zlacker

[parent] [thread] 9 comments
1. tunesm+(OP)[view] [source] 2023-09-01 19:29:06
It's fun reading through the paper he links just because I've always been enamored by taking a lot of those principles that they believe should be internal to a computer, and instead making them external to a community.

In other words, I think it would be so highly useful to have a browseable corpus of arguments and conclusions, where people could collaborate on them and perhaps disagree with portions of the argument graph, adding to it and enriching it over time, so other people could read and perhaps adopt the same reasoning.

I play around with ideas with this site I occasionally work on, http://concludia.org/ - really more an excuse at this point to mess around with the concept and also get better at Akka (Pekko) programming. At some point I'll add user accounts and editable arguments and make it a real website.

replies(3): >>tomoda+61 >>french+9g >>high_p+lh
2. tomoda+61[view] [source] 2023-09-01 19:35:40
>>tunesm+(OP)
So basically a multi-person Zettelkasten? The idea with a Zettelkasten (zk for short) is that each note is a singular idea, concept, or argument that is all linked together. Arguments can link to their evidence, concepts can link to other related concepts, and so on.

https://en.m.wikipedia.org/wiki/Zettelkasten

replies(2): >>tunesm+a2 >>gitgud+vN
◧◩
3. tunesm+a2[view] [source] [discussion] 2023-09-01 19:40:49
>>tomoda+61
Sort of except that it also tracks truth propagation - one person disagreeing would inform others that portion of the graph is contested. So the graph has behavior. And, the links have logical meaning, beyond just "is related to" - it respects boolean logic.

You can see some of the explanation at http://concludia.org/instructions .

replies(2): >>coucha+j4 >>quickt+lq
◧◩◪
4. coucha+j4[view] [source] [discussion] 2023-09-01 19:52:07
>>tunesm+a2
> it respects boolean logic.

Intuitionist or classical?

replies(1): >>tunesm+b5
◧◩◪◨
5. tunesm+b5[view] [source] [discussion] 2023-09-01 19:57:00
>>coucha+j4
Intuitionist. Truth is provability; the propagation model is basically digital logic. If you mark a premise to a conclusion false, the conclusion is then marked "false" but it really just means "it is false that it is proven"; vitiated. Might still be true, just needs further work.
6. french+9g[view] [source] 2023-09-01 21:08:10
>>tunesm+(OP)
I've had the same idea (er, came to the same conclusion) but never acted on it. Awesome to see that someone has! Great name too.

I thought of it while daydreaming about how to converge public opinion in a nation with major political polarization. It'd be a sort of structured public debate forum and people could better see exactly where in the hierarchy they disagreed and, perhaps more importantly, how much they in fact agreed upon.

7. high_p+lh[view] [source] 2023-09-01 21:15:43
>>tunesm+(OP)
I don't think this is the goal of your project, so let me ask this way. Is there any similiar project, where we provide truths and fallacies, combine them with logical arguments and have a language model generate sets of probable conclusions?

Would be great for brainstorming.

◧◩◪
8. quickt+lq[view] [source] [discussion] 2023-09-01 22:32:09
>>tunesm+a2
You would need a highly disciplined and motivated set of people in the team. I have been on courses where teams do this on pen/paper and it is a real skill and it is all you do for days. Forget anything else like programming, finishing work, etc.
replies(1): >>xpe+nO
◧◩
9. gitgud+vN[view] [source] [discussion] 2023-09-02 03:05:52
>>tomoda+61
Isn’t this what Wikipedia is in essence? Ideas, concepts linked together, with supporting evidence
◧◩◪◨
10. xpe+nO[view] [source] [discussion] 2023-09-02 03:19:51
>>quickt+lq
I'd be stoked if you wrote more about this experience and shared it somewhere.
[go to top]