—–
Now to something far more useful.
Learning by guessing. This is how the mind seems to work and is *also* used as a ‘trick’ by salesmen and politicians (or salesmen politicians) and news people in order to convince you by leaving just enough vagueness to allow you to feel that you’ve finished the puzzle they present, which you have, but used uncritically (that is, if you 100% trust them), can mislead you in whatever direction you agree to go with them in.
Used in Neuro-Linguistic Programming (NLP) a well. [some people call it hypnosis but you could simply call it the power of “if”]
—
Moving on to :
Volume II: Patterns of Plausible Inference
which is what I wanted in the first place anyway.
Volume I: Induction and analogy in mathematics first 50 pages were interesting and useful but the last 200 were math-y examples going over induction, analogy and the third important thing that I didn’t need.
It’s in DJVU format (ugh) but my reader converts to PDF so I can at least read it more comfortably.
—
The problem with “single source” information.
I posted something about the problem with right-wing news making claims picked up by Fox picked up by Trump broadcast on Twitter, reported by other news, seemingly confirming right-wing news making the claims.
But this is nothing new. This is from 1988, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference by Judea Pearl, which shows the same process and the same problem in a different context.
You can find examples of this in many areas but news (from whatever leaning) is a prime example.
——
Yeah, I use “whether or not” even though I know it’s wrong. It “feels” more right even though it’s not. I think we do it to sound less arrogant by stepping back a little with “or not”. False humility for social reasons.
—–
Going through network probabilistic reasoning and one strange that keeps bothering me is this idea of causation without time. [that is, the relationships themselves provide the cause and effect, with time effectively becoming “next action” regardless of how long it takes.
But that’s not how things work. It may simplify the models but it takes time for processes to work, for ‘before/now’next” “this/thus/that” to occur.
On a purely theoretical basis I suppose it’s fine to entertain the ideas but time matters, despite being told all our lives that “time is irrelevant” by somebody or another and believing it, at least once or twice or now or might later.
====
oh interesting! I was thinking more of a mathematical thingie (or logic thingie or programming thingie or neural network thingie) but maybe it’s really a gene edity thingie!
—-
h yes! Genetic algorithms applied to networks… or genetic network theory.. or … something that sounds smart like that.
Nice! I trust my intuitions (that is, I follow them – has nothing to do with them being right or wrong true or false just that i need to follow them]
====
Changing minds. 0.0804 is the equiprobable point in massive Bayesian networks.
Different from humans in this way and yet, points to a similarity, how a seemingly minor piece of evidence is sometimes enough for us to abruptly change our minds.
—
—-
“Entropy is a measure of the cost of removing uncertainty”?
Interesting. I hadn’t thought of it in that way before.
====
==
Reading a nice defense of “why” they used logic for the text I’m reading and going through several of the complaints about logic.
One of them that I haven’t heard in some time but have used myself in arguments is the “brittleness of logic”.
The author intuits (correctly from 1988, although it wasn’t a far stretch then or even before then) that connectionist (brain-like) will ‘win’ over pure logic in AI and also correctly intuits that new languages and descriptive methods will emerge from there (as they have).
Brittleness of logic. I’m going to remember that.
—-
I blame excluded middle because contradiction is a fact.
====
Boolean – Never had that level of confidence in it 
🙂 Too many things could go wrong.
===
Love the concepts of closed-world / open-world assumptions. Of course, in this case you’re the open world.
====
Other places where the closed-world opens up is the hardware, ,networking, operating systems, errors in transmission (ok that’s under networking), user-fault etc.
Then again, those can be closed too up to a point.
====
I remember once at Schering-Plough so long ago when a new employee in the IT dept one Tuesday morning, decide to upgrade the ODBC drivers on all the systems. He did a “push” that cascaded across thousands of computers in different countries at different times.
Of course some of us built systems that were dependent upon whatever ODBC drivers we had at the time, and systems began to fail immediately.
Took a day and a half for most of the updates to rollback but not until after a tremendous cost in loss of $$$ via loss of productivity.
====
Lack of persistence on the web (although then there’s staleness of the content itself but that’s another issue) is something that always bothered me, hence my ongoing interest in Internet Archive. At least it’s something.
====
For myself, I left the company long before it went under. I feel bad for your friend though, as it really stinks to have all of this intellectual and technical ‘capital’ just tossed aside like that
—
I don’t care about the company I used to work for in any case. I didn’t follow the buyout but their relationship was established for decades prior and even when I was there, the rumor mill always talked about imminent buyout. But it didn’t happen ’til maybe 8-10 years afterwards.
—