Jump to content

Rich Stowell talks first principles with a ChatGPT robot.


Garfly

Recommended Posts

 

 

https://www.communityaviation.com/hubfs/Nine Principles - Rich Stowell and ChatGPT - 27Sep2023-1.pdf

 

 

 

 

And starts a new blog (which, he reckons,

was written by a human  ;- )

 

WWW.COMMUNITYAVIATION.COM

We're building a network of experts in every field of aviation that you can access for knowledge and hands-on training.

 

 

rich_stowell_picture_web_page_copy_copy

Rich Stowell.  Rich took his first flying lesson in 1982 and began his career as a full-time flight instructor specializing in spin, emergency maneuver, and aerobatic training in 1987. He is a recognized subject matter expert in loss of control in light airplanes, the 2014 National FAA Safety Team Rep of the Year, and the 2006 National Flight Instructor of the Year. A 20-year Master Instructor, Rich is a Charter and Life Member of the Society of Aviation and Flight Educators, and a 35-year member of AOPA, EAA, and IAC. He has logged 10,200 hours of flight time with 9,100 hours of flight instruction given, 34,700 spins, and 25,700 landings.

 

 

 

 

 

 

 

 

Edited by Garfly
  • Agree 1
Link to comment
Share on other sites

Well, Rich has tried to give us SOME idea of what it knows - and what it doesn't (yet).

 

Meanwhile, from today's NYT:

 

 

WWW.NYTIMES.COM

At a U.K. summit, 28 governments, including China and the U.S., signed a declaration agreeing to cooperate on evaluating the risks of artificial intelligence.

 

 

 

 

 

 

 

Edited by Garfly
Link to comment
Share on other sites

21 hours ago, facthunter said:

It is fairly basic aircraft Knowledge. I have NO idea what the robot knows.   Nev

An AI only knows what it can access in its allowed databases.  It will search these databases and output relevant statements.   However if anyone  is interested here is an example of AI vs preprogrammed. 

Suppose that we make an artificial leg, identical to human leg in every way.

1.  We get Beckham to kick a ball and record how each muscle reacts.  We then program the computer to control the muscles in an identical way.

2. we use AI.  We make a target and give the computer a feedback of how close the kicked ball is to the target.  We then tell the computer to kick the ball as close to the target as possible.   Then tell the computer to start.  It will start flexing ,"muscles" in a numerical way.  When it finally gets a feedback that it has hit something it will store the value of the distance of the target.  It will remember all the actions of the muscles to achieve this result.  It will then (preprogrammed) to star small variations to these muscle actions.  If through iteration it gets a closer score all the muscle actions and target distance will replace the previous one.   This will keep making small changes until it gets the target.

 

In fact the AI will get a better score than Beckham.   But it will take a very long time to do it.

 

AI uses comparisons for the various questions asked and store a result (answer) when it gets some agreement to its given task.  It can only use the database access it was given. If it were to use Facebook and someone had said that Mooneys can fly backwards then it would believe it if it had not been challenged by someone on Facebook.  Industry use only verified scientific databases for AI work, reduces error. 

  • Informative 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...