Show HN: Thesaner.com – a troll-proof platform for harder discussions
3 コメント
‘Troll-proof by design’ is a very bold claim. Anyone who has spent time managing online communities or game servers knows that trolls are incredibly adaptable. If you force a strict logical structure, won't you just end up with 'bureaucratic trolls'—people who format bad-faith arguments or sealioning perfectly within your system's rules? How does the platform differentiate between a well-formatted bad-faith argument and a genuine complex discussion?
This is a very important question. I think there is an answer in the incentives of the platform but will have to bear out in practice. The platform incentivizes the development of an argument. Coming up with an argument is already a difficult exercise because of the structural demands of consistency and plausibility inherent in it (something in stark contrast to the ease of faking a fact). Incentivizing an elaborative structure makes it much harder by preventing inferential sleights of hand (bad-faith use) by forcing those inferences to be explicit and open to scrutiny. The most obvious cases should be detectable with LLMs.
Another important defense is that this platform is going to be built around open discussion. I am not trying to prevent being wrong, just to provide a place on average can correct them, even if the participants are not always operating in good-faith.
Another important defense is that this platform is going to be built around open discussion. I am not trying to prevent being wrong, just to provide a place on average can correct them, even if the participants are not always operating in good-faith.
The website remains very much a work in progress and I welcome all feedback. If you have any doubts, please ask about it.
PS: The website only supports discussions on science, policy, and technology at the moment. I am working on making it general-purpose as soon as possible.