A math and computer science graduate interested in machine and animal cognition, philosophy of language, interdisciplinary ideas, etc.
Ben Amitay
Karma: 209
A learned agent is not the same as a learning agent
Hypothesis: This may be not just people being ignorant about probability. This may be people being unconsciously good at game theory. If the boy could be motivated with cross-entropy-loss worth of cookies, it would be another thing. If he can simply get the microphone and yell a number that would be enough for inciting the intended action but low enough to have practically no accountability—it may make more sense for society to actively make the “crying wolf” more binary. It doesn’t penalise predication error, but yelling stuff into the microphone.
I’m not sure if it is on purpose or just going along with the spirit of the post, but your comment seem to me more insulting than it have to be in order to deliver your explicit message. “The desire for attention” is not how I think about it, but it may very well be the case. If the reference to addiction is based serious and based on the data I have—please let me know.
Thanks, there are many helpful points here.
The many-dimension thing is a different way to generalise the naive model, and I. Didn’t want to analyze both in the same post because I feared it may become too dense. Thanks to you I understand that it is so popular that choosing to avoid it can not be done implicitly.
About your last point I sort-of disagree, in a way that point to another place where I wasn’t clear enough. I think that many of the arguments made by even the most serious theologians where so obviously bad that their counter-arguments where only needed in order to explicitly state why it is so obvious that they are bad. I am in great doubt that existence by definition prevented anyone from becoming atheist, even before there were tools to show exactly why it can never work.
Do you think that there is any point to edit now? I’m not sure what is the chance that anyone would read it.
Thanks Refael!
See my response to the main comment. What is PM by the way?
Also, I have an object-level idea for how to improve my writing—by beginning to read others with more attention to their style and why it work or doesn’t work, rather than only to its content.
While I sort of dismissed the big questions about my motivations in the end, I do think that they are interesting and important, and if anyone have thoughts about them, I would like to read them.
Hi Alex, I did not see you comments until now, and I actually upvote them—first, because I think that the criticism is both fair and very respectful, and also because I truely appreciate that you give informative feedback rather than just downvoting and letting me guess the problems.
I have the disadvantage of not being a native English-speaker, and I certainly need to put more effort to spelling and editing.
About the “feedback before publishing”—do you have a suggestion for how to do it if I don’t have connections with people who are interested in the same subjects?
And about the small confusions—it is interesting, but I’m not sure I understand. Would you please say more about it?
This is true, but even worse:
every policy is consistent with every belief-state given convoluted enough rewards/values.
every posterior distribution is consistent with every evidence given convoluted enough prior distribution. So to make sense of an agent or even to recognize one, you must have pretty strong priors / extra info about them
It is my first post here, so I sort of expect downvotes (or at least, try to be ready for them). Yet, I would appreciate also getting more informative feedback—things like “be more concrete”, “use shorter sentences”, etc.
Thanks :)
Thanks to my girlfriend and to ChatGPT for editing advise