I made a Google Scholar page for MATS. This was inspired by @Esben Kran's Google Scholar for Apart Research. Eleuther AI subsequently made one too. I think all AI safety organizations and research programs should consider making Google Scholar pages to better share research and track impact.

My impression (primarily from former representative Justin Amash) is that individual congresspeople have almost no power. Bills are crafted and introduced at the party level, and your choices are to vote with them or not. If you don't vote with the party and are in a safe district (which most are), the party will support candidates who will go along with the party line in the primary. The only scenario in which an individual congressperson has power is a close vote in a contested district where the party can't take the risk of a primary challenger. First, I want to know if which parts of this are true, and in more detail. One guy tweeting does not constitute proof. Are elected officials sitting around watching Netflix all day, or do they still have the power to privately horse trade to support their goals. Second, assuming it is true, where did the power go? Parties aren't real, the power must be in specific humans or incentive systems. Does it all rest in the top three elected officials of each party? Their staff? Unelected party officers? I'm very interested in evidence and arguments in either direction. 

As I was looking through possible donation opportunities, I noticed that MIRI's 2025 Fundraiser has a total of only $547,024 at the moment of writing (out of the target $6M, and stretch target of $10M). Their fundraising will stop at midnight on Dec 31, 2025. At their current rate they will definitely not come anywhere close to their target, though it seems likely to me that donations will become more frequent towards the end of the year. Anyone know why they currently seem to struggle to get close to their target? 

Reality itself doesn't know whether AI is a bubble. Or, to be more precise: whether a "burst-like event"[1] will happen or not is - in all likelihood, as far as I'm concerned - not entirely determined at this point in time. If we were to "re-run reality" a million times starting today, we'd probably find something that looks like a bursting bubble in some percentage of these and nothing that looks like a bursting bubble in some other percentage - and the rest would be cases where people disagree even in hindsight whether a bubble did burst or not.[2] When people discuss whether AI is a bubble, they often frame this (whether deliberately or not) as a question about the current state of reality. As if you could just go out into the world and do some measurements, and if you find out "yep, it's a bubble", then you know for sure that this bubble must pop eventually.[3] And while there certainly are ways to measure properties of bubbliness of different parts of the economy, it could well be that what looks like a bubble today may either slowly "deflate" rather than burst, or reality around it catches up eventually, justifying the previously high valuations. Uncertainty is sometimes conceptually split into two parts: epistemic (our limited knowledge) and aleatory (fundamental uncertainty in reality itself). My claim here is basically just that, when it comes to bubbles bursting in the future, the aleatory component is not 0, and we shouldn't treat it as such. In other words, there is an upper limit in how certain a rational person can become at any point in time on whether an AI bubble burst event will occur or not. Sadly, knowing where that limit is is in itself uncertain, which makes all of this not very actionable. Still, it seems important[4] to acknowledge that we can't just expect that doing any amount of research today will lead to certainty on such questions, as reality itself probably isn't fully certain on the question at hand. Ultimately, whether any burst-l

Quasi-beliefs This shortpost is just a reference post for the following point: It's very easy for conversations about LLM interpretation to get derailed by questions about whether an LLM can validly said to believe something, or to have a goal, or to hold a value. These are valid questions! But there are other important questions about LLMs that touch on these subjects, which don't turn on whether an LLM belief is a "real" belief. It's not productive for those discussions to be so frequently derailed. I've taken various approaches to this problem in my writing, but David Chalmers, in his recent paper 'What We Talk to When We Talk to Language Models' (pp 3–6), introduces a useful piece of terminology: I expect to largely adopt this terminology going forward, linking back to this shortpost as needed. For convenience, I expect to extend it slightly to include terms like 'quasi-goal' and 'quasi-value'. Additionally, unless otherwise specified, if I use a term like 'quasi-belief', later occurrences of 'belief' in the same text should be read as 'quasi-belief'.

I'm noticing evidence that many of us may have an inaccurate view of the 1983 Soviet nuclear false alarm based on reading "Did Stanislav Petrov save the world in 1983? It's complicated". The article is worth reading; it is a clear and detailed ~1100 words). I've included some excerpts here: I haven't made a thorough assessment myself. For now, I'm adding The Dead Hand to my reading list.

DeepSeek v3.1's "self-aware suffering AI" persona seems to have a tendency to write like /[poetic mood description]/ [suicidal message]