Skip to main content
22 events
when toggle format what by license comment
Feb 10, 2025 at 13:04 history edited Starship CC BY-SA 4.0
added 27 characters in body
Feb 7, 2025 at 15:54 comment added Alex @Starship If it were OpenAI, then they would have announced it, because that's already out in the open. Whoever this is, it almost seems like they aren't seeking a long-term relationship-- just an experiment, to collect as much AI training as people are wiling to give-- and once the AI quality stops increasing, the experiment will be terminated, and this entity will walk away never having been identified. If they were seeking to make the results in the public domain, then they would require a much higher rep threshold for vetting answers. But when the goal is AI training, any layman can do it.
Feb 7, 2025 at 14:57 comment added Starship @Alex Who it is isn't much of a mystery. They literally announced a partnership with OpenAI...what's strange is why they don't say so
Feb 7, 2025 at 14:55 comment added Starship @starball Not all robo-reviewers are doing it for rep/badges. For example, I've seen some mods complain of robo-flaggers who are just trying to increase their helpful flags count, even after the Marshal badge
Feb 7, 2025 at 6:15 comment added Alex One Q. I can answer based on my MS in CS. "Why do you feel it okay for you guys to post "checked" AI generated answers while it is suspendable for users to post checked AI generated answers" The idea is, they're trying to build an LLM specifically off the SE users. Using another AI for answers would introduce another AI's data into SE AI's, increasing the required space without adding value. That's probably also why they are concealing the source of the AI -- it's probably a partisan figure like Musk trying to absorb SE's quality into an AI he's building, and revealing would turn users off.
Feb 5, 2025 at 13:49 comment added controlgroup Keep in mind, too: “a database of AI content” is not functionally much different from “ChatGPT with a search bar”. SE’s usefulness stems from the fact that it’s better than a database of generation, not that it could be one.
Feb 5, 2025 at 9:20 comment added tenfour SE can be a community which coalesces expertise of those eager to teach & communicate, OR it can be a database of AI "content". You cannot have both; one cultivates a community and the other minmaxes answers to questions, which will push out the most valuable authors of the site. I fear SE doesn't care about the community.
Feb 5, 2025 at 7:32 comment added Resistance Is Futile "Why do you feel it okay for you guys to post "checked" AI generated answers while it is suspend able for users to post checked AI generated answers" The mentioned user posted unchecked and incorrect AI generated answers.
Feb 5, 2025 at 5:46 comment added starball Mod I didn't see anything in the post or help center page that would make me concerned about robo-reviewers- unless I missed it, there's no mention of rep or badge rewards for reviewing these. so my question has been and still is... why would anyone spend their time reviewing this? if anything, I expect the company to run into a problem on their end of lack of interest in reviewing this stuff. fingers crossed that they don't add rep as a reward for reviewing, because that would attract robo reviewers like mad.
Feb 5, 2025 at 0:39 comment added Starship @NotThatGuy Posted in 2024 could be 13 months old...thats not that different from a question posted 3-4 years ago
Feb 5, 2025 at 0:35 comment added NotThatGuy "Posted in 2024 or 2025 - Why" - no-one, especially not the asker, cares any more about questions from years ago that no-one paid much attention to (a lot of those questions would've been autodeleted at 365 days). If lots of people paid attention to it, but no-one managed to answer it, AI is most likely just going to produce garbage. For newer posts no-one paid much attention to, most likely only the asker is going to benefit from an answer that anyone can get from ChatGPT (which isn't great for providing future value to anyone else).
Feb 4, 2025 at 20:54 comment added Andras Deak -- Слава Україні The only "unfavourable" is "risk of loss of profit" for the owners of the company. So I guess that happens when the partnered AI providers are unhappy with the services of the flesh slaves.
Feb 4, 2025 at 20:44 comment added Starship @KevinB Hence why I asked. Wishful thinking. But also sometimes "everyone is really mad at you" does lead to some changes. 5-10% of the time, but it does occur
Feb 4, 2025 at 20:42 comment added user400654 "If it is the company, as I suspect, is "everyone is really mad at you and thinks its terrible" considered an unfavorable result?" isn't "everyone is really mad at you" the standard state now days?
Feb 4, 2025 at 20:36 history edited Starship CC BY-SA 4.0
added 338 characters in body
Feb 4, 2025 at 18:37 history edited Starship CC BY-SA 4.0
added 288 characters in body
Feb 4, 2025 at 18:30 history edited Starship CC BY-SA 4.0
added 219 characters in body
Feb 4, 2025 at 18:28 comment added user400654 "Users with at least 50+ rep (default value) on the specific Stack Exchange site will be able to see and evaluate the private answers. This reputation requirement can also be customized per-site."
Feb 4, 2025 at 18:28 comment added Starship @KevinB Source?
Feb 4, 2025 at 18:27 history edited Starship CC BY-SA 4.0
added 423 characters in body
Feb 4, 2025 at 18:27 comment added user400654 "And who will be able to see these and mark them correct anyway?" anyone with 50 or more rep, unless otherwise defined on the given site. so you and I will be able to see and potentially approve these on sites we've never participated on.
Feb 4, 2025 at 18:24 history answered Starship CC BY-SA 4.0