Hello friends,
I know it’s a rough opening, but we stand on the precipice of more war. I’m talking specifically about Iran, where the UK has withdrawn its embassy staff in anticipation of an attack by the US. In preparation for retaliatory attacks, American embassy employees were advised to leave Israel today. And right now, the administration and our military leadership are engaged in extremely public histrionics over any limitation on its use of AI for mass domestic surveillance or fully autonomous weaponry.
When I launched this project, I wrote a note to myself, saying that I wanted it to be,
“a way to keep a handle on what’s happening without drowning in the churn…. to validate grief and rage without getting entirely swamped in it… a steady voice in the storm.”
So this was Year 2, Week 9. Steady as we go, yeah?
I recently came across a fantastic thread from Molly Crockett that introduced me to the term epistemic vigilance. It describes the way we protect ourselves against being misled, and it has two necessary subcomponents: scrutiny of the content itself and scrutiny of the person offering you the content. She writes, “When I get info from another person I can ask: how does the content relate to their expertise? Their lived experience? Their interests and biases?”
Her thread is specifically about how we think about LLMs, and why we must fundamentally treat the products of agents like Claude differently - no matter how carefully we scrutinize the content they produce. She’s also published on how LLMs create an illusion of understanding that is particularly insidious in research contexts. That paper closes with the reminder that knowledge production “is a fundamentally social practice that is shaped by the norms of its institutions.”
Right now, those norms are changing radically - and not for the better. I had already been planning to write about how NSF is retooling its priorities to be all about AI and quantum computing. I think it is entirely reasonable to hypothesize that the disruptions in the GRFP are linked to new budget priorities that subsume everything else.
It’s the same pattern over and over again - unqualified people in positions of power are diverting all the resources into executive priorities that simply do not align with the best interests of the vast majority. It would be tedious if it wasn’t so dangerous.
I will leave you with some questions tonight, rather than my usual exhortations: What do we do once the pattern is clear? What are the perks of realizing that the chaos is all-too-predictable? And how might asking those questions help process the next onslaught of news?
Liz