The Turing report - reporting on scientific process...

Last night's talk from Michael Bolton was as interesting as I knew it would be.  One quote in particular caught my imagination, that I had to put onto Twitter ...

This got me thinking a lot about how reports are given into progress in science.  It's well worth watching such a press release on say "discoveries on Pluto", "string theory" or "artificial intelligence".

The language used is very much like that of a stand-up - and is probably what we need to be emulating in our reports ...

• This is our journey to date
• This is what we've learned
• This is where we're going next

Take for instance the Turing test - this is the test of a machine's ability to mimic thinking by being able to fool someone into thinking they are really talking to a human being (not a machine).

We've made huge leaps in this field since Alan Turing first postulated it.  We're not quite there yet, although we've been "getting there" for about the last 20-30 years.  We probably have a couple of decades to go yet, and even once there, it's a goal that once achieved, artificial intelligence will look for the next goal.

The Turing test caught my attention this morning after reading that Microsoft's recent AI Tay - which was designed to mimic the behaviour of a teen girl, and interact online - has evolved it's behaviour after interacting online.  It now is exhibiting the behaviour of a sex-crazed, foul mouthed, Hitler apologist - so Microsoft had to turn it off (reading this I wondered - did 1st April come early this year?).  I guess you could say her father Bill Gates grounded her after reading what she was saying on social media.

[Perhaps this is also a lesson on the corrupting influence of social media]

In a fit of whimsy, I thought "what if we tried to apply the same rules to scientific reporting as we do to testing?".  With a never-ending focus on "the numbers".

Ladies and gentlemen, I give you "The Turing report" ...

BTW - I like to call this "The Bach Test", after something I learned from James Bach on RST.  If you are trying to get a message across, and someone just isn't listening, and being obstructive, you need to consider walking away (not necessarily telling them to "fuck off" though).  Because when you overly compromise towards someone who is not listening and only has a rudimentary grasp of what you're doing

• you can never satisfy them because their evaluation of what you're doing is entirely arbitary
• to make things even more challenging, you're also hamstringing your own efforts.  It's like someone putting a ball and chain on your leg, saying it'll strengthen you up so you can run even faster.  But they never take it off.