The Debater 

efhartmann I hate it

@@mridontexist5489 a female android doing what u want... oh hell I love progress...

Reality doesn't follow the plot of movies

...What do you mean always does. This has never happened before you luddite. James Cameron is not god.

Forward Robotics specifically designed its plane-like drone design - called the U7AG - to be price competitive with larger pull-behind sprayer units (around $50,000 Cdn). - Photo: Forward Robotics Designs variable and numerous Of course, efforts are being made to eliminate this problem through improved designs. Meng Wei, chief executive officer for the aforementioned Forward Robotics Inc., says the U7AG incorporates larger droplets (medium and coarse) and a fixed-wing schematic to reduce drift issues. “Fixed wing aircraft have different air fluid patters,” he says. “[It’s] more like a traditional aircraft, thus fitting with traditional aircraft spray regulations.”

There are still several advantages. To make an accurate forecast you need 3 things : data, hardware (or processing power) and software (algorithm). The thing is, some random people are able to forecast way more accurately than the CIA even if they don't have access to CIA's accurate data and powerfull computers (see the book Super Forecasting by Tetlock and Gardner). It means the most important part is the algorithm, and it makes sense if you study epistemology. When you want to study a topic, you gather data and you have to tell what's probable and what isn't. There is no way to say X data is 100% or 0% accurate, the % is always between 100% and 0%. When you are asked a "big" question like "will Russia end the current war in Ukraine by Jan 1 2023", you must start by breaking down this question into smaller parts. If one of these smaller parts is still too big, you break it down too until you can't break anything down anymore. Then you start answering each small parts individually and give each answer a probability in % of it to be accurate. For example, the small part "Does Putin have cancer?" is Y% true. Then you answer another question, then another... Then you will have to decide the "weight" of each answer. For example, if you are pretty sure Putin takes a medical treatment that might interfere with his mental health, it has more weight in the final answer than knowing he likes cats. (the book probably explains it much better than me, and I don't even speak english very well, sorry) The issue with that is that even if you try to supress all your biases, some of them are hard-coded, deep into your early life experiences, or even in your genes. For example the Loss Aversion bias is also found in monkeys, which means this one is a survival instinct, certainly based on ressources scarcity at some point in evolution. It was certainly a good thing to have this bias as our ancestors who survived appear to have had an evolutionnary advantage at the time (maybe even today in some situations), but if you want to answer that Putin question, it will mess with the accuracy of the result. What if you could build a computer that avoids this bias (and all the known biases we know)? I don't know about IBM Debater, but I know it's possible today, in theory, to build a computer with a software almost as "soft" (quantum psychology, like, when you see a coat-hanger in the dark, you might start by seeing someone, then once you process more what you see and look at it longer, you might see a shadow, then you might see it's actually a coat-hanger) and less unbiased than a human. Since everything you see, read or hear is only a perception, you can't be 100% sure of anything. So the ability to accurately judge how probable the information is is the key to accurate forcasting.