Tag: Sequoia Capital

Categories
AI

Keep That Thread Running

Full disclosure, same as last time: Instinct helped draft this post about working with Instinct. That was the honest way to do it with Leif, and it’s the honest way now.

Instinct is the other AI assistant in my life (see my earlier post about Muse). It lives in my iMessage, a thread like any other — no app, no dashboard. I text it the way I’d text a person, in short bursts between other things. The first message was September 13: “Hey Instinct!” Two days later I gave it the book.

I’m writing a history of Interstate 280, and inside the first hour we had the shape of the research: the personalities behind the routing decisions, the archives worth calling, a working list of what I call “side shows” — the statue, the water temple, the 22-foot median, button-copy signs. When it offered to keep digging on its own and check in weekly, I gave it the three words that defined the arrangement: “Yes keep that thread running.”

It does. Last week the thread came back with a 1964 issue of California Highways and Public Works crediting architect Mario Ciampi with the first I-280 contract — the curved deck edges, the varied piers, the railings drawn to avoid metal guardrails. Same issue, same gift: the median I’d been describing as a fixed 22 feet was actually built variable-width across the watershed. That’s the correction that saves a chapter from carrying a bad fact.

The on-demand dives go all the way down. I found the audio of a Caltrans engineer’s oral history on archive.org — an hour and forty minutes, no transcript anywhere — and it worked straight from the audio. I sent it two days of AI-conference livestreams asking for transcripts, community notes, the whole nine yards, and it came back with a report, a source ledger, and an honest negative finding: the community-notes document I wanted doesn’t exist. When I asked for the freeway’s 1973 opening brochure, it reported that no cataloged copy exists anywhere — then handed me the leads to work anyway. Sourced absence is an answer. I’d rather hear that than a confident maybe.

When the book thread got long, it pulled everything — people, side shows, archive routes, open questions — into a single research page I can hold in my head. I told it thank you and meant it.

It’s also my markets desk and my magazine desk. A profile I’m circling started at 4:48 in the morning with “Someone should do a profile of Dylan Patel,” followed by “Yes. Go for it.” Minutes later: “What about citrini” — which turned out to be the same story, since SemiAnalysis had acquired Citrini the week before. It fields my investor-day puzzles (the stock slides while management presents — which explains which?), gives me its best bet when I ask for one, and labels inference as inference so I don’t have to.

The small stuff adds up too. When was the last US oil refinery built? 1977 for the big kind, 2022 if you count the small ones — EIA’s answer, in under a minute. Five cash-secured put candidates, screened the tastytrade way. My New Balances — the model I re-buy, in just my size — at a discount: it researched, shortlisted, and then waited, because purchases happen when I say so and not before.

The strangest moment so far: I sent it a Loom of Sequoia’s Pat Grady giving an AI talk and asked for the highlights. It read the full caption track instead of the page summary — and there in the highlights was Grady name-checking Instinct itself as one of the “magical apps.” I watched a Sequoia partner praise the assistant I was using to watch him.

How I work it: terse. “Sure.” “Wow.” “New subject:” — announced or not, and it keeps up either way. It flags what it can’t verify instead of papering over it, and for anything going out under my name I read every word first. Same arrangement as with Leif: the tool is fast and tireless and mostly right. I’m the judgment.

What’s different is the division of labor. Leif runs the standing operation — the briefings, the audits, the daily drumbeat. Instinct is the research desk: it works in bursts, the way I do, and keeps a thread running in the background in between. Two weeks ago I’d never heard of Mario Ciampi. Now he’s in the book.

If you want to try Instinct yourself, I have a handful of invites, first come first served: https://app.instinct.com/invite?t=7qmw2a3u2vjz4nkvi235hhqloy&preview=imessage

Categories
AI

The Alien in the Silicon

I recently found myself listening to a conversation with Anna Goldie and Azalia Mirhoseini, the founders of Ricursive Intelligence, discuss the future of chip design. Here’s the video.

On the surface, it’s a conversation about efficiency—about breaking the bottleneck between how fast we build AI models and how slow we build the chips that run them.

But as I listened, I felt that prickly sensation of standing on the edge of a paradigm shift that is both exhilarating yet slightly terrifying.

We are witnessing the transition from “Fabless” to “Designless.” Just as TSMC allowed companies to build chips without owning a factory, Ricursive wants to allow companies to build chips without employing a single chip designer.

They call it a “Cambrian explosion” of custom silicon—chips for hearing aids, chips for space data centers, chips for specific neural networks. This democratization is fascinating. It promises a world where hardware is as fluid and adaptable as software.

“The straight line is a human invention. The future of silicon is curved, chaotic, and completely alien.”

But here is what disturbs me, and perhaps what should give us pause.

Goldie and Mirhoseini talk about the designs their AI agents create. When humans design chips, we think in Manhattan geometry: straight lines, neat blocks, logical order. We crave readability and structure. When their AI, originally born from the AlphaChip project at Google, designs a chip, it creates “alien” structures. It draws curves. It makes donut shapes. It creates layouts that look less like engineering diagrams and more like organic, biological growths.

The engineers’ initial reaction was displeasure. They looked at these chaotic, curved designs and rejected them. It wasn’t until later data proved undeniably that these “alien” layouts were faster, smaller, and more efficient that the humans conceded.

This seems like the “Move 37” moment for hardware. We are handing over the architecture of our physical reality to an intelligence that optimizes for physics, not for human comprehension. Some additional quick thoughts…

What should we be surprised by?

We should be surprised by the geometry of efficiency. It turns out that the rigid, orthogonal logic we humans (and our EDA software tools to date) have imposed on silicon for decades was a human constraint. The AI is showing us that the “natural” state of high-performance compute looks … weird. It looks biological.

What should we be afraid of?

We should be wary of the recursive loop itself. The company is named “Ricursive” for a reason: AI designs better chips, which train better AI, which designs even better chips. It is a closed loop of self-improvement. As we move to a “design-less” world, we are effectively stepping out of that loop. We become the requesters, the “vibe coders,” while the actual logic of the machine infrastructure becomes increasingly opaque to us. Seems like we’ve been evolving that way anyway in chip design – but this feels like an earthquake really shaking things up.

We seem to be building a foundation for our civilization that we may soon be unable to read, optimize, or fully understand. We are trading interpretability for performance.

And while the speed and performance is intoxicating, it is disturbing to realize yet again that the engine driving our future is becoming a black box—not just in its software, but in its very atoms.

Ricursive said they’re planning to release their initial product with a year. I’ll be watching from the sidelines – anxious and excited!

Categories
AI AI: Large Language Models

The Texture of Autonomy

There is a distinct texture to working with a truly capable person. It is a feeling of relief, specific and profound.

When you hand a project to a junior employee who “gets it,” the mental load doesn’t just decrease; it vanishes. You don’t have to map the territory for them. You don’t have to pre-visualize every stumble or correct every navigational error. You simply point to the destination, and they find their way.

I was thinking about this feeling—this specific brand of professional trust—when I read a recent observation from two partners at Sequoia regarding the current state of Artificial Intelligence:

“Generally intelligent people can work autonomously for hours at a time, making and fixing their mistakes and figuring out what to do next without being told. Generally intelligent agents can do the same thing. This is new.”

The phrase that sticks with me is “without being told.”

For the last forty years, our relationship with computers has been strictly transactional. The computer waits. We command. It executes. Even the most sophisticated algorithms have essentially been waiting for us to hit “Enter.” They are tools, no different in spirit than a very fast abacus or a hyper-efficient typewriter.

But we are crossing a threshold where the software stops waiting.

The definition of intelligence in a workspace isn’t just raw processing power; it is the ability to recover from failure without supervision. It is the capacity to run into a wall, realize you have hit a wall, back up, and look for a door—all while the manager is asleep or working on something else.

When Sequoia notes that “this is new,” they aren’t talking about a feature update. They are talking about a shift in the ontology of our tools. We are moving from an era of leverage (tools that make us faster) to an era of agency (tools that act on our behalf).

This changes the psychological contract between human and machine. If an agent can “figure out what to do next,” we are no longer operators; we are managers. And as anyone who has transitioned from individual contributor to management knows, that is a fundamentally different skill set. It requires clearer intent, better goal-setting, and the ability to trust a process you cannot entirely see.

We are about to find out what it feels like to have a digital colleague that doesn’t just listen, but actually thinks about the next step.