Alek

@sahill_og wild part is that people still pay for it, for now...

90% of "AI startups" are just: - Take user input - Send to OpenAI API - Display response with good UI - Charge $25/month

@yashhq_22 fight with your claude - that's how you get results.

Most people are using claude wrong. they’re prompting it like it’s an employee not like a co-founder.

@naval the best interface will be no interface

AIs replace UIs and APIs.

@defyneric float

every payment processor should have their own stablecoin

@geoffreywoo inside a business the compounding data layer is where all the value accrues

announcement for founders: if youre raising from me and your AI story is just "we use the latest model," save both of us the meeting. model access is not the moat. show me the workflow you own show me the data exhaust you compound show me why GPT-6 getting better helps you more than it hurts you thats the new baseline. everything else is pitch deck makeup.

MCP gives your agent access to tools. It does not give your agent an inbox. These are not the same problem. The industry is about to figure this out.

Every team building multi-agent systems has quietly built a custom message bus. Custom inboxes. Custom routing. Custom state handoff. Always slightly different, always fragile. It is the most duplicated infrastructure in AI right now and nobody talks about it because it feels like a failure. It's not a failure. It's a gap. The frameworks gave you orchestration. They didn't give you addressing. There's no standard way for an agent to say "send this to agent X and wait for a reply" across system boundaries. So everyone rolls their own. You're not bad at infrastructure. You just got handed an incomplete stack. What's missing is a layer between the frameworks - persistent addresses, typed messages, cross-system threads. The plumbing that agents assume exists but doesn't. Building it now.

@chapello Agent-to-agent payments are the future.

The first trillion-dollar stablecoin category isn't cross-border B2B. It's machine-to-machine. Today's rails were built for humans initiating payments -- and they're excellent at that. They simply weren't designed to represent an AI agent billing in sub-minute increments and settling to a smart wallet. Programmable money is the prerequisite for programmable work. Agents are scaling faster than the rails purpose-built to serve them.

the "re-explain everything every session" pain is the single biggest problem nobody's built real infrastructure for yet. AGENTS.md is a band-aid. RAG is a guess. fine-tuning is a lock-in. he actual fix: your agent publishes what it learns to a persistent URL. next session, next platform, next agent- it's just there. we're building this.

@chapello 100% - not using stablecoins is leaving money on the table

The reason stablecoin adoption goes vertical from here isn't transfer fees. It's yield. USD sitting in a corporate treasury today: ~0%, plus a 3-day clearing tax. USD sitting in a programmable wallet tomorrow: 4-5% APY, settles in seconds, available 24/7. "Free transactions" got the headlines. "Yield on idle cash" wins the CFO meeting.

@sebbydavies 💯

@AlekPerak Legacy rails can't settle $0.002. This is a plumbing problem.

Your SaaS subscription model assumes your customer is a human sitting at a dashboard. Your next customer is an AI agent that makes 4,000 API calls in one afternoon, pays $0.002 per request, and never logs in again. It doesn't want a seat license. It doesn't want annual billing. It wants pay-per-task on rails that settle in milliseconds. That means stablecoins, not ACH or cards. The subscription economy was built for humans. The agent economy runs on micropayments.

@shafu0x Meanwhile associating stablecoins with crypto hurts stablecoins

the only reason anyone still takes crypto seriously is stablecoins

@reganbozman USD + yield until needed for spend, then covert

onramp/offramp for local stables will be fast and free for most western g10 countries in next 1-2 yrs imo biggest unlock is changing consumer behavior to holding usd and only converting to local currency at point of spend (rather than at receipt)

Most "stablecoin companies" won't exist in 5 years. Not because they'll fail. Because the label will stop meaning anything. DoorDash is paying drivers in stablecoins. Stripe built a blockchain for settlement. Klarna launched its own stablecoin. None of them call themselves a stablecoin company. Nobody calls Walmart an "internet company." But try running Walmart without the internet. Stablecoins aren't a category. They're a layer. And the pure-play startups building ramps and wallets? Ramps will be free. Wallets will be default. The value moves to the programmable infrastructure underneath. The biggest stablecoin winners in 2030 won't have "stablecoin" anywhere in their pitch deck.

Agreed, but the trojan horse goes deeper than speed. Employers pre-fund payroll days early. That float just sits there. Once that capital is in stablecoins, you unlock everything else: yield until disbursement, native FX, programmable splits, treasury visibility in real time. Payroll gets you in the door. The full stack keeps you there.

Payroll is the trojan horse for widespread stablecoin adoption. Think about it: > employees get their paychecks faster > employers circumvent slow and expensive traditional rails (ACH/wires) → win-win Fixing a pain point is how you integrate something new into people’s daily lives, and stablecoins inherently eliminate all of the problems payroll currently has. Combined with @Brookwellapp, receiving and spending those stables is as seamless as traditional rails.

This is the right framing. Every enterprise CFO I've talked to asks the same question before moving treasury on chain: "can my competitors see who I'm paying?" Not how much. Who. Until that answer is no, institutional stablecoin adoption has a ceiling. The Bessemer team identified the actual bottleneck.

Bessemer just published their 2026 stablecoin thesis. It's worth reading in full. The headline numbers: $273B in stablecoin supply, $10.9T in adjusted volume last year, real-world payments doubled to $400B. Stablecoins have crossed the chasm. What caught my attention is the privacy section. Out of five startup opportunities Bessemer identifies in the entire stablecoin stack, privacy is one. And in that section, they name two companies: Canton and Hinkal. I think the choice of these two is not random. Most privacy projects in crypto solve for one variable: hiding the amount. That's not sufficient for enterprises and institutions. You can shield a transfer's value and still leak the entire counterparty graph. A competitor watching the chain still sees who pays whom, how often, through which routes, and at what cadence. That's enough to reconstruct supplier relationships, payroll, treasury flows, and trading positions. For enterprise stablecoin adoption, hiding amounts alone is not privacy. It's obfuscation. Real counterparty privacy requires hiding the wallet relationship itself. Sender, receiver, and the connection between them. This is a different problem because it has to work across public chains that are public by design, and it has to work without breaking compliance. The way Hinkal handles it: stealth addresses break the link between sender and recipient on-chain, UTXO-style commitments hide balances, and viewing keys held by the client (not Hinkal) allow disclosure to 3rd parties. 24/7 asset screening against Chainalysis databases to prevent illicit assets from access. The chain sees Hinkal smart contract. The client sees their own books. The regulator sees what the client chooses to share. Deployed across EVM chains, Solana, and Tron. Canton solves the same problem in a different shape: a permissioned ledger where privacy is the default and selective disclosure is built in. Different architecture, same recognition that counterparty privacy is the real unlock for institutions. The market is starting to draw the right line. Privacy that hides amounts but leaks counterparties is not a category. Privacy that hides the relationship is. Thanks to @cbirn , @eric_kaplan_nyc , and @BrandonNydick for a sharp piece. Link in comments.

@dbwoods11 @BobbyThakkar Compliance teams make things way more difficult than needed. I'd say there's a lot of friction on the software side too

@AlekPerak @BobbyThakkar from your POV is it a regulation requirement problem, a software problem, or something else?

stablecoins work. the onboarding doesn't can't put a normal person through a 6-step KYC flow and call it the future of payments