Hacker News Reader

Comment by CyberRabbi | Hacker News Reader

Hmm doesn’t seem credible. Everyone knows that brain size has no effect on intelligence.

Yea, I was talking to a fruit fly yesterday...

Wait no, that’s not accurate because while larger brains don’t inherently equate to more intelligence, there is a minimum size for any given level of intelligence. That said, this wasn’t a test of intelligence, just how cells grow in response to specific genetic alterations.

But larger brains in monkeys do equate to larger social groups (dunbar's number. monkeyspheres) and could be a necessary precursor to civilization.

It’s important to separate changes across evolutionary timescales with individual mutations. Brian’s have a metabolic cost which promotes efficiency over time, but efficiency is just one of many relevant factors. A whale has significantly less relative metabolic cost per Lb of brain than a mouse. Further, individual mutations may be a net negative.

I didn’t realize that was scientific consensus. So that would imply that brains under a certain size are incapable of higher levels of intelligence? I don’t think that’s right... I thought the consensus was that brain size has no strong correlation with intelligence at all. That makes sense considering how neurons work, the connections they form seem like they would be more causal.

Its a complicated question. Within the human species, the correlation is not that strong at all. In cross-species comparisons it tends to be quite strong, especially when considering brain to body weight ratios. Bird brains may be an interesting exception however, which I suspect can be quite intelligent, depending on how intelligence is measured.

Part of what makes that interesting is we are very good at some tasks that might be quite expensive, and really bad at some tasks that could be simple. In other words, our personal views on difficulty could be wildly off. For example assuming computer vision should be easy and chess hard.

The scientific consensus is that from the evolutionary perspective, brain/body size is absolutely correlated with performance. This doesn't have to be general intelligence, but can be sensory processing ect.

That is not to say that structure isn't important too. This is especially relevant because the study observed some of the structural changes through to be associated with increased intelligence as well.

Information processing is limited to the capacity of the network. So it is easy to see that very small brains cannot be very smart (there is no set of connections that can make fruit fly human-intelligent). However, having a large capacity network doesn’t mean it will result in high intelligence. So it’s necessary but not sufficient.

You'll never play Doom on a single NAND gate. Huge efforts in getting toasters, microwaves and pregnancy testers to run Doom notwithstanding.

But brains aren’t made of digital logic elements

You'll never play Doom on a single neuron.

But you surely can on the computer simulating that single neuron :)

You’ll never play doom on any amount of neurons... brains are not cpus

Why not? You could manually process the assembler on pen/paper. Sure the FPS might be a bit on the low side.

I’m pretty sure there’s sufficiently complex cellular machinery inside a neuron to run Doom if arranged in the right way.

https://en.wikipedia.org/wiki/Brain-to-body_mass_ratio

With an extra caveat:

> Recent research indicates that, in non-human primates, whole brain size is a better measure of cognitive abilities than brain-to-body mass ratio. The total weight of the species is greater than the predicted sample only if the frontal lobe is adjusted for spatial relation.[18] The brain-to-body mass ratio was however found to be an excellent predictor of variation in problem solving abilities among carnivoran mammals.[19]

Another win for the heuristic "if someone says 'everyone knows X' it probably means X is false"

Reply on news.ycombinator.com