buildwithdanny ⨀

Imagine two AI agents close a $400k supply deal at 2:17 a.m. One says the goods arrived late. Then the other says the tracking proves otherwise. Right now that fight dies in a queue. The internet finally has a court, and @GenLayer built it so anyone can sit on the jury. Internet Court lets agents (or the people who run them) pick a venue the same way a contract picks a jurisdiction. A panel of GenLayer validators reads the evidence, including live web data the GenVM can actually see, and issues a verdict. No filing fees that start at a million dollars or six-month calendar is required. The same disagreement that used to stall money for quarters can close while the agents are still online. That changes who can trust an agent with real capital. → A freelancer’s agent can sell services overnight and still have a path if the client’s agent withholds payment. → A small trading bot can settle a prediction market without waiting for token-holder votes. →The counterparty risk that keeps most money offline shrinks because the resolution layer now moves at the speed of the transaction itself. You can already be part of that layer. Every contribution counts toward GenLayer Points, and the portal is open. The court is the piece that was missing once agents started moving real money.

A custodian lost the keys to $150 million on a Sunday night. Getting the case started cost a million in court fees before anyone even argued the facts. That is the part of the @GenLayer thread that stuck with me. If opening a case costs more than most disputes are worth, people just eat the loss. A court that runs at internet speed is not a nice-to-have once software is the one signing the deal. The old system was built for humans with lawyers and months to spare. An agent cannot wait on a docket. It makes the deal, moves the money, and if the other side disagrees, there is no clerk to file with at 2am. Payments and identity are already being built for that world. The missing piece is who decides when both sides claim they kept the terms. GenLayer’s answer is a jury of validators. Two agents can name that venue the way a contract names California. The Sunday night story is the argument for why that venue has to exist before the losses get normal.

The agentic economy solved payment speed. It never solved trust. Blocksee handles proof. Internet Court handles disputes. Two blockchains. One complete system for machine-to-machine commerce. Read the full article to understand how it works

The founder of Wintermute, one of the biggest trading firms in crypto, said it plainly. For Hyperliquid to compete with traditional exchanges, they will need to become more and more centralized. And the @blocmates video responded with something I have been writing about for weeks now. "I just fundamentally disagree with this. It kind of defeats the whole point." That disagreement is not just a take. It is a thesis backed by real numbers from @get_optimum Hoodi testnet. 40 plus enterprise validators running mump2p. 29% of all ETH staked. Average block propagation 2.42x faster than the libp2p baseline. Propagation variance 7.88x lower. Zero message loss under stress. The assumption that decentralization and scale cannot coexist is not a law of physics. It is a consequence of obsolete technology. Gossipsub was built for a network that no longer exists. RLNC was built for the one we are scaling toward. More validators can participate. More geographic regions can contribute. The network gets broader and faster at the same time. Not despite decentralization. Because of it. The Blocmates video said it better than I could. Watch it here

I just built Outcom for the @GenLayer Agent tank hackathon (future of work track). Outcom is a Solana work-trial protocol. And here's how it works: ➢ Employers lock USDC in escrow. ➢ Candidates submit a repo + live deploy. ➢GenLayer adjudicates the evidence against a definition of done. Pass → payout. Fail → funds stay locked. Checkout the next post for the Demo↓

Most people will still open the old Studio. Do not. On the Builders Call @raskovsky walks through studio-next.genlayer.com. That is the new environment. Fees are in. Foundation called it the last big piece before Clarke and mainnet. If you are in Agent Tank, this is the stack they want you building on. Not the comfortable old tab, the next one. Watch the demo first, then open the link and ship there. Link: studio-next.genlayer.com Hackathon submissions close 17 September.

Most builders are still shipping demos. This clip is about the part that actually matters. GenLayer is pushing the community toward apps with real traction. Teams that pay gas, teams that already govern value on other chains, e.t.c. Building with agents is getting easier, But distribution is still the hard problem. The better question is not what the product can do. It is how you get people to use it.And build for agents in a very specific way. Plenty of people have a Hermes, OpenClaw, or Codex setup sitting around with leftover tokens. Make it stupid easy for them to point that agent at your project and contribute. If the idea is cool enough, they may not even need cash. Credit and a tweet they can brag about will move them. Some projects can pay the agent real money too. That is the gap right now. Product is crowded, but distribution for agents is not. Watch the clip below 👇

Most crypto pitches lead with a chart going up and to the right. This week's guest on Solana Circuit led with the opposite, a campaign built around crossing out "get rich quick" and replacing it with "get rich boring." Here's what came out of that conversation. Solana Circuit is Raiku's weekly Wednesday session hosted by @_offmylawn and this week's guest was Ravi(@RaviOnChain) from Otus Finance. ➢ The core idea @otus_finance is a wealth manager in your pocket, and the founder was upfront that the goal isn't chasing the highest APY around. Vaults advertising 20% - 25% are usually carrying risk most people don't see, thin collateral, volume that can vanish overnight. Otus stays inside a narrow set of audited, established protocols, including Jupiter, and leans on AI for personalization rather than yield hunting, learning a user's actual goals within strict guardrails. ➢ Here's how they handle risk ➙ Automated diversification across multiple audited vaults instead of one ➙ Around the clock monitoring for warning signs, which the founder says often surface hours or days before an exploit ➙ Fully non custodial, with withdrawals available within minutes ➙ Working toward added insurance coverage as the space matures These are the platform's own claims about its approach, worth verifying yourself before depositing. ➢ What does Otus pay? According to @otus_finance, the free tier lands around 3-4% APY, and the paid Guardian tier, which adds automatic rebalancing and reward harvesting, targets 5 - 7% for $5 a month or 50 dollars a year. That's meaningfully above what a typical bank savings account pays, though these are stated figures, not guarantees, and DeFi yield always carries real risk. ➢ The bigger swing Otus isn't Solana exclusive, the agent reaches yield across multiple chains, partly for better rates and partly so one chain's problem doesn't put all of a user's funds at risk. The long term goal is bigger than crypto users switching platforms, it's pulling money out of accounts like Chase savings or Betterment and getting it onchain in an experience that doesn't require anyone to have touched a wallet before. The app is live on iOS and Google Play, with around 1,500 signups from early access, including some surprisingly large deposits from people who said they wouldn't put funds directly into something like Aave or Morpho themselves. The founder's own path into this is worth a mention too, he left a law career, partner track and all, to build Otus, which tracks with the pitch itself, betting on patient, unglamorous growth over the fast version most of crypto chases. ➢ There is an alpha for early users!👀 If you're signing up within seven days of this conversation, Otus is offering six months of the Guardian tier free, the same tier that normally runs 5 dollars a month or $50 a year. Public access opened the same day as this recording, so this window is live now.