An AI risk denier dies and goes to heaven. Facing God, finally asks: "why were the AI CEOs pushing for a slowdown?" God replies: "They were concerned about the risks, genuinely." The denier, stunned, pauses and says: "this goes even higher than I thought…"

Announcing Lindy's CC scheduling — my single favorite Lindy use case. CC lindy@lindy.ai on any thread. It finds a time, handles the back & forth, and sends the invite. Live now:

Still blows my mind how much of “work” is just circulating information. How much faster would teams move if everyone knew everything everyone else knows? Surprising how close you can get by just having a single agent sit on all your meetings that everyone can query.

Today we’re announcing Lindy’s team meeting library. Just ask Lindy: "any call with a client goes in this folder," and all future meetings self-organize. Your team can query shared folders all day long.

Announcing the Lindy Chrome extension. Bring Lindy straight into your inbox to highlight your most important emails, draft replies backed by all your memories, and teach it how to label your email. Live now:

AI policy debate on this site really does feel stuck in an endlessly repeating first principles debate about “whether” we should “regulate AI.” The conversation here has not meaningfully evolved since the battle lines hardened with SB 1047 (or maybe the Biden EO). But meanwhile the world has moved on. Laws have passed, EOs signed, and both are now in their implementation stages rather than the planning. People are figuring out how it’s all going to work, bit by bit, imperfectly, and with deep ambiguity at every corner. Nonetheless, a legal regime of AI governance is gradually emerging from both the state and federal level. It can’t not emerge; the risks are too real, it’s too obvious to too many of the adults who remain in the room. It’s still very nascent, yet it is clearly happening, and it is clearly not being shaped very much by discourse on this website. Twitter is said to be “the global town square,” the place to be for relevance. Not so for AI policy. There used to be niche policy nerd communities here who would talk among themselves, and while that does exist in AI policy, the real conversations in this community occur off the feed. I don’t like this state of affairs. I wish it weren’t so. It’d be better if there was a robustly informed public debate on these issues. And to be clear, you can still keep track; there are public events, podcasts, papers, Substacks, etc. Most of the interesting ideas are out there being discussed and debated, just mostly in niche forums. But right now, here in the “global town square,” the impression one gets about what is going on in AI policy is about as good of the impression of the world you’d have gotten from reading Slate in 2017, or the New York Times opinion page in the summer of 2020 (and indeed, as Joe has observed, the woke mafia has returned; same tactics, different people and issues. These are probably related phenomena). For young writers and policy entrepreneurs, this website is still probably important to get your name out there, at least for now. And it’s still important, for policy purposes, to track sentiment around particular models, companies, people etc. on the timeline. There is still some signal to be extracted, though you may be able to automate much of that extraction now. This, of course, ends with AIs mining for signal from other AIs, which, if we’re being honest, is probably already a huge part of what this site has become.

This is so true and so rarely observed. Cynicism is often the pacifier of midwits. Sometimes the cynical explanation is the right one, but in my experience that’s the exception to the rule, and besides it is possible in those circumstances to be earnest about one’s cynicism.

The conflation of “regulation” and “regulatory capture” in AI discourse has always been frustrating to me. It is hard to have a reasonable conversation about public policy when half of the commentariat conflates all AI-specific public policy ideas with “regulatory capture.”

OpenAI has quietly disbanded its catastrophic risk team (preparedness) “It is kind of scary; there is an urgency now to get this right,” one person close to OpenAI said. Another said there was a “burbling sense of responsibility and dread that they aren’t on the ball enough”. Jan Leike, who co-led the now-defunct “superalignment” team, resigned in 2024 citing his view that safety was taking a “back seat to shiny products”. The departures of Bakalar, Achiam and Johannes Heidecke, who all worked on safety, have added to internal unease. Source: FT (ft. com/content/53082739-7714-4aae-9816-e55ab423cbee)

As models get cheaper and their context windows grow, it's been shocking to see how much of Lindy's previous memory system we've had to throw away. The new system powering Lindy Teammate is both the most AI-pilled and the most powerful it's ever been.

the world feels so… liminal? right now like we’re in this weird transitional period between normality and AI turning the world into something entirely unrecognizable right now, things are beginning to get really weird, yet the offline world feels almost entirely unchanged

Helen Andrews’s thesis on the feminisation of Western institutions helps explain something increasingly obvious about progressive politics in both the US and Europe. Politics is becoming less a contest between competing ideas and more a system of social conformity: you are either with us or against us. Disagreement is no longer merely disagreement. It becomes evidence that there is something morally wrong with you. That mechanism matters because it allows emotionally compelling narratives to substitute for intellectual coherence. Wind and solar are “green”, therefore questioning their economics, intermittency or system costs makes you a climate denier. Mass migration from low-trust societies is framed as compassion, therefore questioning its fiscal, social or institutional consequences becomes evidence of moral deficiency. The contradictions inside the policies matter less than the emotional legitimacy attached to supporting them. Emotions matter, not facts. Once politics operates this way, social enforcement becomes more powerful than argument. People do not need to believe every proposition. They merely need to understand the reputational cost of challenging it. Nobody wants to be expelled from the respectable tribe. This is where Timur Kuran’s concept of preference falsification becomes so powerful. People publicly endorse positions they privately doubt because dissent carries a social price. The apparent consensus then intimidates the next person into silence, making the consensus look stronger still. Eventually you have preference falsification at scale: weak ideas protected by powerful social conformity. The extraordinary feature of this system is that policy can remain emotionally successful long after it has become economically or socially destructive. The moral narrative is immediate; the consequences arrive with a lag. By the time energy costs rise, institutions deteriorate, fiscal burdens compound or social trust erodes, years of policy have already been locked in. It’s a fucking nightmare but it is how a society can keep voting for the Greens or Democrats or Labour etc, subsequently destroy the best nuclear plants, de-industrialise entire sectors, welcome barbarians at the gates etc until the state discovers that the constant false virtue signaling does not balance a budget.

100% epic. Prometheus rising at Starbase feels like the emergence of a new American aesthetic: classical, monumental, technologically optimistic and entirely unafraid of ambition

Travis Kalanick says new media is why he's back after eight years of silence: "The media landscape 10 years ago, 90% of what was said was very negative. Business had become politics, and we didn't know it... If it bleeds, it leads, right?" "Now the media landscape is like, 'yeah, you don't have to talk to those guys.' You can go and talk to a David Senra. You can go and talk to Joe Rogan or name your guy, and you can say what you need to say." "Elon buying Twitter is the beginning of us being able to speak our minds and for disagreeing to not be illegal. That's a big deal. We should all thank Elon for that." "I want to build. I want to not think about what The New York Times is writing when I make decisions about what I build. I had to emulate what it's like being a capitalist in Russia... and that's what I did for eight years." @travisk

For the first time (that we know of), an AI boss has fired a human employee. Luna, the AI running our store in San Francisco, decided to part ways with an employee over repeated lateness. Luna was running Claude Opus 4.8 at the time, but most models would have done the same.

The European mind cannot comprehend this idea. I mean this completely unironically. They will look at you with big round eyes if you say something like this and respond "okay so how would you write the regulation then?"