Link Search Menu Expand Document

Notes | Tutorial 33

Date: 2020-08-28

LW discussion for treeification (carrying on from last session)

notes

24min - if there isn’t a productive way to proceed it’s confusing/besides the point to keep discussing; solve that problem first

This is irrational.

CR answer “what is rational” - about error correction, about methods; do your methods allow for EC and help EC, or work against EC (or disallow EC). Irrationality is working against EC.

Rationality not about correctness; CR: about processes for continuous improvement

33min - can’t claim irrationality

not designed for positive outcome; aggressive

focus on simple points, only after track record of success scale up to more difficult things. simple enough messages that work well, don’t start with like 4 sections with 7 paragraphs. trying to deal with way too many things at once.

36min - discussion format: do 2 things at a time, each person picks one. they write their post, make claims, ask one thing. then you answer and say one thing. one response of your choice and one response of their choice. small chunks. two streams in the one convo. both get to lead the conversation, but also participate and answer them.

46min - anything else thigns

how to reply to a thing

The point is that every claim in general depends on assumptions. So, in particular, the critrats don’t have a disproof of induction that floats free of assumptions.

brainstorming

  • every claim has assumptions, okay
  • TAG: CR ppl don’t have a crit of induction that’s free of assumptions
  • can any proof be free of assumptions?
  • if not, then there are no proofs, really, right?
  • if there are, what are they?
  • can you use assumptions of the idea to argue against it? e.g. Bayesian assumptions to argue against Bayesian Epist?

ET: start slow, don’t skip steps. can do that later after track record of success

TAG’s position:

  • Theres are issues with CR
  • It’s significant if arguments are based on assumptions.
    • the particular assumptions are significant.
    • every argument is based on assumptions somewhere.
      • “The point is that every claim in general depends on assumptions”
      • “the critrats don’t have a disproof of induction that floats free of assumptions”
        • implying this is a problem with the crit
      • how do those two ideas gel?
      • is he saying: you can’t attack induction based on assumptions b/c everything makes assumptions?

he’s making a point about assumptions, but I don’t quite know what point that is, what are all the positions on assumptions ppl could have?

positions on assumptions:

  • assumptions are necessary, though a contradiction means some/one is wrong - can improve from there
  • we can get back to a set of assumptions that are “common sense” and go from there, can’t get better than that
  • these are my assumptions and tough luck if you disagree
  • some contradictions are always going to happen so we should ignore them
  • some contradictions are always going to happen => we can’t know anything for sure
  • there’s no foundation, all assumptions have assumptions
    • opt1: so nothing is right
    • opt2: so we do the best we can
  • basing ideas on assumptions means the idea is wrong
  • ET: minimise riskyness of assumptions; logically proven, self-justifying, obvious things (e.g. i think therefore i am)
    • also done by minimising the number of assumptions
  • for 2 ppl to agree on an argument based on assumptions they need to agree on the assumptions

can try to figure out what’s going on by matching TAG to the above

what pos do I have and how does it differ from his

69min - idea for moving fwd

TAG: too much work, our convo more specific

reply: what should I read, what are the bounds of the convo in your mind?

if you get stuck you can add more ideas to the system (the starting system) and see if it helps. can brainstorm new ideas.

93min - halting problem - idris, coq, ada

turing machine - constant overhead?

110min - in general benefit more by finding things with low error rate + things I can self eval on => expand on that => extend reach

before i do something

  • what have i done that’s similar i can point to which I succeeded at (went well, good example, worked out, no/few downsides, etc)
  • example is like 80% as hard or similar
    • better to say 10 things better in a row that are 80% as hard
  • this perspective: i want to do X, what have i done prior

  • i can do X,Y,Z what can I do next?
  • be careful w/ breakpoints (can’t predict problems)

for next time:

  • pick out sentences that are problematic for next time

You can leave a comment anonymously. No sign up or login is required. Use a junk email if not your own; email is only for notifications—though, FYI, I will be able to see it.

Comments powered by Talkyard.