This sounds good, but how do we spin up a massive hype cycle that sways people mandating LLM use (like my boss)? They're already not paying attention to evidence from the early studies that indicate negative effects

I appreciated this article about it - theintercept.com/2025/08/16/n... it is maddening

Looking into the paper a little bit iceberg.mit.edu/report.pdf (the link is broken on their website), the first appendix seems to say "we used an LLM to match up which skills LLM tools offer with which skills humans use at work" which also feels shoddy

Here’s a handy protocol to determine whether AI could be helpful with your problem or project.

What if we paid the agents to quit? A la the (failed) "Clarence Thomas retire now" offer that John Oliver made a while ago. Seems like encouraging even a few defections in various field teams would gum up the works

At the time I was very confused, didn't know about the drone strikes or much of his middle east policy at all. He had won a nobel prize - wasn't he a big departure from W and the others?

I got an odd reversal of this: in 2010, someone in a cafe in Germany pulled me aside to tell me that Obama was worse than H*tler for what he was doing in the Middle East.

Are you wondering about the flight path or type of plane? I was on a flight to MDW today that followed the same approach