hari raghavan

This post absolutely nails a concept I've been noodling on for a while. A major dimension that separates the left and right is the "O" on the Big 5 personality traits: liberals tend to be high on openness to new experiences, conservatives tend to be low. This means that creativity, variety, art thrive in left leaning areas. Visiting SF, NYC, Toronto, LA... Damn, the food and drink is incredible. There's also just so much to do, experiences to take in. It's a real vibe. The cities, however, are almost unlivable day to day. On the other hand: Miami has middling food, (the 99th percentile restaurants in Miami are like 70th percentile in NYC / 80th in SF), no great theater, just-okay coffee, and virtually no great ethnic food (besides Latin American, for obvious reasons). There is like one good Thai restaurant, one decent Indian restaurant. But boy is it a joy to live here. The streets are pristine. City inspectors during a remodel can be scheduled next day, not in 2 weeks. I've accidentally left my garage door open overnight. Nothing missing. In SF I had thieves sneak into a secured garage, use bolt cutters to on a bike cage, and a saw to cut off my bike lock to steal it. There is no state income tax, and yet: it draws far less from federal coffers per capita, because it spends far less per capita and... It all just kind of works. It works FAR better than it ever felt in California. Yeah. I think the left is better at beauty because they are more open to new experiences, and the right is better at order and governance because they are not. Separately: I can't believe Spencer Pratt -- this poet, this philosopher, this absolute king -- didn't win the LA mayoral election. I can't believe he didn't even make the runoff. What an indictment of direct democracy. What an indictment of the LA electorate.

Y'all should follow Clara! She knows private markets as well as anyone out there.

Hyperscalers should simply ration usage and latency (including social media) by distance to the closest data center. I bet that would turn the tide on data center buildout amongst politicians and the electorate real quick.

This very-incorrect premise (Notion being a Bending Spoons target, or at all comparable to Airtable) surfaces from time to time. I don't think @NotionHQ should engage with it (it's a leading question and even the comparison to Airtable does Notion a huge disservice), but as an early investor and big Notion bull, I'm happy to do so. This is not based on any insider info / not endorsed by the company but I know them pretty well outside-in. (I have a 3,000 article on Notion I wrote a couple of months ago that I need to finalize and publish, but in the meantime this is a brief TL;DR.) 1. Notion is growing far, far faster. Airtable was at 15-20% towards the end. Notion's is not publicly disclosed but all signs point to 70%+. I'm pretty sure they're approaching $1B in ARR. 2. Airtable's attempts at building AI into the core product mostly all sucked. Like Omni was just *really* bad. The first thing they had that sort of worked (maybe?) was Hyperagent, which was a totally different product which didn't benefit from Airtable, or pour benefit back into Airtable. Notion AI on the other hand is fundamentally woven into Notion and actually completely changes the way Notion can be used (this is part of the longer post, but some examples below). 3. Notion has completely rearchitected the product for primitives that are AI native. MCP, CLI, API — basically every bit of Notion can be used from inside Claude / Codex etc. Airtable never got there. 4. UI is really important because most people aren't systems thinkers in their head. They're visual thinkers. I would argue UI is *more* important in an agentic era, not less. 5. Airtable's main building block was a user-friendly database... it's unlucky for them but this was the least-defensible thing in the era of agents + Supabase etc. Notion has the building blocks needed for the context graph (company + personal) — dozens of primitives like a markdown editor (deceptively hard to build well), databases, tables, html blocks, gantts, charts, dashboards.... these would make no sense for each company to build for itself. Sure you can go use an open source thing or string together a bunch of components but this is really really annoying and <<1% of people want to build or maintain this (speaking as one of those people... I've built all of these components before for startups I founded). 6. Notion is leaning into this — the context layer / store of knowledge & skills for your agents — if you look at their announcements and new releases. TBH the main problem with Notion (as a company / product) is storytelling. They don't explain how you can use it well — to be fair, this is hard since there's so much there. But nevertheless, this is the company's responsibility to educate users. I consider myself a power user and I still feel like I'm barely scratching the surface. 7. The best products — even if they're hard to grok and use — will win in the agentic era, because agents are smart enough to understand the best tool for the need and adopt it. The importance of educating users might actually decline. This would be the best thing to happen to Notion, because a) they have an incredible product, b) their weakness has been educating users so that may gradually becoming irrelevant, and c) it's agent friendly. 8. On usage & adoption: 70% of Notion's user base is outside the US (a post by Ivan just yesterday). It has a ton of penetration and those who love it really love it (like me!). AI usage / MCP usage is rapidly ramping — 9x and 60x in the last year (again, see the post by Ivan yesterday). So I would argue all of these users are really early in the curve. 9. Look at the talent they're attracting. There have been many, many talented ex-founders who have joined the company in the last 6-12 mos. These folks have their pick — they could start a new company, or join one of the model labs, etc. They would not join a dying company! And they're choosing to join Notion. 10. Finally: Notion did a tender in Dec 2025 at $11B — this was the price investors paid for common stock. This was GIC, Sequoia, and Index, so about as blue chip as it gets. They don't raise a bunch of additional capital in tranched rounds because they've been profitable for a long time and they don't need to.

I understand why Zitron is the way he is, now. He's British, and the accent always sounds smart (to non-British people). He pairs this with being the purest form of midwit: someone who speaks with a great deal of confidence, and understands slightly more than the median person. Like a 103 IQ individual. But certainly not smart enough to understand second-/third-order effects (which usually kicks in around one standard deviation higher). So he's had nearly 20 years in the US of being told "you're really smart" (I mean sure, compared to a teapot) and "you see things others don't see" (so do schizophrenics). This leads him to be unerringly incorrect, an arrogant asshole, and utterly oblivious to those things. All at the same time, an impressive feat! — which is probably the only impressive thing about him. (I hesitate before being mean to people on the internet, but a) he's a much more mean to others, b) I tried to engage in good faith and he resorted to name calling and blocked me, c) "public" figures like him are fair game.)

I've been waiting for Stripe to launch a closed-loop payments network for almost a decade. But then I figured the dependence and platform risk with V/MC was too strong. Curious what changed. I don't think it's stablecoins or agentic payments, per se. Maybe it was simply time.

I think the thing that made it *really* fuzzy is that the "is it recurring" question re: Annual Recurring Revenue doesn't always have a binary answer. There are lots of "reoccurring" revenue companies that effectively look like recurring revenue. Stripe is a very good example. If it's SO like-recurring, shouldn't they get "credit" for it in the business model description? I could see a good argument for that. So that's one line crossed. But then the thing that's changed is not usage-based pricing, but rather the weakness of the moats. One could argue that OpenAI / Anthropic *would* have had effectively-reoccurring revenue (even absent enterprise contracts) but there are no switching costs. Another line crossed. And then we get to things like usage-based SaaS reselling tokens at like 10% (or even negative!) margins. Ok well that's not that different from Ant/OAI. Another line crossed. And then when you finally get to data brokers... it's basically just GMV, but there have been too many lines crossed and nobody knows what the fuck ARR means anymore. So here we are.

Sometimes entire industry norms get set in moments like this: by the right person calling out an apparent-to-insiders truth at the right time.

This is utterly off. Ramp was a hot seed stage company out the gate. They raised capital at $40-100M within a few months of incorporation. At $100M, a 1-2% grant is $250-500K / year in preferred-equivalent terms. Early stage startups simply do not grant that. It's off by more than half an order of magnitude. This is why benchmarks for early stage comp -- especially equity comp, when stated in percentages -- are worse than useless. They are actively harmful, and everyone (founders, people ops, candidates) would do better to completely ignore them. Carta and Pave are useful for base comp, but they are doing the world a grave disservice by pretending to "benchmark" equity (at early stage through Series C/D, it's objectively incorrect). Source: I've spent easily 10,000+ hours working on comp spanning founding through Fortune 500 companies, dating back to 2013. I can't wait to publish a new version of my open source comp tool. (It'll be free, just a public benefit.)

Now that The Bear is done, I think The Diplomat is the best all-round show on TV. I've seen many of the other acclaimed shows currently running: Slow Horses, White Lotus, Severance, Foundation... Very solid shows. But nothing nearly as riveting or well-done as The Diplomat.

Everyone then: oh elon, you're always saying such crazy things Everyone now: WTF

Someone in a WhatsApp chat: I have a $5M block of Anthropic. Me: "I thought we were discussing Ant secondary blocks, not secondary blocks for ants. It's gotta be at least... three times as big!"

It's really unfortunate Elon is mostly untouchable by scandals at this point because "elongate" is just sitting there unused.

Been a customer of ampersand for a while. Great product!

Objectively funny.

"talking to a machine should feel as natural as talking to a friend or coworker" uh, no.

SO STOKED FOR THIS

This is probably how @OpenAI should monetize their soon to be released personal assistant, O. They've already laid the foundations with ads.