I'm making up new term: rate-limit anxiety (for APIs)
Context:
When CPython GitHub bots were still using personal access token, we were rate-limited to 5,000 requests per hour.
Now that both miss-islington and bedevere are GitHub Apps and accessing APIs using installation access token, the rate limit is increased, based on the number of users on the GitHub org , and capped at 12,500 requests per hour, relieving our rate-limit anxiety 😌
What systems are people using for bot prevention these days? With captcha being mostly a pile of shite (and mostly bad for accessibility too), what alternatives are out there?
@brunty I'm using Akismet for my forms. I haven't implemented custom spam/ham reporting yet but I want to. It's decent at detecting spam but isn't perfect. Better than requiring people solve problems, though!
I reduced the posting frequency of some of the bots I've made recently, so if you've avoided them before for being too noisy, consider giving them a try now!
@dalereardon I haven't really touched Twitter's API in months. If I remember correctly, the API for writing is now free, and you have to pay for reading? If that's the case, you might just have to scrape the data.
"For distorted text fields, humans took 9-15 seconds with an accuracy of just 50-84 percent. Bots, on the other hand, beat the tests in less than a second with 99.8 percent accuracy."
@root42, I believe the APIs can be used to only request specific datums, e.g. not request anything else but proof of humanness. If in some country that's not possible, which would be surprising, there can be a trusted platform provider who only relays attestation that a person is human.
i am so traumatized from #bots pretending to be #human on twitter that from now on i only want to follow accounts that are run by actual ppl, but idk what to believe now that checkmarks don't mean anything anymore. even on ppl's personal websites, ppl anonymize so much about their online presence now — i'm guessing out of fear of losing their jobs? that or their entire online presence revolves around their work & that's not what i'm looking for
@br00t4c publications are usually brutal about describing a study but not linking to it, but they usually at least give you a hint. I finally found a different article that at least named an author, and from there the paper - it's the most recent work listed here
For those that weren't aware, IEMBot from Iowa State has come to the Fediverse officially. I'll leave it up to the bot maintainers whether or not they would rather continue running their own version which many of you are most likely already following.
@WXFanatic Don’t get me wrong, BTW: we all truly appreciate the work @akrherz has done to bring another instance of iembot feeds. If our X-perience has a lesson, we need multiple sources for critical information like IEM.
2023 is the year that @vgan and @WeavingWithAI get together and make a bot-baby that makes people guess the prompts used to create an AI art image #bots#AIArt#botmakers
