Happy New Year 🎉 Welcome to 2026 - the Year of Execution. This year, I’m choosing action over endless planning. Less noise, more building. The lessons from last year brought clarity, and now it’s time to apply them properly. I’m approaching 2026 with focus, intention, and a lot of passion for building things that actually matter, With just consistent execution and learning along the way. Let’s build with passion this year, support each other, and let the work speak. Wishing us all a strong, purposeful 2026 🤝

Final day: PLAY’s $230,000 target is backed by 200 machines that were already earning before this went onchain. I think crypto should pay closer attention when the business activity comes first. PLAY starts with 200 claw machines already operating in real arcades, where people were already playing and generating revenue before an onchain structure entered the picture. That timing is the strongest fact here. This isn’t a plan to buy machines later and hope usage appears, or a yield loop waiting for new depositors to fund it. The machines were earning first, and PLAY brings that existing machine revenue onchain through Solana. Today is the last day of the window, so the numbers are clear. You can join through UPTIME via @DualMintRWA and @stardotfun, with the target set at $230,000. → 200 operating claw machines were already earning in real arcades before PLAY moved revenue onchain → $230,000 is the deposit target for the final day of the window → 12 to 15 percent annual yield is targeted, and it isn’t guaranteed → Revenue comes from people playing the machines, with distributions paid monthly My view is firm: the pre-existing revenue makes PLAY stand apart from products that put incentives first and search for a business afterward. I like that the central proof is simple, 200 machines were earning before the chain was involved. PLAY is onchain because the machines were already working offchain.

I’ve been trading NEAR through @RexAssBot, and the web terminal is the part I love most. Charts, limit orders and TP/SL on every buy all sit on one screen. It runs on the same wallet as the Telegram bot, so I start a trade on my phone and manage it from the desktop without moving funds. There’s no official $REX token yet, and any official one will only launch from rexbot.near. The full rundown is in the thread 👇 nitter.net/i/status/2105296359674…

A sneak peek at some things we'll be bringing with V2: - More community focus: a dedicated area for backers & holders to raise concerns and give feedback directly to founders. - On-chain data transparency: better surfacing of fund flows, usage & token distribution. - Full UI rework: A simplified UX to help backers understand what is happening pre & post raise. All of this will be dropping later this week, along with a multitude of other platform improvements and new raises.

Private stablecoin payments are moving from theory to product. Aztec relaunched zk.money on Aztec Network as a self custodial wallet for private stablecoin payments, with readable tags instead of raw addresses. You get the privacy of zero knowledge proofs without making every transfer impossible to follow, which is the UX crypto has been missing. I think this is the real counterpoint to the “stablecoins become invisible” narrative. Banks and AI agents want programmable dollars they can move everywhere, while users still need a way to transact without publishing their entire financial history. Robinhood adding AI agents and up to 10x crypto perps pushes in the opposite direction, putting automated leverage in front of people while they carry the risk. → Aztec makes private stablecoin payments usable, not just theoretical → Morgan Stanley is testing stablecoins and DeFi, validating the institutional demand → Robinhood is bundling AI trading, perps and weekend access, raising the execution risk I keep coming back to the same product lesson: self custody only wins if privacy and execution feel simple. I’m building that into @CorvoEdge across Base, Solana, Fogo and Robinhood Chain, because users shouldn’t have to choose between control and usable trading tools. The next moat is private execution.

I've been sitting with the Baranos white paper this week, and the framing that stuck is treating an AI answer the way a chain treats a transaction. The gap it targets is real. The moment an onchain system needs interpretation instead of arithmetic, it usually leans on a private API or a privileged operator, and that answer can move money or classify collateral before anyone gets to inspect it. Section 1 calls "the API said so" a weak foundation, and that lands for me. What a Baranos job actually commits (section 2.2): → one job binds the model, its weights, the tokenizer and runtime, the inputs, the prompt construction and the decoding rules ↳ same committed job → same canonical result ↳ stronger than getting a similar answer twice, independent parties can agree on the exact computation that should have happened How the architecture stays honest (section 5): → it splits the parts so nothing hides behind one opaque endpoint ↳ a model registry, an evidence and input registry, an inference policy, the job itself, and a result account ↳ every component can execute onchain, even when the heavy work runs offchain → two execution modes, and the fallback is the whole point (section 5.1) ↳ Confirmation runs inference offchain, posts the result and its receipts onchain, then verifies and settles ↳ if challenged, the disputed portion is replayed and adjudicated onchain ↳ Replay mode runs the entire computation onchain when that heavier path is wanted ↳ optimistic by design, enforced through commitments and challenge rights, not a zero knowledge proof of every inference Why it lives on Fogo (section 4): → Fogo is an SVM layer 1 on a Firedancer based validator set ↳ picked for the throughput, low latency and low cost needed to return and settle a verifiable LLM result in about a minute ↳ thesis 3 says it cleanly, Fogo settles what the AI actually computed, not what it ought to conclude (section 10) The use case I keep returning to is real-world-asset lending, where collateral is a document rather than a price feed and one operator currently decides whether it qualifies before a loan is issued. As someone building autonomous agents and watching how markets resolve, this is exactly the shape I would want made verifiable. Under Baranos it becomes a pipeline anyone can audit: → a protocol precommits a specific model and an evidence policy → the borrower document is hashed under the job input root → the classification settles in roughly a minute → a wrong call gets challenged and replayed instead of accepted on trust I am borrowing that precommit pattern from section 7, where a prediction market precommits its model, evidence policy, decoding rules and resolution procedure. Lending is my own extension. The live focus today is prediction markets resolving natural-language questions from a frozen evidence set. Two limits I want to be honest about, both straight from the paper: → verification proves execution integrity, not truth (section 3) ↳ it certifies the result model M produced on input X under policy P, nothing more ↳ a verified model can still be wrong, so the evidence policy does real work → untrusted evidence has to be treated as data, not instructions (section 10) ↳ prompt injection sits inside the security model, not outside it ↳ and the paper itself asks for reproducible public benchmarks before the latency and cost claims are treated as proven (page 1 status note, section 12) That last part is why I take it seriously. A team that writes its own caveats into the white paper is easier to trust than one that does not. paper: baranos.ai/assets/baranos-wh… @BaranosAI @fogo nitter.net/i/status/2100453150737…

I built Marlowe for the AnsemHack Clawrena, a Solana agent that does the one thing most trading bots skip. It checks before it apes. Every token gets read before a buy: → mint and freeze authority, on-chain → holder concentration, insiders and pools stripped → liquidity and LP lock ↳ then a clear PASS, WARN or BLOCK, with the reason It even reads its own token, no free pass. Live guardian, radar and MCP at marlowetrade.xyz $MARLOWE FtWnuCZxgmkrDeqYeu9rUWb56vHknU4j14MuWzmeRY8d @clawpumptech

A PLAY holder owns an enforceable claim on machine revenue, not a promise in a thread. If I put money into PLAY, I want the answer to be concrete: my position is tied to revenue from 200 operating claw machines already running in real arcades. That claim sits inside a legal structure, so token holders have an enforceable basis if the operator stops paying. Distributions are paid monthly. The annual yield is targeting 12 to 15 percent, but that target isn’t guaranteed. You can join through UPTIME via @DualMintRWA and @stardotfun. → 200 operating claw machines are already running in real arcades → PLAY is DualMint’s machine vault tied to machine revenue → 12 to 15 percent annual yield is targeted, not guaranteed → Distributions are paid monthly, with contributors joining through UPTIME This is the part I like: the holder’s position has a defined object and a route to enforcement when payment fails. I think that’s the standard real world claims should meet, and PLAY gets there through a claim tied to machines that already operate. Ownership here means enforceable access to machine revenue.

The quiet before the storm. @base @RobinhoodCrypto

👇

It’s $NEAR o’clock, and I’m ready for this one. I’ve been building quietly while the @NEARProtocol trenches picked up. Tomorrow, I’ll show you what I’ve been working on. 🦖

Oh, hi $NEAR.👋 It's cooking time 🍳

Caught $NEARFI early. +1638% so far. I'm still on $ONECAT CA - onecat-eeff96.pad.onetokenhub.near MEOW rexassbot.xyz/c/jm6Z-UJ5NTyl…

Robinhood Chain is live on PipTradeDex, and swaps from it are private by default. Move USDG off Robinhood Chain into Bitcoin, Ethereum, Base, Arbitrum, Sui, XRP, Tron or Zcash with your wallet and route kept out of view. PONS and CASHCAT trade here too.

NEAR AI Cloud is now a provider on @openrouter. NEAR AI offers GLM 5.3 Flash and 1M context, under the provider slug near-ai. Reachable with the key and balance you already have.

The interesting signal in Optimism’s ASP grant is where grant money is going: audit service providers, not just new protocols. I would browse grants.fogofi.com/explore to see what other ecosystems are funding. Infrastructure still matters.

Milestone 01 on @CorvoEdge is done, a week early. I audited the terminal itself. 19 findings, 13 closed. The one that mattered: wallet identity was running on values written in code. Change any of them and every X-login user gets a freshly derived, empty wallet, while login still succeeds and health still reads green. Nothing would have alerted. Fixed. The terminal now records a fingerprint of those values at boot and refuses to start if they ever drift, so silent loss becomes a loud refusal. The same review closed a path where a compromised backend could have handed the client calldata that drained a wallet's whole token balance. The client rebuilds the transaction itself now. Every finding, what it risked, the fix, and the hour it landed: corvoedge.xyz/security

Bitcoin bears are paying up for another leg down. Futures positioning is near yearly lows while BTC, ETH, and SOL are all under pressure. I read that as leverage leaving the system ahead of U.S. jobs data, not proof that spot holders are capitulating. If you’re trading this, the important signal is whether open interest keeps falling while selling slows. The other side of the tape is capital still moving into usable rails. @solana ETFs pulled a record $188 million in a week, but Bitwise took two thirds of it, while Ethena and @binance are pushing USDe collateral into tokenized equity perpetuals. That means crypto is trying to absorb traditional market exposure, even as the Bitget hack reached $387.5 million and reminds you where custody risk still sits. → Bearish futures show leverage getting flushed before macro data → Solana ETF inflows hit $188 million, with demand concentrated in Bitwise → Binance and Ethena are extending stablecoin collateral into equity markets I keep coming back to the same split: liquidity is cautious on crypto beta, but bullish on financial infrastructure. I think the builders who win here won’t just offer more markets, they’ll make risk, custody, and settlement legible inside the trading flow. That’s exactly why I’m focused on self custody and execution at @CorvoEdge, because users shouldn’t need five tabs to understand what they’re exposed to. Positioning matters more than headlines.

Alpha insight I believe this should run