• The Efficiency Question Nobody Wants to Answer
• Jamie Dimon: The King of the Flip-Flop
• The Private Blockchain Paradox
• Censorship Resistance: What They Promise vs. What You Get
• The Oracle Problem: Where Trustlessness Goes to Die
• DeFi Usage: Following the Numbers
• What Blockchains Actually Provide
• ERC-8004: When Blockchain Meets AI Hype
• The Honest Framework
• The Way Forward

I’ve spent years teaching people about Bitcoin—breaking down the complexities until folks with zero background could explain it to their grandmothers. I believed in the vision. I still believe in parts of it.

But I have a problem with lies.

Not the kind of lies people tell on purpose. The kind that get repeated so often they become truth. The blockchain industry is full of these. Trustlessness. Decentralization. Censorship resistance. These words get thrown around like confetti at a parade, but when you actually try to grab one and hold it up to the light?

It crumbles.

I’m going to tell you what blockchains actually do versus what people say they do. I’ll use ERC-8004—Ethereum’s new standard for AI agents—as a case study. But the lessons apply to almost everything in this space.

§The Efficiency Question Nobody Wants to Answer

Let me put this plainly: blockchains are slow. Catastrophically slow.

A regular database on regular hardware can process millions of transactions per second. Ethereum does 15–30. Bitcoin does 7. That ain’t a small difference. That’s like comparing a Ferrari to a shopping cart with a broken wheel.

Every time someone says they’ve “solved” blockchain scaling, they’ve made a trade-off that undermines the whole point. Layer 2 rollups? They batch transactions but introduce centralized sequencers. Sidechains? They sacrifice security. “High-performance” chains like Solana? They achieve speed by requiring expensive hardware, which means fewer people can run validators, which means… more centralization.

The Blockchain Trilemma says you can only pick two: decentralization, security, or scalability. People keep trying to prove it wrong. Nobody has.

Here’s the truth most builders won’t tell you: If you control both sides of a transaction, or if you can agree on who runs the database, just use a database. It’s faster. It’s cheaper. It’s simpler. It’s been battle-tested for decades.

§Jamie Dimon: The King of the Flip-Flop

I need to talk about Jamie Dimon for a second. I actually built a company that worked with his Bank back in 2017. Because this man is the perfect example of everything wrong with how the financial establishment treats blockchain.

In 2017, the CEO of JPMorgan Chase called Bitcoin a “fraud.” He said he’d fire any trader caught touching it “for being stupid.” Compared it to Tulip Mania. Said it would collapse.

In January 2025, he called Bitcoin a “Ponzi scheme” with “no intrinsic value.” Linked it to sex trafficking and money laundering. Said if he were the government, he’d “close it down.”

By July 2025, he compared owning Bitcoin to smoking cigarettes—you have the right to do it, but he doesn’t think you should.

Strong words. Consistent message. Bitcoin bad.

Now here’s the thing.

While Dimon was publicly trashing Bitcoin, JPMorgan was quietly building blockchain infrastructure. They created Quorum. They launched JPM Coin. They built a whole blockchain division called Onyx, later rebranded to Kinexys. In 2025, they filed a trademark for JPMD—their own deposit token. They launched a tokenized money market fund. They started letting institutional clients use Bitcoin and Ethereum as collateral for loans.

Wait. What?

The same man who called Bitcoin a Ponzi scheme is now running a bank that accepts Bitcoin as collateral? The same man who threatened to fire employees for trading it?

And now Jamie Dimon says “blockchain is real.”

Oh, NOW it’s real? After you spent eight years calling it a fraud?

Here’s what changed: He figured out how to make money off it without giving up control. JPMorgan’s blockchain is private. Permissioned. They control who gets in and who doesn’t. They set the rules. They run the validators.

He hates the part of blockchain that was actually revolutionary—the permissionless, censorship-resistant, nobody-can-stop-your-transaction part. But he loves the buzzword. He loves the “innovation” label. He loves charging institutions for access to a glorified database with blockchain branding.

This is the game. Dismiss the actual innovation. Co-opt the language. Build something that defeats the entire purpose. Call it progress.

Jamie Dimon isn’t a convert to blockchain. He’s a banker who realized he could sell the same old product with new marketing.

§The Private Blockchain Paradox

This is where it gets funny. And by funny, I mean sad.

The whole breakthrough of blockchain—the thing Satoshi figured out—was how to get agreement between strangers who don’t trust each other without needing a middleman. That’s it. That’s the innovation. To achieve this, the technology sacrifices speed, storage efficiency, and privacy.

So what do enterprises do? They build “private blockchains.”

Think about that for a second.

A private blockchain reintroduces the middleman. The moment you have a gatekeeper who controls who can read or write to the ledger, you have removed the need for a blockchain. If you know and trust your validators, you don’t need Proof-of-Work or Proof-of-Stake. You just need a distributed database with cryptographic signatures.

Building a private blockchain is like buying a tank for your morning commute because you want to feel safe—then removing the armor and the gun to improve gas mileage. Now you’ve got a vehicle that’s heavy, slow, expensive to maintain, and hard to drive. Minus the protection that made you buy it in the first place.

It sounds crazy.

But consultants made billions selling this to corporations. And corporations bought it because “blockchain” was a buzzword their boards understood.

§Censorship Resistance: What They Promise vs. What You Get

The pitch goes like this: “No one can stop your transaction. No government. No bank. No corporation.”

The reality?

Tether has frozen hundreds of millions in USDT at law enforcement request. Circle blacklisted Tornado Cash addresses within hours of OFAC sanctions. After Ethereum switched to proof-of-stake, around 70% of blocks became OFAC-compliant because major validators filtered sanctioned transactions. The network has partially self-corrected, but the structural vulnerability remains.

The moment you touch stablecoins or use a chain with concentrated validators, you’re back to trusting institutions. You’ve just added extra steps.

Bitcoin is genuinely harder to censor. Mining is physical and distributed. Miners can’t be easily identified and pressured. But even there, neutrality ain’t absolute. We’re already seeing “soft” censorship where specific mining pools refuse to process transactions containing Ordinals or Runes. They call it “spam” or “dust.” But who decides what’s spam? If a majority of hashrate decides your transaction type is illegitimate—for technical, ideological, or regulatory reasons—the open door slams shut.

The only chain with real censorship resistance is Bitcoin. And even that is a spectrum, not an absolute.

§The Oracle Problem: Where Trustlessness Goes to Die

Here’s the thing blockchain marketing never mentions: smart contracts can only see what’s on the blockchain.

Any real-world data—prices, weather, whether a package got delivered, whether an AI agent did its job—requires an oracle. An oracle is an external service that feeds information into the chain.

You are trusting the oracle operator. You are trusting their data sources. You are trusting their infrastructure.

Chainlink is the biggest oracle network. It’s essentially a federation of node operators. Better than trusting one company? Maybe. But it ain’t “trustless.”

The entire DeFi ecosystem depends on price oracles. When oracles fail or get manipulated—which has happened many times—contracts execute wrong. But they execute “as programmed,” so nobody’s liable. The trustlessness was an illusion. You moved trust from a bank to Chainlink. The trust still exists. It’s just been relocated and obscured.

For something like ERC-8004—Ethereum’s proposed standard for AI agents—this problem is devastating. Agent reputation? Off-chain. Validation results? Off-chain. Metadata about what the agent does? Off-chain. The blockchain part is just a pointer. All the actual trust assumptions live somewhere else.

§DeFi Usage: Following the Numbers

The narrative says blockchain adoption is growing and the future is decentralized.

The numbers tell a different story.

DeFi total value locked peaked around $180 billion in late 2021. It crashed to about $35–40 billion. It has only partially recovered. And most of that “usage” was circular—deposit a token, earn yield in another token, deposit that token, leverage the position, repeat. When the token incentives dried up, the usage collapsed. Because the organic demand was thin.

NFT trading volumes are down over 90% from their peaks. Daily active users on most chains remain embarrassingly low given the market caps.

What looked like adoption was mostly speculation and incentive farming dressed up as utility.

I’m not hating. I made money in DeFi. But I’m also not going to pretend the emperor has clothes when he’s standing there naked.

§What Blockchains Actually Provide

I’m not here to say blockchains are useless. That would be a lie too.

Strip away the marketing and you’re left with a specific set of properties that traditional databases genuinely cannot offer:

Coordination without a trusted third party. When competitors need shared state but won’t let each other run the database and can’t agree on who should. A neutral chain sidesteps that negotiation.

Credible neutrality. Infrastructure that can’t be captured by any single interest. Ethereum can host applications from competing companies because nobody owns it.

Permissionless composability. Anyone can build on top without asking permission. This enabled DeFi protocols to plug into each other without business development deals.

Settlement finality. Once a transaction is confirmed, it’s historically immutable.

These properties matter for a narrow set of problems. For most software, they don’t.

§ERC-8004: When Blockchain Meets AI Hype

ERC-8004 is Ethereum’s proposed standard for AI agent coordination. It creates on-chain registries for agent identity, reputation, and validation.

The case for it: If Google, Microsoft, OpenAI, and Anthropic all want their AI agents to work together, who runs the central registry? None of them will trust each other with that power. A neutral blockchain sidesteps the standoff.

That’s legitimate. I get it.

The case against: Most agent interactions will happen within trusted environments anyway. Companies already use APIs and contracts for B2B coordination. The “trustless” benefit evaporates when you’re still relying on off-chain data for reputation and validation. Gas costs add friction that discourages experimentation.

The cynical read: This is an attempt to make Ethereum relevant to the AI hype cycle. The technical benefits over a federated database with digital signatures are marginal. The real motivation might be creating a narrative that connects ETH to the AI boom.

I don’t know which interpretation is right. But I know that asking the question is more honest than pretending the answer is obvious.

§The Honest Framework

When someone pitches you a blockchain solution, ask these questions:

Who needs to trust whom? Trace the trust assumptions through oracles, validators, and off-chain components. “Trustless” almost never means what it sounds like.

Is it a private blockchain? If yes, stop. Use a database. You’re paying for security features you’ve disabled.

What’s the actual censorship resistance? Who runs the validators? Where are they located? What happens when the government calls?

Could a database with digital signatures solve this? In most cases, yes. The remaining cases where blockchain adds genuine value are narrower than the industry admits.

What happens when the token incentives end? A lot of apparent adoption disappears when you stop paying people to use the system.

§The Way Forward

I’ve spent years in this space. I’ve taught thousands of people about Bitcoin. I still believe in the core vision of financial sovereignty—money that no government can inflate away, transactions that no institution can block.

But I’ve also watched an industry lose its way. Watched marketing replace engineering. Watched narratives replace results. Watched people get hurt because they believed promises that couldn’t be kept.

Blockchains are a real technology with real properties. Those properties are useful for a narrow set of problems involving coordination among people who don’t trust each other, can’t agree on a central operator, and need some degree of censorship resistance.

For most software applications, blockchains add cost, complexity, and fragility while subtracting performance. The industry’s inability to say this plainly has damaged its credibility.

The technology doesn’t need saving through marketing. It needs honest assessment of where it actually belongs.

I’d rather tell you the truth and have you make good decisions than hype you into something that wastes your time and money.

That’s love. And to me, Love and Justice are on the same side.