I’m deeply uncomfortable with this because while it may be an effective tool in general for identifying AI writing, false positives can be SUPER harmful

Show Post

The first time I saw Patrick Stewart in person was at the Cambridge Union. He talked movingly about how, as a kid for whom the formal education system did nothing, he had sat on a hill looking down on his school on a day he was supposed to take an exam - and how he'd got from there to here.

Show Post

For balance, when I saw Patrick Stewart in Oxford, at a fundraising thing at the Playhouse, he was wearing a spectacular pair of cowboy boots and spent the first 15 minutes talking about them. Also very, very memorable.

Show Post

I see in this example a really clear example of the tension between years of progress towards open data, tools, and science anyone can participate in, and the tendency of LLMs to swamp systems with low quality content.

Show Post

However, the important bit is this: 'We acknowledge that most of the submissions come from people acting in good faith, who wish to participate in the process of astrophysics and discovery. Their work is enabled by the effort put in over many years by many people to provide open tools and data.'

Show Post