulting 11-foot by 4-foot textile is the only large piece of cloth made from natural spider silk existing in the world today."

If I'm reading it correctly, you have 82 people extracting silk for four years to make a single 11x4' piece of cloth that weighs 2.6 lbs. On top of that you need someone to weave the cloth.

Here is another tidbit
"Camboué built a small, hand-driven machine to extract silk from up to 24 spiders at once, without harming them."
The question is whether this machine could be constructed using Low-Tech (TL0-TL4) technology.

More
"Fourteen thousand spiders yields about an ounce of silk"
"Once the spiders had been milked, they were released into back into the wild, where Godley said it takes them about a week to regenerate their silk."

It is one of the best articles I've seen for pulling out information that modellers can use. You wouldn't believe the frustration when writing Low-Tech. A lot of the practical information we needed for trying to model various ancient technologies was incredibly hard to find.

[Flyndaran]

Originally Posted by Anthony

Oh, that's probably true; there are on the order of 100 trillion synapses in the brain, and it's unlikely you can model them on less than a byte each (actually, just marking each synapse with a unique ID for each associated neuron would be about 10 bytes).

This assumes that the human brain is orders of magnitude near its optimal efficiency. I call shenanigans on that. Evolution is a sloppy trial and error method that comes with numerous faults. It's just had billions of years to blindly increase complexity.
There have been people suffering damage that maintain normal or even high intelligence with only 10% functional brain tissue. Much of the important stuff is in the thin outer layer.
All in all, there is a huge amount that can be improved.

[Anthony]

Originally Posted by Flyndaran

This assumes that the human brain is orders of magnitude near its optimal efficiency.

It's probably subject to some degree of optimization, but evolution does a surprisingly good job at a lot of things.