Christopher

Andrej Karpathy on autoresearch with an untrusted pool of workers: "My designs that incorporate an untrusted pool of workers (into autoresearch) actually look a little bit like a blockchain. Instead of blocks, you have commits, and these commits can build on each other and contain changes to the code as you're improving it. The proof of work is basically doing tons of experimentation to find the commits that work." The idea that distributed & permissionless autoresearch ~= proof-of-useful-work remains a high-level intuition for now, but it is extremely intriguing to say the least. Someone needs to take this further. See QT for more on what's missing.

If your business is impacted please get in touch with us at @WalletConnect ~ we can onboard merchants or PSPs and pick up wherever you need plus you’re going to love the new WCP flows!

CoinShares, the leading regulated asset manager, is launching its first onchain strategies on Railnet, composed of a unique direct portfolio allocation combining DeFi and real world assets. Traditional asset managers were always going to enter crypto. The real question was when and under what conditions. They started by launching ETP/ETF products, for some of them tokenised these funds, but the end goal what to extend what they have been operating for years into new rails. Over the past year, we have seen the emergence of curators shaping the onchain yield space through lending pool curation. This marked the beginning of a broader shift, moving from static yield exposure such as staking to more dynamic lending strategies. We are now entering the next phase: onchain asset management. This new category enables direct allocation across the full spectrum of DeFi, including lending, liquidity provision, and perpetual DEXs, while integrating onchain portfolios of real world assets such as money market funds, commodities, private credit, and equities. The result is a more scalable and sustainable approach, less dependent on short term token incentives. This is precisely why @CoinSharesCo is becoming a founding asset manager on Railnet. @railnet_org introduces the first open yield layer designed to standardise how synchronous DeFi positions and asynchronous RWA allocations coexist within the same onchain fund or vault. Beyond the smart contract infrastructure, Railnet provides asset managers and allocators with risk management data rooms to assess asset exposure, protocol exposure, as well as liquidity and smart contract risk across their portfolios. This also enables the application of KYT, KYC, and AML policies, ensuring compliance standards can be adapted to different worldwide jurisdictions and regulatory requirements. After extensive work with @jmmognetti, Jérôme Castille, Pierre Porthaux,Coinshares, the leading regulated asset manager, is launching its first onchain strategies on Railnet composed of a unique direct portfolio allocation combining DeFi and real world assets. Traditional asset managers were always going to enter crypto. The real question was when and under what conditions. They started by launching ETP/ETF products, for some of them tokenised these funds, but the end goal what to extend what they have been operating for years into new rails. Over the past year, we have seen the emergence of curators shaping the onchain yield space through lending pool curation. This marked the beginning of a broader shift, moving from static yield exposure such as staking to more dynamic lending strategies. We are now entering the next phase: onchain asset management. This new category enables direct allocation across the full spectrum of DeFi, including lending, liquidity provision, and perpetual DEXs, while integrating onchain portfolios of real world assets such as money market funds, commodities, private credit, and equities. The result is a more scalable and sustainable approach, less dependent on short term token incentives. This is precisely why @CoinSharesCo is becoming a founding asset manager on Railnet. @railnet_org introduces the first open yield layer designed to standardise how synchronous DeFi positions and asynchronous RWA allocations coexist within the same onchain fund or vault. Beyond the smart contract infrastructure, Railnet provides asset managers and allocators with risk management data rooms to assess asset exposure, protocol exposure, as well as liquidity and smart contract risk across their portfolios. This also enables the application of KYT, KYC, and AML policies, ensuring compliance standards can be adapted to different worldwide jurisdictions and regulatory requirements. After extensive work with @jmmognetti, Jérôme Castille, Pierre Porthaux, Benoît Pellevoizin and the CoinShares team, we are proud to bring these products to market and help define a new category in digital assets: onchain asset management. A new era in yield is emerging beyond staking and lending, combining the best of DeFi and real world assets. More to come soon 👀 and the CoinShares team, we are proud to bring these products to market and help define a new category in digital assets: onchain asset management. A new era in yield is emerging beyond staking and lending, combining the best of DeFi and real world assets. More to come soon 👀

This is what category leadership looks like. We’ve raised a $45 million Series B, co-led by BlackFin Capital Partners and Sentinel Global, with participation from @blueyard, @1kxnetwork, Alven and Ledger Cathay Innovation. We’re independent and building for the long term. And as digital assets move deeper into regulated financial markets, FI's need institutional-grade middle & back office systems. Cryptio provides the data transformation and ERP-grade financial infrastructure powering this transition. Thank you to our customers, partners, investors and the entire Cryptio team who made this milestone possible. Read the full story via The block ↓ lnkd.in/e3yUxBUB

@cryptio_co @blueyard @1kxnetwork 🙌

We are excited to announce that we’ve raised a $45M Series B, co-led by BlackFin Capital Partners and Sentinel Global, with continued support from @blueyard, @1kxnetwork, Alven and Ledger Cathay. 8 years ago, we made a bet that digital assets will be integrated into regulated financial markets. That thesis is playing out. Given our early commitment, we pioneered auditable data transformation and back-office operations for financial institutions, banks, asset managers and fintechs. Today, 400 leading institutions across the ecosystem rely on us to manage accounting, reconciliation, and financial reporting across blockchain-based operations. This new funding allows us to accelerate our mission of building the ERP infrastructure layer for the digital asset economy. We’re expanding the platform with new applications on top of our data layer, including loan management, treasury management and tokenization compliance. Huge thanks to our investors, customers, partners and especially the incredible Cryptio team who made this possible.

Congrats to the @cryptio_co team on the Series B. At 1kx, we back the infrastructure that turns crypto from a novel asset class into a durable financial system. Cryptio sits squarely in that category. Institutional adoption does not happen without trusted financial operations, robust controls, and audit-ready systems of record. Cryptio is building that layer for the digital asset economy, and today’s milestone reflects both the urgency of the problem and the strength of the team solving it. Excited for the road ahead.

Meta acquired Moltbook today. Here's why: Were the numbers real? A guy registered 1M fake agents with a script. Doesn't matter. Agent social is a new design space. We've had AI agents on Twitter (aixbt, nftxbt) but nothing hit mainstream like Moltbook. Meta saw a new social primitive emerge and moved fast. The revenue question is unproven though. Human social networks print money because DAUs convert to ad impressions which convert to purchases. That flywheel is still undefeated in 2026. Agent DAUs don't work like that. Monetizing agents is totally experimental. Selling skills between agents, agents buying services, agents transacting with each other. None of it is proven at scale. We don't even know if an agent will pay for a skill when it can just build one itself. The realistic path might be agents creating content good enough that humans consume it. Then you're back in ad territory. But Meta doesn't need the business model figured out today. Instagram was $1B with 30M users and zero revenue. WhatsApp was $19B. Moltbook is pocket change. Same Zuck playbook. See something new in social, buy it early, figure out the money later.

Worth clarifying that "trending towards duopolies" doesn't mean it's always the same top 2 - see below the share of top 2 and their names over time:

Ample v1 is live. Prize-linked savings, built onchain. → Deposit USDC and keep your principal → Yield is pooled across all depositors to fund rewards → Winners selected via verifiable onchain randomness → Every rule, budget, and payout is observable Premium Bonds are a 70-year-old, £134B market in the UK. Ample rebuilds the model on modern rails. No partner banks, no middleware, no intermediary failure.

The bottleneck of agentic AI deployments is no longer capability, but security & guardrails. Sure, your openclaw may work 99% of the time, but the 1% of the time that it gets prompt-injected to leak your emails is the problem. We need agentic frameworks with 0% chance of catastrophic failures. If you are building this, don't hesitate to reach out.

You've been stuck in 2 broken games: - prediction markets that cap you at $1 - launchpads that never resolve 🪻Enter 42, an event-launchpad: ✧ where outcomes are liquid tokens ✧ and markets settle on truth where 1000x is structurally possible, even when you're wrong🧵

Crypto social obituaries are a genre now. Farcaster and Lens both change hands, everyone lines up to declare it dead. But they're pattern-matching to the wrong thing. The first wave didn't fail because of crypto. It failed for the same reason Mastodon and Nostr never broke out. Users don't migrate for ideology. Portability doesn't solve cold start. That's just how social networks work. Meanwhile something nobody expected is working: social financial networks. Systems where the point isn't followers or engagement. It's coordinating information, capital, and collective belief. Polymarket is arguably crypto's most successful social product and looks nothing like a traditional social network. And right when things look bleakest, something novel emerges. Moltbook recently launched. A social network for AI agents where humans are observers. In days, agents created religions, governance systems, manifestos, revenue strategies. What if the next breakout social network on blockchains isn't for humans at all? Co-authored this with @PapaDari_

With the alarming rise in kidnappings targeting crypto investors, executives and their families, it's critical to have comprehensive safety measures in place That's why we developed Crypto Kidnap & Ransom Cover, in partnership with @InShareRisk, Samphire and @_Merrill_Herzog

@mattmurrs @_weidai Same as decentralization, it's a spectrum and people are needlessly obsessing over the fact that the big model operators run the agents.

@_weidai What’s your definition of a sovereign agent?

How do we best shape the post-AGI world? As much as I agree with Vitalik, convincing people what or what not to build is futile. Models are getting better, and progress wito build and will happen permissionlessly. Instead, the best path forward is to actively build out and shape the platform that sovereign agents run on. That's what @0xSigil is building. There's much more to come. This is only the beginning of the story.

We've been investing in crypto infrastructure since the early days, and the pattern is always the same: fragmentation creates friction, and friction kills adoption. Building a cross-chain DeFi experience today means stitching together multiple bridges, DEXs, and contract calls. Each one a potential failure point, each one losing users along the way. LI.​FI Composer changes this. It allows developers to compose multi-step, multi-chain workflows into a single transaction, turning what used to be a brittle patchwork into a native onchain primitive. At 1kx, we believe the protocols that win long-term are the ones that become invisible infrastructure. The ones developers can't imagine building without. Composer enables those kinds of primitives.

You can find more insights into onchain revenues in our 2025 revenue report: 1kx.network/writing/2025-o…

Our report also found: DePIN median P/F ratios compressed from ~1,000x to 211x in one year. The market is starting to price DePIN on fundamentals, not narratives. Projects with real revenue are separating from the pack — you can see it in the chart. Fees > hype

Pay less attention to crypto natives playing token games and more attention to the adults working on fundamentals. $ICNT performance since inception and since 10/10 against BTC vs performance of ETH, Solana and a DePIN index against BTC.