You are viewing a single comment's thread from:
RE: Are bots bad for Steemit? - 🤖 BOTS! 🙀 Act 2
A bot is just a tool, like a knife. You can prepare a meal with a knife or kill someone.
There are a couple of good ones like @cheetah
Most of them are bad
Why are most of them bad?
Because they auto vote, not really adding value to the platform.
Not all the bots which are not @cheetah are auto vote bots, but probably most of them are.
I think "adding value to the platform" means different things to different people. The job of curation by voting is, in my view, adding value to the platform, even if done by bots. My question is really is it harmful when bots do the same things as humans. Because there is nothing a bot can do that a human can't, though they can do it faster, tirelessly, and by their own programmed logic. So my question is, is there a negative effect when bots act (vote for example) compared when humans act?
My own auto vote bot is useful to me for finding posts I'll probably like in the sea of posts made each day and voting on the "best" of those, according to some rough judgements based on the various data of a post. I also have the best possibility of voting early on it to give some of the curation reward back to the author in the reverse auction feature.
One of the things the whitepaper (page 18) outlines is that distributing the currency in any way is valuable, even when this constitutes abuse.
I wouldn't consider the use of bots and abuse at all, but some might. Using bots to up vote is not a direct human contribution, that's true, but using one's stake to vote, even when delegated to a bot, is valuable I would argue.
I'm sorry I have to disagree
A bot can't tell the difference between Davinci's Mona Lisa and a picture of a turd. Excuse my language.
I think you got my logic backwards with that statement. Worded differently (but saying the same thing) is that whatever a bot can do, a human also can do. I think you think I'm saying bots can do everything a human can. I was careful with my wording here and it doesn't mean that, but I'm sorry if it was confusing, perhaps I should reword it.
The point of it was that the API allows only a certain set of operations (voting, posting, Steem transfers, etc.) that are equally accessible to humans or bots, in fact there is no difference from the API's perspective.
I decided to try your example anyway, because a bot actually could tell the difference between images if it was programmed to well enough. I tried the Wolfram Language Image Identification Project for it.
I used the one and only Mona Lisa and this picture of a turd. It thought the Mona Lisa was a person and the turd was slug! Haha, so it could tell the difference but got the actual classification wrong.
I shows we have some ways to go, I'll concede.
Points taken, thank you!
Cool project. Didn't know it existed. Sooo, you could use this to even better filter posts, before manually going through the filtered list and decide as a human to upvote, what percentage upvote it gets and even decide if it is worth a comment. If classification is good enough, your bot may even send an auto comments making some intelligent remark to kick start a conversation. I still would argue, anything to do with reward distribution, let that to be decided by a human individual.
Filtering: it would also be very good if the bot operator balances the filtering rules such that the entire community can have posts and comments in the "check this list of posts and comments". When for instance filtering on high payout authors, or only on newbie than a bot can have a unbalanced influence on the reward distribution.