Unfollowing here and subscribing their blog helps.

What I mean is that the most skilled and resourceful use custom feeds and labelers and their experience is very much not the default, which implies that that which is popular here is also kinda mid.

I see your roasted brussels and raise you roasted brussels with pomegranate seeds and feta in a pistachio vinaigrette.

Saying Bluesky is anti-something is like saying Times Square and Tiananmen Square and Tahrir Square are against something. Bluesky contains multitudes, none of which should be taken seriously just on the basis of numbers.

Remind me in 6 months to check whether those authors got the hint or whether they're still spamming slop everywhere they can.

They didn't even use a current model. More recent ones do a lot better.

The most underutilized feature

Take a well-deserved break and when you get back (or if!), hit me up and I can share tips from personal experience with this sort of thing.

Thanks for fleshing out the concept! How do people think about the problem where proxy measurements like clicks or viewing show satisfaction but only because the user is choosing from a limited menu & qualitative study shows they aren't actually getting what they want?

Interestingly, high self-reported confidence is associated with lower accuracy. This is dissimilar to most of the literature.

In our setting, we can track individual forecasters over time. And thus we can observe: this result is driven by overconfident forecasters.

🏆 Institutional: The Brazilian Reproducibility Initiative is a nationwide effort to evaluate research results in laboratory biology & the largest coordinated replication effort in the field worldwide, showcasing the potential of country-level research improvements. @redebrrepro.bsky.social (3/5)

Same thing happened on Quora when they rolled out the creator program and didn't pay amounts that the best writers would care about.