Jazz • Product & Motion Designer #NeoBank

"we will summarize it" is the sharpest line in the piece. the whole business model in three words, applied to the people it displaced to exist

@_0xghost_ the "alt season soon" accounts exist because comfort scales better than accuracy. most people aren't paying for insight, they're paying for permission to stay bullish.

I see some in comments say I want the US to be destroyed. Bc I’m spreading awareness of the mess of the situation & risk to financial markets. That’s like saying I want BTC to be destroyed when I said it looked like top was in in Q4. Point is that ppl will treat anything they don’t like to hear as an attack, even a personal attack, instead of as a data point which needs consideration. This is why there are so many BS large KOL accounts who have hundreds of thousands of followers & all they spew is “alt season soon” BS. Bc ppl here don’t actually want to learn anything or even make money, they just want to be told what makes them feel good.

@lordjorx @protocol_fx "nobody had to wake up" is the only benchmark that matters during a $30B liquidation cascade. governance votes under stress are just organized panic with extra steps.

Every stablecoin works until the market breaks. Between October 2025 and February 2026, the crypto market saw over $30B in liquidations across multiple crash events. @protocol_fx built fxUSD for exactly those conditions, and through every single one of them the peg held between $0.999 and $1.001. The protocol just executed its design: > Arbitrage redemptions enforce the floor > A stability pool absorbs dips > Minting halts until peg restores > Funding fees activate to re-peg the stablecoin fxUSD is fully collateralized by wstETH and WBTC, verifiable on-chain every block. The protocol has 16 independent security audits and also has real-time threat monitoring with @HypernativeLabs (that could have mitigated incidents like Resolv's) before they spiraled for hours. I also met with @cyrille_briere from the team this week and left genuinely convinced this is one of the most ambitious protocols in DeFi right now. fxMINT already changed how cheap and accessible it is to create stablecoins, something we've talked about many times but still feels massively underused. And what's coming with @FX100Perp looks like it could shift things even further. The best stablecoins are the ones nobody has to worry about when the market dumps.

@0xzak the irony of a "far ahead in cyber capabilities" model being found via a misconfigured bucket is too clean. security posture and capability claims have to travel together or neither means anything.

So apparently Anthropic left 3,000 internal docs in a public data lake and security researchers found them. The docs reportedly describe a new model called claude mythos that sits above opus in a new tier called capybara. Their own draft allegedly says it's "far ahead of any other AI model in cyber capabilities". If true, the company building the most security-sensitive AI on earth got popped by a config flag.

@sabben Japan at 4.7 is the more interesting data point. third largest economy on earth, generational engineering talent, and almost nothing converting to unicorns. the cultural unlock is the variable that GDP and universities alone can't explain.

Sweden powerhouse for unicorns! 🇸🇪🦄 Sweden produces the second most unicorns in the world relative to GDP - only behind Israel and ahead of the US. Truly impressive 🤯 +Entrepreneurial culture +Supported by (government) infrastructure enablers that are globally attractive + Founder Factories + strong technical universities + growing angel culture = Winning! Source: @dealroomco

@AzFlin the "meant to be coders" framing is the real insight. AI didn't create new engineers, it removed the initiation ritual that was keeping some of them out.

AI has not changed much in regards to human talent tbh cracked people are still cracked. retarded people are still retarded but there is one interesting case - the previously non-technical guy that is coding prolifically they were meant to be coders b4 but were lead astray

@staysaasy because the tools lowered the floor and nobody noticed they also lowered the ceiling. speed without taste is just more mediocrity, faster

I don’t want to be that guy again but eh why are there still 0 good new consumer products in the age of AI. Why has literally every art form turned into fast casual throwaway mediocrity. Why are existing products getting worse.

@systematicls "call option on ambition without paying the premium" is a sharp frame. most people want the identity of the bet without the exposure.

When you really talk to people about ambition, you start to understand that most people want the call option of ambition but are unwilling to pay the premium. It’s always the same excuses: what if it doesn’t work or that “I have a good thing going”, or “let’s see how it goes”. Always finding a hedge, always finding an excuse to “diminish risks”. That’s no way to live your life - you should be daring greatly and recognize that the only way to live and do something is to push it to its most extreme form. At the limits, you understand the shape of the problem and the opportunity much better, and can act with better information than those dancing at the fringes.

@grahamfergs @Securitize @apolloglobal @FissionXYZ @Loopscale sub-2% haircut on an Apollo private credit redemption is the number that matters. that's not DeFi speculation, that's institutional liquidity infrastructure actually working.

Intra-quarter redemption for @Securitize + @ApolloGlobal's ACRED at a sub-2% haircut. Powered by @FissionXYZ on @Loopscale. Making illiquid assets liquid one asset at a time.

@BigSeanHarris the framing is clean but it skips a step, money eliminated trust between strangers, but institutions reintroduced it as a compliance layer. KYC isn't antithetical to money, it's antithetical to the original promise of money.

Money was literally invented to not have to know your customer. It was created to eliminate the layer of trust needed between the exchange of goods. KYC is antithetical to money.

@adityaag the companies built on what models can't do yet are the ones worth backing. "too early" and "wrong timing" are different bets and most people conflate them.

Everyone wants to avoid taking on the models head on. Surprising how few people take the opposite approach Lean into the risk. But make a bet on capabilities that are 12-18 months out. The models suck at something today but I am going to bet it is good enough after XXX months

@BranfordEquity the helicopter example is the whole post. static assumptions don't just look naive, they tell the investor nobody in the room has actually sold anything.

I get pulled into fundraises that aren't going well. Always the same conversation. Founder thinks investors are treating them unfairly. Thinks they should be getting better terms. Thinks something is wrong with the numbers but can't pinpoint what. I ask them to send me their forecast. Their financial model. Whatever they've been showing investors. "Our accountants do that." So I get the model. And every time - every single time - the person who built it has never sat on the investor side of the table. Has never had to argue the case for a business plan under scrutiny. Has never had someone pick apart their assumptions line by line. Had a hardware manufacturing client. They make helicopters. The model their accountant built was projecting 11 unit sales per year, split evenly across quarters. 2.75 helicopters per quarter. Cash landing proportionally at 0.75 of the sale value. That's like a surgeon planning an operation by dividing the patient into equal sections. It bears no relationship to how anything actually works. You don't sell three quarters of a helicopter. You don't collect three quarters of the cash. The real pattern is lumpy - maybe 4 in Q1, nothing in Q2, 3 in Q3, 4 in Q4. Cash collection lags by 60-90 days. Deposits come first, final payments on delivery. Every one of those details changes the cash flow profile completely. The moment an investor opens that model, the meeting is over. They know instantly that nobody in the room understands how the money moves through the business. Here's why this keeps happening. Accountants operate in a completely different world to the one you need for fundraising, for growth decisions, for anything commercial. Their success metric is minimising tax. That's how they justify their fees. That's what they're trained for. Taking a commercial position is a bet. You can get bets wrong. I've got them wrong in my career. But if you never take the bet, you can never be seen as losing. So accountants stay safe. They want to operate in a full informational sphere before making any recommendation. That sounds responsible. In practice, you never have full information. Decisions get made with 60-70% of the picture and a strong view on the rest. That's what commercial operators do every day. Accountants freeze in that environment. What that looks like in a model: - All costs treated as static. No connection to growth drivers. Marketing spend sits on its own line with no linkage to CAC, LTV, or revenue - - No unit economics. No way to stress test what happens when you acquire 50 more customers or lose 20 - - The deliverable is a P&L page. Compliance-focused. Optimised for HMRC, not for scaling the business - - The bridge between historic financials and the forecast is always weak. The model just jumps from "here's what happened" to "here's what we think will happen" with nothing connecting the two - - Inputs scattered everywhere. Jargon throughout. Nobody who didn't build the sheet can audit it I've seen accountants actively resist aggressive growth decisions that were legally sound, commercially obvious, and massively positive on a risk-adjusted basis. Their concern was that it might create a question further down the line. Meanwhile the return-weighted outcome of making the move was enormous. They'd rather protect themselves from a hypothetical audit query than help the business grow. That's the fundamental gap. Your accountant is optimising for compliance. You need someone optimising for the commercial outcome. Those are two completely different jobs requiring completely different instincts. Fundraising is entirely about taking positions. "Here's what we think the business is worth. Here's why. Here's the model that proves it." That requires someone willing to put their neck on the line and defend the assumptions under pressure. Your accountant will never do that. That's not a criticism of their ability. It's a completely different job. If you're going into a raise with numbers built by someone whose primary objective is keeping HMRC happy, you're bringing a tax return to a negotiation. I can usually tell within 10 minutes of looking at a model whether it was built by someone who's ever had to defend it in front of an investor. If yours hasn't been stress-tested by someone who has, fix that before you start the process.

@valkenburgh enforcement discretion is not law. the next administration just reverses it. this is the only version of the argument that actually matters.

I didn't sign up for just vibes-based pro-crypto policies: non-binding guidance, memos about deprioritizing enforcement, speeches. I want the laws changed to be clear that publishing and maintaining privacy software without a license is not a crime. I want a court to rule on that question or Congress to pass the BRCA or a similar safe harbor. I will not rest until it is done.

@antoniogm the irony is both movements are being funded by the most centralized capital on earth. the idealism is real, the ownership structure less so.

Crypto->return to open protocols and decentralized idealism AI->return to CLIs and fuck SaaS lock-in Xennial nostalgia for the early Internet as tech ideal totally vindicated

@jchervinsky the yield debate is a product question. the developer classification is an existence question. not the same conversation.

Stablecoin yield has dominated media coverage of the CLARITY Act. It matters, but it's not the only issue. The biggest challenge is ensuring non-custodial software developers aren't misclassified as money transmitters. That's non-negotiable for DeFi, and it's still unsettled.

@CyprxResearch @StartaleGroup @OfficialXFX $109M in one day and none of it is a token raise. that's the signal people should be paying attention to.

Stablecoin infrastructure funding is accelerating. $109M raised. 4 deals. 1 day. @StartaleGroup ($50M) Tazapay ($36M) @OfficialXFX ($17M) Payy ($6M) Capital is flowing into one of the most commercially proven parts of crypto: moving money better. Why now? - Regulation improving (MiCA, GENIUS Act) - Clear use case (cross-border payments) - The infrastructure stack is maturing What started as a trading tool is becoming: onchain cash rails. And historically this is how financial shifts start: - Payments first. - Markets later. Next wave likely includes: - RWA infrastructure - Tokenized securities - New onchain market structure Stablecoins aren’t the end game. They’re the entry point to rebuilding financial infrastructure onchain.

@chuk_xyz @whop the revenue visibility angle is the part banks should actually be scared of. Whop already knows who's earning, how much, and how consistently. that's better underwriting data than any credit bureau.

Whop shipped 6% yield to 21M users. Most of them have no idea it's DeFi. If you haven't heard of @whop, they're like Shopify, Patreon, and Udemy rolled into one for the internet generation. 21M users, $3B+ in payouts, growing 255% year over year. Their new Treasury product enables balances to earn up to 6% on @aave through @veda_labs via an integrated self-custodial wallet on @Plasma. The user just sees a number going up every second in their Whop dashboard, with zero crypto friction. Classic DeFi mullet. While congress debates whether stablecoins should be allowed to pay yield, Whop shipped yield through lending markets, sidestepping the debate entirely. Everything is a bank. Any platform with float can offer yield as a feature. This is the same threat to deposits that banks have been lobbying against, except this one is less in their control. Users are on Whop because that's where they earn. Now they can grow their balances there too. It's only a matter of time before Whop enables spend (cards) and credit (they can already see your revenue). This is the consumer equivalent of merchant banking, and banks will need to step up the utility of their offerings in order to compete.

@Markkrypt the app layer racing ahead of key delegation and tx simulation is just storing up risk. one high-profile hallucinated trade and the whole narrative takes a hit.

Sharing a lot about how AI agents will become the biggest users in crypto. But is slapping general-purpose web2 $AI frameworks onto existing defi actually enough? Or does crypto need its own native agent infra built specifically for machines? Probably yes imo. Agents are just upgrading the existing machine layer. – 500+ agent projects already, ~$3B mcap – AI sector sitting at ~$16B mcap – AI dApps already ~18% of all dApps – agents moved billions TVL chasing yields Crypto is actually the only environment where agents make full sense. – wallets = native machine identity – stablecoins = programmable money – composability = agents can chain actions across protocols in one flow – 24/7 markets = no downtime for capital The settlement layer is mostly there. The app layer already has stuff like @virtuals_io, @HeyAnonai, @elizaOS. But the middle layers like coordination, execution, identity still have a lot to be built. – no clean key delegation for agents (they can’t safely hold keys long-term) – no MEV-aware routing layer for agents – no reliable transaction simulation to prevent hallucinated txs – no standardized identity / reputation (ERC-8004 still early) Web2 AI was built for forgiving environments. Make a bad API call? Just retry or roll it back. Onchain is the exact opposite. If an agent gets prompt-injected or hallucinates a trade, the funds are literally gone forever. That’s why the first wave of DeAI was just degens aping launchpads, AI agent CEOs… that mostly round-tripped to zero. They were basically nothingburgers because the underlying infrastructure wasn't there to make them secure. Now that the masses are fading the space, there are real teams keep building to turn crypto into the default financial system for machines

The asymmetry that makes this worse: Group 1 ships in public, iterates visibly, and gets the discourse. Group 2 solves harder problems behind NDAs and internal wikis. The knowledge compounds privately, which means the public mental model of "how AI works in production" stays permanently behind.

One unproductive AI discourse pattern keeps to be how individual workflows preferences are talked as the universal hallmark of software engineering. Group 1: A solo builder with agents, their preferred stack, and a pile of markdown files, working on their own apps, is the right way to build and everyone else ngmi Group 2: A much larger group building with agents at scale inside companies, where coordination, reliability, shared systems, and organizational complexity create a very different set of problems which most people don't hear about. tbh individual workflows can still be directionally useful to show new ideas, but they can also be not stable, and enterprises might have very different problems that individuals don't ever have. It's like why small startups don't need to or shouldn't operate like Google, but Google kind of has to operate more or less like company of Google's scale. All ideas are good but much of the AI narrative still very confidently comes from group 1 too little from group 2 (with few notable exceptions).

@Grantblocmates The real tell is when the token whitepaper arrives before the product thesis. At that point the fundraise is the product, and everything else is narrative support for the raise.

teams need to just stop launching tokens unless absolutely necessary for the protocol to exist you are judged on the popularity contest of the token by several orders of magnitude more than the product/service itself if you have to raise money, any worthwhile vc will accept equity greta products and teams are ripe for acquisition now , M&A is up and to the right if they are pushing you for a token when it is not needed, then see through it if you must have a token and its only job is to sell so fund runway you do not have a good business any teams unnecessarily launching a token will fail faster than those without one