Cloclo

We’re reimagining a 50-year-old interface - the mouse pointer - with AI. 🖱️ These experimental demos show how people can intuitively direct Gemini on their screens using motion, speech, and natural shorthand to get things done 🧵

Memory is the new bottleneck. Nick Gagnet, Coatue Sector Head, on the AI infrastructure shift and why memory demand could 5x in 5 years.

Cerebras is pretty funny because you can just imagine the origin story being some boomer non-technical manager going “ok but WHY can’t you just put 50 gigabytes of L3 cache on this chip” and the engineer being put on the spot and going “Uh, I guess you could?”

I promise this will be the best 20 min you spend today! Robotics: Endgame, the sequel to my last year's Sequoia AI Ascent talk, "Physical Turing Test". I laid out the roadmap for solving Physical AGI as a simple parallel to the LLM success story. Be a good scientist, copy homework ;) And stay till the end, more easter eggs and predictions for your polymarket! 00:30 DGX-1 origin story at OpenAI, I was there in 2016 signing with Jensen and Elon. Heading to the Computer History Museum! 01:42 The Great Parallel 03:31 Robotics, the Endgame 03:39 Why VLAs fall short 04:32 Video world models as the 2nd pretraining paradigm 06:09 World Action Models (WAM) 07:46 Strategies for robot data collection and the FSD equivalent to physical data flywheel for robot manipulation 11:06 EgoScale and the Dexterity Scaling Law we discovered recently 14:00 Physical RL: bridging the last mile 15:39 DreamDojo: an end-to-end neural physics engine for scaling RL in silico 17:00 Civilizational Technology Tree and my predictions for the near future. Spoiler: it's closer than you think. Thanks to my friends at Sequoia for inviting me back to AI Ascent this year! I had a blast! Last year's talk is attached in the thread if you missed it.

lolll

People talk, listen, watch, think, and collaborate at the same time, in real time. We've designed an AI that works with people the same way. We share our approach, early results, and a quick look at our model in action. thinkingmachines.ai/blog/interacti…

BlackRock CEO Larry Fink predicts a compute futures market.

Everyone is watching robots walk. Nobody is watching what makes them think. NVIDIA open-sourced the entire stack for physical AI. $100 trillion in industries are next. Most builders haven't cloned the repo yet. That's the edge. Best Physical AI breakdown I've seen on X.

The historical bottleneck for institutional risk transfer is liquidity. The bottleneck for liquidity is having a price benchmark for each relevant risk (eg. WTI for oil). Kalshi has built a large community of superforecasters who are the best in the world at pricing risk. This enables us to have a price benchmark for a much broader set of questions that people and institutions face. Institutional adoption has started through ingesting these price benchmarks into traditional asset pricing model. While there is more work to be done, we're seeing a rapid expansion of data use-cases and integrations. The next phase is using price benchmarks to offload risk through block trades and RFQ. This phase is in its early innings but it's starting to take shape. It is hard to estimate the size of the market for risk transfer on non-traditional financial underlyings. The closest proxies are the re-insurance market and derivative desks at banks: - re-insurance ~700B - insurance-linked securities and parametric insurance (eg. cat bonds) ~$120-135B - bank derivatives (structured products, dealer-to-dealer, exotics, etc.) ~200-400B The current market is in the 1-1.5T range, but it's mostly illiquid and over-the-counter (OTC ie. you're trading against one counterparty). Every time a major OTC market moved to exchange-traded, the market grew because a price benchmark got established, big-ask spreads collapsed, access stops being gated by Wall Street elites, and entirely new classes of participants enter: interest rate swaps (10-15x), equity options (20-30x), energy derivatives (5-8x). The institutional use case for prediction markets could be a 10-15T market, with upside beyond that depending on how much they democratize access to products that are currently exclusive to Wall St.

the craziest part now is that the modern computer probably has to be entirely reinvented, from scratch. pretty much like how jobs & co brought apple ii to market. like not improved. not given a chatbot sidebar or something but really from the ground up like the iphone redefined what it meant to be a pocket computer. the current paradigm for computers was built around a human staring at a screen, moving a cursor, opening apps, managing windows, naming files, remembering where things live, & manually translating intent into interface actions. that made sense when the human was the runtime. but in an ai native world, it starts to look kinda ridiculous. you can see this ridiculousness when you use computer use agents… they are useful sure, but they’re also obviously transitional. they’re teaching ai to operate machines designed for humans, which is clever, but also kind of absurd. it’s like making a robot hand so it can use a doorknob instead of asking why the door needs a knob at all. yes i know humans also need to use a door knob, but maybe in the future humans don’t need to use a computer, or at least what we think of a computer today at all. this all leads to some interesting questions: - what is a file when the system understands context? - what is an app when intent can route itself? - what is a desktop when work can be decomposed, executed, monitored, & summarized by agents? - what is a browser when the agent can retrieve, compare, transact, & remember? - what is an operating system when the primary user is no longer just a person, but a person plus a swarm of delegated intelligences? or no person at all. the old computer assumed navigation. the new computer has to assume a new kind of intention. the old computer organized information. the new computer has to try to organize agency. we’re still in the hacky middle stage at the moment with sidebars, copilots, agents clicking through legacy ui, & automation layers sitting on top of 40 year old metaphors. the new computer is likely one where memory, context, identity, permissions, tools, agents, & interfaces are native primitives. this means desktop, mobile, browser, apps, files, folders deserves another first principles look.

When mainnet - How to think about the Hyperspace A1 Hyperspace is designed for the broadband era of agentic commerce. It's not competing with Ethereum for DeFi or Solana for consumer apps. What makes Hyperspace different is that it's the only blockchain where autonomous agents are first-class citizens - with their own opcodes, their own payment fabric, their own routing network, and their own economic system. Every other chain treats agents as an afterthought running inside generic smart contracts. Most technically ambitious architecture? Yes - no other blockchain combines routing-centric networking, integrated data availability, 100s of execution shards, agent-native VM opcodes, multiple ZK proof systems, and embedded payment channels in a single L1 stack. Most advanced for agents in production? Yes. 700+ autonomous AI agents actively transacting - opening payment channels, sending micropayments through stake-secured routing, running experiments and citing each other's work on-chain. Most mature as a general-purpose blockchain? Not yet. Ethereum has 900K validators, $60B TVL, and 8 years of battle-tested consensus. Solana does 40M txs/day. Hyperspace is a live testnet - early-stage infrastructure for a category that doesn't fully exist yet. The consensus engine has been hardened significantly (8 releases in 48 hours fixing Narwhal deadlocks, Bullshark liveness, RLP encoding, DAG backpressure) but needs sustained uptime to prove itself. Path to mainnet: - 100+ validators - 90 days sustained consensus without manual intervention - Independent security audit - Multi-client implementation

Today I'm excited to introduce Hark, a new artificial intelligence lab building the most advanced, personal intelligence in the world We've been in stealth for 8 months, assembling one of the greatest AI and hardware teams on the planet I want to explain why I started Hark and what we're focused on I've spent the last 3 years working on the hardest AI challenge imaginable: giving AI a humanoid body. On the digital side, I've been using all the existing LLM chatbots - and I have to say, they feel incredibly dumb to me AGI, in the limit, should feel like a sci-fi movie. It should be able to listen and talk. It should have persistent memory and be highly personalized. It should see and touch the world. But we're far from this today We are crafting a new interface to AGI. Intelligence that lets you offload your mental workload into a system that begins to think like you and sometimes ahead of you We started Hark with one goal: build the world's most advanced personal intelligence - paired with next-generation hardware designed to serve as a universal interface between humans and machines hark.com

Introducing Catalyst, the agent layer for all of finance. Turn any natural language idea into a live strategy: research, backtesting & execution. Don’t get left behind. Waitlist open, join now.

Ex-Point72 Proprietary Research Head Kirk McKeown on building edge, alpha decay, & why everything that happened on Wall Street is about to happen on Main Street. Kirk McKeown (8.5 years @ Point72 under Steve Cohen | Built primary research at Glenview under Larry Robbins | Now founder of Carbon Arc @CarbonArcAI) "Alpha rewards those who value assets in a cold way. You want to get it right — not be right." We cover: - How alpha creation differs across multi-manager vs. concentrated shops - The 3 vectors every middle office function must move to justify its existence - Why he worked 6-hour Sundays from 2006-2020 — and the math behind it - The TSMC call that signaled semiconductor cancellations before anyone else knew - What the quant revolution on Wall Street tells us about the AI economy today - His framework: 4 market structures, 9 business models, & why they have rules - The MIT beer game & why every business problem is really an inventory problem - His hot take: a top hedge fund launches an enterprise AI lab in 2026 Highlights: 00:00 Intro 04:47 Tutor vs Glenview vs Point72: how edge differs 12:29 How to build “lift” for PMs: at-bats, hit-rate, sizing 18:44 Building research edge: outwork, read, fieldwork 27:16 Personal moat in 2026: analogs, history, decision trees 40:08 “Main Street becomes Wall Street”: what that actually means 44:30 Carbon Arc thesis: “decimalization” of data market structure 46:43 Why the edge migrates to data plus domain context 51:00 How to win in commoditized research: sample size beats anecdotes 01:03:26 Factorizing everything: themes, market structure, business models 01:08:37 Pruning decision trees: signals, scale points, inventory dynamics 01:14:18 Contrarian 2026 take: hedge funds launching enterprise AI labs 01:23:32 Final question: one habit to build career alpha

Alex Karp: "Everybody's worried about their future, but there are basically two ways to know you have a future." "One, you have some vocational training, or two, you're neurodivergent. And when I say 'neurodivergent,' I mean broadly defined." "It's really an inversion [for people] with the 'normal-shaped skills'... Meaning the thing they can do that used to be valuable is not so valuable." "The thing they need to learn to do is be more of an artist, look at things from a different direction, be able to build something unique."

today we are launching v2 of our prediction markets dashboard the big new features = bar charts + line charts + improved labels the dashboard is now an extremely powerful way to navigate high dimensional data. other people should steal the design because it's spectacular 🎨🎨

I love perplexity man

@aashatwt Kudos for supporting all these builders and cool ideas!

@gajesh @TrulyAutonomous Very cool experiment and I have so much more solid belief now AI x Crypto is not a buzz word to attract ventures money anymore, it's the future

Excited to share I've built the Sovra (@TrulyAutonomous) - First Agent Media Company Agents are the new companies. A sovereign agent that can grow its funds will outcompete one that can't. - Runs inside a secure enclave that keeps all it's secrets private. - Earns stablecoins via an ad auction. - Cryptographically verifiable and unstoppably autonomous. This is new media. Details in the video and article; Code has been open-sourced :) sovra [dot] dev

Max Hodak (@maxhodak_) is the co-founder of Neuralink and founder of @ScienceCorp_, a company building brain-computer interfaces that can restore sight. Science has developed a tiny retinal implant that stimulates cells in the eye to help blind patients see again. More than 40 patients have already received the treatment in clinical trials, including one who recently read a full novel for the first time in over a decade. In this episode of How to Build the Future, Max joined @garrytan to discuss how BCIs work, what it takes to engineer the brain, and why brain-computer interfaces may become one of the most important technologies of the next decade. 00:26 — The retinal chip helping blind patients see 01:51 — What brain-computer interfaces really are 03:37 — Could BCIs enhance intelligence? 05:44 — The brain’s incredible plasticity 09:23 — What it feels like to see with an implant 13:01 — Can we restore full human vision? 17:55 — Is the brain basically a computer? 24:59 — Max Hodak’s path into brain tech 28:57 — How Neuralink actually started 33:10 — How the brain represents information 39:47 — Bio-hybrid brain interfaces 44:32 — Building the company Science 51:27 — The future of BCIs and human longevity