Someone earlier called Hormuz trumps Waterloo but it’s Waterloos for days.

The nature of tech and the resulting offense/defense imbalance empirically seems to have ended the era of hegemony. I suspect and hope it will be a powerful deterrent to great power conflict in general. The imbalance currently favors defense but probably will just be ambiguous most of the time.

Every time I read an Asimov story the first time, I rejected the core premises and over time I think I may have un-rejected every one? * basic alignment is hard * AI additionally creates x-risk * robots rights will become an issue * macro fates can overwhelm any individual action * more?

After listening to this podcast yesterday, I've already shared it with like ten friends. The idea is basically "what if all the things we call biological disorders are actually just behavioral adaptation and reinforcement?" Huge if true, and there is a lot of evidence. overcast.fm/+AA-yu9d6Vyw

California: Stay safe and support your community by reporting any potential abuse by Donald Trump's federal agents ⬇️

Robert Mueller pursued justice and the truth. May he rest in peace. www.npr.org/2026/03/21/n...

"Over the last few years, American tech oligarchs have become the greatest argument against their own existence. The sheer scale of their erratic behavior and political interference proves that extreme wealth concentration is a structural threat to society." -Joni Askola

1/2. He took the greatest military force in history, lost a war to a middle power in a week, begged the world to save him, & demanded the media lie about this & everything else. I try, but at a simple human level I do not see how anyone mistakes this man’s almost supernatural weakness for strength.

pmarca reading Scanlon: "oh man, they got this all screwed up!"

This is a good lesson for journalists: Use plain language to state true things clearly.

Coefficient Giving directed over $1b in 2025, the most in our history. I wrote a more personal annual letter this year, focusing on what a decade of grantmaking has taught me, how I'm thinking about AI, and how we're grappling with giving at this scale. 🧵 Linking the PDF here: tinyurl.com/3sh3mwhp

if anyone ever shows him the impact he’s having on nominal GDP we are just going to be cooked forever

Late Sunday entry to consciousness discourse: the hard problem is only known that way in western philosophy. In eastern, it is well solved since Buddhism is “science of mind”. To them the answer is dependent origination / no independent self. To me that clearly implies a consciousness gradient. 1/2

Holy shit. Crude oil futures going nuts tonight.

This is a pretty good point. Trump thinks the way to win wars is big shows of force and none of the other ways of winning a war are going to work in this case. His goal isn’t regime change because they have no one to anoint; it’s submission.

The incidents referred to from prior years made me lose faith in institutions in general. I don’t even understand why they did it at all.

Also it was gutted as soon as Trump took over, illegally, thus ensuring the payoff never arrived.

We really deserved a longer conversation about this but I think Americans should assume it’s over now. Sam Altman decided this was a fine issue to retroactively decide he did “sloppily” and there’s no going back.

OpenAI head of robotics just resigned over company deal with the Pentagon saying… “Surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got”

SHOULD THE U.S BLOT OUT THE SUN AND USHER IN AN AGE OF ETERNAL DARKNESS? - Gallup 2/20/26-3/5/26 NO - 41% YES - 38% UNSURE - 21%

Say it plain: This is a terrible jobs report. The U.S. economy lost 92,000 jobs in February, nearly double the expected loss of 50,000. It’s the third payroll decline in the past five months. 📌Note: This is one of the worst monthly payroll drops since Covid.

their brutish invaders, our heroic military operationers

where attention (tokens) goes, knowledge grows

are people seeing @why.bsky.team flowcoding lately? he's post vibe now

If these trends continue, "head of meta ai" will be the most common professional role in our post-ASI future.

Seven ai orgs for the dwarven heads of AI in their halls of glass.

I’ve been notified about another death in ICE custody. This marks a stunning 40 deaths during the second Trump administration. This cannot continue. There must be real accountability for ICE and DHS.

These are four particularly good long reads on the topic of military software provisioning from the past few days. www.hyperdimensional.co/p/clawed thezvi.substack.com/p/secretary-... jessicatillipman.com/what-rights-... www.astralcodexten.com/p/all-lawful...

Ben Thompson making a full-throated case for fascism here stratechery.com/2026/anthrop...

"If this event contributed anything, it simply made the ongoing death more obvious and less deniable for me personally. I consider the events of the last week a kind of death rattle of the old republic, the outward expression of a body that has thrown in the towel." Don't skip the intro.

--dangerously skip international law

openai employees: you can, indeed, just quit.

Your regular reminder that the official a16z principle on ethics is “anti”.

Effective Altruist 2016: someone known for making choices based on extremely detailed ethical views, often rooted in obscure ideas like expanding the circle of moral concern, strategic cause selection, and efficiency on the margin Effective Altruist 2026: someone with ethical opinions (secdef def)

come on people like the existence of a Dario in this universe didn’t always strongly imply the existence of a Wario. it was right in front of us the entire time.

tired: OpenAI has lost the Mandate of Heaven wired: OpenAI has gained the gates of hell

does anyone need a used monkey’s paw? I’m all finished with my wish to vindicate effective altruism by regulating frontier AI labs

The QP is a way to use Asana inside Claude Code but I just happily stumbled on this new improvement to the connector on Claude.ai I just asked it to create a project with tasks and it did this automatically (after authorizing). I heavily leverage Asana to decompose agent work and you can too.

my persona here is presumably something like the third thing, i.e. AI practitioner, but really inside me there are two wolves

This headline really undersells a wide ranging conversation with Anthropic's Jack Clark about the future and policy/societal issues. If you listen closely, Jack points at some pretty concrete answers, while also just being candid that they aren't actually yet in the overton window.

aaaaaah updated METR study updated METR study update on the new version of the study that people wrongly cite as proof that LLMs slow developers down. they're having trouble because it's impossible to find participants willing to work without AI for the comparison lmao

"Illicitly distilled models lack necessary safeguards". - short quote from a recent blog post people haven't read but are commenting on Either you take the problems seriously or you don't. If you are pro-distilled models, are you totally ambivalent about Claude's constitution too? Ok with Grok?

Did you know that two of the top two leading AI labs use Asana to help organize their work? Sadly neither are publicly nameable.

Important reminder of the devastation wrought by the Musk/DOGE dismantling of USAID. Hundreds of thousands dead, many more to come. The greatest crime of the 21st century (so far).

As AI labs increase the pace of model release cycles, people are going to have to get used to the underlying guts of their companion bots being regularly replaced. Call it a friendship of Theseus.

does the AI safety field have an alignment solution for future, more capable models? the answer is no. but does it have a detailed and pointed plan to develop a solution? also no. but! does the research community have determination & espirit de corps?!! wait, why are you crying? come back!!

Brainwashing, 2026 edition. This paper shows how X's algorithmic feed shifts people's views rightwards. It's a sophisticated, highly effective form of reorientation. And it is utterly chilling. If you're still on that platform, unhook yourself now. www.nature.com/articles/s41...

About 15 years ago, a group of my friends built a geodesic dome from lengths of conduit. I always thought it would be easier with a model to reference, and I finally built it, just a little late.

If I have seen further it is by standing on the shoulders of Ants.

If I have seen further it is by standing on the shoulders of Opus.