Self hosted trail mapping and GPS tracking with Wanderer

@selfhosted

I deployed an instance of Wanderer and defined a route in preparation for my upcoming trip to The Great Smokey Mountains National Park next week. This service uses #ActivityPub but I haven't been able to follow my account from #Mastodon yet... Even so I'm glad to add this to my #selfhosted arsenal. I will be live streaming on #Owncast during my hike, and I'll also edit the footage for #PeerTube.

https://wanderer.eom.dev/trail/view/@eric@wanderer.eom.dev/7eb769b9157c8c7?share=ojps3c4mo6q0v169l14sj4i5dkje867s

you are viewing a single comment's thread
view the rest of the comments
[–] [S] 2 points 1 day ago (2 children)

@selfhosted

@curbstickle I didn't make Wanderer, I just deployed it on my servers. Weather or not they used AI is no business of mine and I don't think it is reasonable to ask me to investigate every service I deploy for their use of #AI. I don't really care if they used it or not. It isn't my project.

  • source
  • hideshow 4 child comments
  • [–] 1 point 13 hours ago (2 children)

    @eric @selfhosted @curbstickle

    If you choose to selfhost an app, I think its fair for people to criticize you for it being vibe coded. Choosing to run it on your servers and then telling people you've done so is an effective endorsement you can't just handwave by saying you didn't code it.

    Also as the owner and maintainer of your server, you should do some due diligence on what you host. Code without a community backing it is more likely to have security flaws go unnoticed, especially if the code isn't even being written by a human.

    The same way emdashes and negative parallelisms get overused by LLMs, so will security flaws. Someone will find an exploit and that same exploit will then be usable against tons of vibe coded projects, in a way human written code, which is more varied even if "worse" in other ways, would have prevented

  • source
  • parent
  • hideshow 4 child comments
  • [–] 1 point 13 hours ago

    That really isn't the issue, and no, that would not be permitted here. It would be one thing to note that an LLM was used in some part of the project, it would be entirely different to criticize the poster based purely off whether or not there were any contributions involving LLM.

    That would be a rule 1 violation.

    If you want to comment on a known vulnerability, thats one thing. Commenting that its vulnerable because an LLM had a pair level patch is not.

  • source
  • parent