Forget ripping off brains for AI. Butterflies and worms could lead us to self-repairing intelligent robots, says prof
During the whole metamorphosis process, the memories of the caterpillar remain intact and are transferred to the butterfly. “Things that the caterpillar has learned, the butterfly still remembers,” Levin explained.
Is there actually any proof of this at all, or is it just anthropomorphising something which is an innate instinct?
"...anthropomorphising something which is an innate instinct"
Well, what IS instinct? We tend to use it as a collective way of saying "This animal shows that behaviour and we have no real idea of how that happens, except that it seems to be inbuilt". And if it's inbuilt behaviour that means there is a mostly genetic component to it.
The twist here is that while 30+ years ago it used to be thought that genetics were mostly fixed, now we know that certain memories / experiences can be genetically encoded and even passed on to offspring. So it's not unreasonable to surmise that a some memory / experience is coded as genetic change, which can survive the metamorphic process.
Of course all that is highly speculative on my part and I am in no way an expert on any of this stuff... and of course, yes, they could be anthropomorphising
"Is there actually any proof of this at all, or is it just anthropomorphising something which is an innate instinct?"
Butterflies merely sip at nectar etc to feed themselves and this is clearly a learned behaviour since in their earlier "very hungry caterpillar" stage they gained memories of how eating one piece of chocolate cake, one ice cream cone, one pickle, one slice of Swiss cheese, one slice of salami, one lollipop, one piece of cherry pie, one sausage, one cupcake, and one slice of watermelon will give it stomach ache.
"This animal shows that behaviour and we have no real idea of how that happens, except that it seems to be inbuilt".
Thank you. After posting my comment I had that exact thought. It's very easy to say something is an instinct but actually it mostly, as you say, means we don't know where it comes from.
And that begs the wider question of how, and why, organisms manage to expand, modify and store behaviour over generations, or through metamorphoses as in this case.
I would say though, that to call this "intelligence" is a bit of a stretch.
"Is there actually any proof of this..."
I don't think there's even any evidence for it: the two body-forms, lifestyles and primary functions are so different that I can't think of anything that a caterpillar might learn that might be relevant to a butterfly, nor how it might be expressed.
"...now we know that certain memories / experiences can be genetically encoded..."
I think you have to be careful what you mean here: a learned experience can't be genetically encoded but a genetic change can result in changed behavior, which can then be inherited.
If the genetic change is beneficial and confers an advantage it can be mistaken for learning and thus it can appear as though an organism has learned something that has then been passed on to its descendants.
Is there actually any proof of this...
"I don't think there's even any evidence for it: the two body-forms, lifestyles and primary functions are so different that I can't think of anything that a caterpillar might learn that might be relevant to a butterfly, nor how it might be expressed."
Indeed. And perhaps the fact some of the "learned" fear of predators wouldn't be helpful - Imagine the inner conversation of a butterfly as it prepares for its first flight, wrestling with the urge to fly out into the open world to procreate, and its equally compelling memory of hiding under a leaf. No, natures goals are, most times, best served with less thinking and more doing.
I think most creatures, like us, are as much driven by chemistry, as they are cognitively. And if that is true, then a small brain needs only concentrate on its fundamental task of mixing up cocktails of drugs to stimulate urges, to achieve basic goals. The butterfly doesn't think, it just reacts to stimuli. It is chemistry that is passed down the generations, not synaptic pathways. After all, a foetus hasn't got a telepathic connection to the mothers brain, but it does swim in the very same chemicals, and it is there that new recipes, outside of the shared DNA can be passed on.
I would say though, that to call this "intelligence" is a bit of a stretch.
which leads us back to "what is intelligence " ............
trying to complete AI from a human brain start was highly unlikely to be attained without some form of a EUREKA moment, and it would appear that the natural world MAY have answers that we need to explore first, it may lead us to true AI, it may not, but we won't know until we try .............
OR our metallic overlords attain sentinence and start to kick off about all the shit weve done to them over the years, which is the OTHER way to know we have achieved true AI ;oP
I would argue that a learned experience can be genetically encoded and be passed on from generation to generation. This would be a great explanation for a root cause of irrational fears, such as arachnophobia and claustrophobia
I can quite easily imagine one of our ancestors being trapped in an encosed space after an earth quake and for that experience to be genetically encoded and manifest itself in the form of an irrational fear in later generations
DNA is very adaptable, and it does this through feedback of the environment
To disagree with this, is to ignore your own nature
This is what I've said plenty of times, actually. I think there was a paper, but it basically described how a series of relatively simple learned behaviors could be put together to execute complex tasks.
Possibly outing myself if someone recalls it, but I employed the same principle when coding a maze-runner in College Robotics. Move forward, watch for openings. Measure the distance of any detected opening to determine if it is likely a valid turn. If a valid turn is detected, make it by slowing the inside wheel in order to minimize speed loss. And of course, a straightening routine consisting of a jag away and then slightly less back the other way if a wall is bumped. That's all it took to execute a maze run smooth and fast enough to post a time that feels like it may still be the fastest after... well, more than a decade.
Biological robots already exist if you consider the genetic work done with crops and bacteria.
Algorithms must be able to bridge the physical as much as the logical in order to execute. Organisms ultimately end up crowd sourcing outcomes among all their brethren cells birthed from the same DNA instructions.
The body doesn't work from one member and neither should the next tier of AI
Brains are overrated
Human intelligence isn't confined to the brain. The nervous system extends all through the body, the brain constantly solicits - and gets - feedback from the legs, back, guts, loins and every other part. "Thinking with your gut" is a real thing, there are more nerve endings there than in the brain of an adult cat.
This is why sentient brains in jars is still science fiction, and likely to remain so. Also why uploading "yourself" is such a stupid idea. The brain and body can't be separated, at least not without effectively destroying whatever owned them.
All forms of life
There are forms of life on this planet that have survived for billions of years
Humans have only been here for a short while, and some might argue not for much longer
It's also been posited that the reason we dont see 'advanced' intelligent life out in the universe is, that they are likely to have ended up destroying themselves.
Intelligence comes in many forms, and our inability to recognise this shows that we are unintelligent
Ai will be nothing like organic intelligence, it will be a super intelligence, able to make use of a scary array of sensors that our soft organic tissue cant begin to compete, super high frame rate sensors able to percieve in all kinds of wavelengths, infrared, even X-ray, processing speeds which makes moving objects look like they are going slowmo, can come in all shapes and sizes, be highly specialised, or general purpose.
You just cant compare Ai with organic intelligence, its not even the same thing
We need a reference for the flatworm claim. I've taught the flatworm story for years. The 'memory' is just stress hormones (generated when the 'donor' flatworm has to learn a stressful task) which affect the rate at which regenerated (or cannibal) flatworms learn a task. Please take a look at Frank, Stein & Rosen (1970) 169, 339-402. It is a fabulous debunking of 'chemical memory'.