Brian Douglas
2 HN recommenders · recommended in 3 months, 2020–2024
15:28 · 65K views · Nov 16, 2015
Brian Douglas introduces Braitenberg vehicles, simple robots whose sensors are wired directly to motors yet show surprisingly lifelike behavior. He draws on Valentino Braitenberg's book Vehicles.
Recommended as a short, fascinating introduction to how simple sensor-actuator machines can appear intelligent.
3 comments from Hacker News
I am forever indebted to this book for pointing me to Valentino Braitenberg's Vehicles: Experiments in Synthetic Psychology ( [LINK] ) where very simple autonomous sensor/actuator machines combine to exhibit complex behaviour which maybe mistaken for "Intelligence". Makes you really think abut Complexity/Intelligence/Emergent Behaviour/etc. Wikipedia : [LINK] Some Videos: [LINK] this video
Uphill Analysis, Downhill Synthesis is a major point in the book Vehicles . [0] Fascinating thought experiments. Here’s an introduction on YT [1], though it really only scratches the surface. [0] [LINK] [1] this video
At the time I was actually pretty skeptical about how hard it would be to get the AI to work plausibly (and so was Will, but somehow he managed to do it), and I though an online game like The Sims Online would be easier. But leveraging the player's imagination and supporting storytelling was key. Will mentioned a "Braitenberg Machine": [LINK] >A Braitenberg vehicle is a concept conceived in a thought experiment by the Italian-Austrian cyberneticist Valentino Braitenberg. The book models the animal world in a minimalistic and constructive way, from simple reactive behaviours (like phototaxis) through the simplest vehicles, to the formation of concepts, spatial behaviour, and generation of ideas. >For the simplest vehicles, the motion of the vehicle is directly controlled by some sensors (for example photo cells). Yet the resulting behaviour may appear complex or even intelligent. The Braitenberg Vehicles: this video Student: What about from person to person, you talking about the information that's contained within the objects, so there can be information in another person, that you want to interact with in the environment. Have you looked at any reasons why you would want to do that? Will Wright: Oh yeah, I mean, that's that's the hard problem. I mean, simulating ants is hard enough, when you get to people there's really no hope. There are two issues here. You can look at this as a technology. It's not a product right now. And their are a few directions this could go. I could se