But I’m not sure if it is fair to consider Bee intelligence to be a simple matter of number of neurons. Bees are great at being bees and doing bee things, not “general” intelligence necessarily. They are part of a ecosystem where they co-evolved. For example, they have a great sense of smell and that helps finding food sources with little computational power (compared to using vision, for example).
What I wonder is: what if we decided to create a robot that acted like a bee, we put all the sensors bees have, how much processing power it would be needed to “emulate” a bee on its environment?
You don’t actually need to build a robot. You can design a video game (a simulation if you will) and have bee characters.
Rzlufsays:
Over the years I have very similar thoughts. But it still looks like an uncharted question. I will be very grateful, if autor could continue this interesting topic of insect intelligence.
i am agree with you, a bee level ai could easily improve lots of our daily work while still wont made us afraid 😀
Pierre-Antoinesays:
I wholeheartedly agree and reached a similar conclusion, albeit with cockroaches rather than bees.
Contrary to what Solimao and others say, the bee’s intelligence is absolutely general. It is adapted to a different problem-set and a different sensory input than ours but I’m convinced it solves the same problem in the information theory field, if I may.
If we manage to create a simulation of a bee, we will have solved the hard AGI problem and it’s only a matter of processing power and sensory-tweaking before we can simulate a rat, a dog, a monkey, a smarter monkey (aka human) etc.
But I’m not sure if it is fair to consider Bee intelligence to be a simple matter of number of neurons. Bees are great at being bees and doing bee things, not “general” intelligence necessarily. They are part of a ecosystem where they co-evolved. For example, they have a great sense of smell and that helps finding food sources with little computational power (compared to using vision, for example).
What I wonder is: what if we decided to create a robot that acted like a bee, we put all the sensors bees have, how much processing power it would be needed to “emulate” a bee on its environment?
You don’t actually need to build a robot. You can design a video game (a simulation if you will) and have bee characters.
Over the years I have very similar thoughts. But it still looks like an uncharted question. I will be very grateful, if autor could continue this interesting topic of insect intelligence.
Bees may be a bit of a high bar 🙂
https://arstechnica.com/science/2017/02/bees-can-train-each-other-to-use-tools/
https://www.newscientist.com/article/2142884-bees-are-first-insects-shown-to-understand-the-concept-of-zero/
Would an animal with a simpler behavior work better as a start (nematods etc.)?
I don’t know if bees are too high a bar, but it is surely easier to emulate bees than human beings.
i am agree with you, a bee level ai could easily improve lots of our daily work while still wont made us afraid 😀
I wholeheartedly agree and reached a similar conclusion, albeit with cockroaches rather than bees.
Contrary to what Solimao and others say, the bee’s intelligence is absolutely general. It is adapted to a different problem-set and a different sensory input than ours but I’m convinced it solves the same problem in the information theory field, if I may.
If we manage to create a simulation of a bee, we will have solved the hard AGI problem and it’s only a matter of processing power and sensory-tweaking before we can simulate a rat, a dog, a monkey, a smarter monkey (aka human) etc.