This is The Menu: the weekly briefing from the Web3 Dinner Club.

What the market is saying and how the best builders and investors are making sense of it.

In this issue:

Signal, served weekly.

The dinner club is proudly sponsored by Novel Labs.

A multi-award-winning London storytelling studio building the brands of the future in AI, blockchain, and emerging technologies.

Best known for the $100m expansion to the Bored Ape Yacht Club, The Mutant Cartel World.

If you’re a startup or scale-up building a brand and looking for real go-to-market impact from those who have repeatedly built unicorns and category kings as VCs and founders... ask for an intro at the table.

A crypto card looks and works like an ordinary debit card, but it lets you pay using money held in crypto or stablecoins.

When you tap to buy a £5 coffee, the card provider converts enough of your chosen crypto balance into pounds in the background.

The café still receives pounds through Visa or Mastercard; it does not need to accept crypto or do anything differently.

The whole process happens in seconds.

Think of it as a translator between two money systems:

The useful part is convenience.

The important caution is that conversion can involve fees, exchange-rate spreads and, depending on where you live and what you spend, possible tax consequences.

Crypto cards used to be a novelty: a way to prove you could buy a coffee with Bitcoin.

That has rapidly changed.

The latest generation of cards is not just about converting crypto to fiat at the checkout.

It’s about giving users one spending layer across stablecoins, cryptoassets and traditional currencies, with different approaches to custody, rewards, fees and risk.

Kraken’s US launch is a good example.

Its Kraken Card, launched in the United States on 18 August, lets eligible users spend from 600+ cash and crypto balances through a Visa debit card.

This is less “a crypto card” and more “a multi-asset wallet with a card attached”.

There is no single best crypto card. Users are choosing between different trade-offs:

The key change: crypto cards are no longer all trying to solve the same problem.

Some focus on stablecoin spending and international payments. Some target high-volume users. Some are built around DeFi and self-custody. Others are turning rewards programmes into token ecosystems.

A card offering 5–10% cashback can look compelling. But: cashback in what?

If rewards are paid in a platform token, subject to vesting, their value can rise or fall before you can use them.

A “zero fee” card also needs closer inspection. The cost may show up in: FX charges, conversion fees, ATM fees or the spread applied when crypto is sold to complete a transaction

The boring checklist matters more than the exciting perk:

The most useful crypto-card use case may not be spending Bitcoin at all.

It may be:

Crypto cards are becoming part of the wider shift towards multi-asset money.

You may hold pounds, dollars, euros, USDC, Bitcoin and tokenised funds in different places today. The next generation of products is trying to make them feel like one balance sheet, while leaving the user to decide what to spend, what to hold and what to keep separate.

As useful as this all sounds, the old rules still apply:

The crypto card is no longer a gimmick.

It is becoming a consumer interface for a world where money comes in more than one form.

If you build products, raise capital, sell, hire or write online, Robert Cialdini’s Influence is essential reading.

This is not a book of persuasion “hacks”. It is a field guide to the mental shortcuts people use when they are busy, uncertain or overloaded with choice.

Cialdini’s insight: we often make decisions less rationally than we think, and good founders understand those shortcuts.

Persuasion is usually about reducing uncertainty, not making the loudest pitch.

Used well, these principles do not mean manipulating people. They mean making genuine value easier to understand and act upon.

The ethical line is simple: do not invent urgency, fake testimonials, borrow credibility you have not earned, or manufacture a sense of community to stop difficult questions.

Use real evidence. Show genuine demand. Be clear about constraints. Build trust that survives scrutiny.

Influence is one of those rare business books that remains useful because it is not really about business. It is about people.

For founders, it offers a more intelligent way to communicate value. For everyone else, it provides a defence against being rushed, flattered or socially pressured into a decision.

Ethical persuasion is not about making people do what you want.

It is about helping the right people recognise genuine value, while giving them enough clarity to say no when it is not for them.

A curated, seated dinner for a small group of builders working in crypto, AI, and frontier tech.

One table. No pitches. No panels. No ego contests.

Just the kind of conversation that doesn't show up in your LinkedIn feed. The relationships that move capital, talent, and ideas in Web3 don't start at conferences.

They start at a handful of dinners with the same people, repeated over time.
Seats are limited by design.

Proudly sponsored by Novel Labs.

Crypto ETFs make digital assets easier to buy, but they do not make them less volatile.

They bring bitcoin and ether into traditional investment accounts. For investors who want exposure without managing wallets, private keys or crypto exchanges, the appeal is clear.

But do crypto ETFs make crypto safer, or simply easier to access?

Advantages: Easy access through a brokerage account, no private-key management, familiar ETF structure and pricing, potential portfolio diversification.

Disadvantages: Volatility remains unchanged, fees reduce returns, no direct ownership (you own shares in a fund, not bitcoin), limited trading hours while crypto trades 24/7.

Crypto ETFs solve an access problem, not a volatility problem.

They are useful for investors who want exposure without managing wallets or private keys. But the familiar ETF structure can make the underlying risk feel smaller than it is.

Treat a crypto ETF as a high-risk allocation, examine its fees and structure, and invest only an amount that fits your tolerance for loss.

Crypto ETFs may bring crypto into traditional finance, but they do not turn crypto into a traditional asset

At W3DC, DYOR usually means the tech, the opportunity, the wallet, the protocol or the asset.

This week, it also means the rules.

The FATF report on DeFi is a policy document about money laundering, terrorist financing and proliferation financing risk. Not light reading. But the core message matters:

DeFi may be decentralised in technology, but it is not always decentralised in control.

DeFi TVL hit $86.6bn in 2026, up ~85% vs 2023. Institutions, VASPs and regulated entities are increasingly interacting with DeFi.

DeFi brings real advantages: innovation, cross-border payments, transparent transactions, global access.

But its core features also create risk: pseudonymity, permissionless access, smart contract automation, composability and cross-border reach can all move funds quickly and opaquely.

FATF’s stance: standards are technology neutral. They apply when a person or entity provides financial services. So the question is not:

Is this DeFi?

It is:

Who if anyone has control or sufficient influence?

Many DeFi projects look decentralised but retain centralised elements: concentrated governance tokens, admin privileges, upgrade control, economic benefit, development influence, control over infrastructure.

If someone can meaningfully control, profit from, change, govern or operate the access layer, regulators may treat them as responsible.

FATF splits DeFi into three types:

In practice, regulators are behind:

DeFi has moved fast. Regulation is still working out where responsibility sits.

The report flags abuse: fraud, ransomware, money laundering, proliferation financing, chain-hopping, bridges, DEXs, mixers, governance manipulation.

“Controllers” may include developers, governance token holders, core maintainers, front-end operators, investors, oracle providers, foundations or corporate entities.

Responsibility sits in different places depending on how the arrangement actually works.

“Decentralised” is not a magic word.

It is not enough for a protocol to say it is decentralised.

You need to know that control genuinely exists, where it sits, and whether the structure can be explained, verified and supervised.

FATF is not calling for DeFi to be stopped. It is calling for better risk assessment, clearer governance analysis, stronger AML/CFT controls where relevant, better blockchain analytics, more public-private cooperation and more international coordination.

Truly decentralised arrangements may fall outside the standards, but still require risk-mitigation via adjacent points: stablecoin issuers, VASPs, application-layer interfaces.

If DeFi wants to become part of serious financial infrastructure, it cannot rely on slogans.

DYOR has to go beyond “will this token go up?”.

We don`t like lists but It has to include:

Trust in DeFi will not come from saying “trustless” often enough.

It will come from structures people can understand, verify and challenge.

This is one of those stories that sounds futuristic until you realise it is already happening.

Binance has launched Agent OS, a system that allows AI agents such as ChatGPT, Claude and other compatible tools to connect to Binance and place trades across spot, margin, convert and futures products.

In simple terms: You can give an AI agent permission to trade for you.

Until now, trading bots were usually built by people who understood APIs, code and exchange infrastructure.

Agent trading lowers that barrier. A user could describe a strategy in plain language and allow the agent to execute it through connected tools.

Binance has tried to manage the risk through isolated sub-accounts. The agent can trade inside that account, but it cannot withdraw funds to an external wallet. That protects against direct theft, but it does not protect against bad trades, leverage mistakes or liquidation.

And Binance is not alone.

Coinbase, Gemini, MetaMask, MoonPay and Ledger have all launched or developed similar agent-trading products, each with different custody models. Some use exchange-hosted accounts, some use self-custodial wallets, and Ledger’s approach uses hardware-wallet spending caps.

That tells us the industry is moving quickly, but has not yet agreed on the safest model.

The biggest unresolved question is liability.

If an AI agent makes a bad trade, who is responsible?

For now, the answer appears to be the user. The article notes that the terms of service across platforms largely place the risk of agent-driven trading on the person using the tool.

Quite the contradiction you may think.

These products are being built for autonomous execution, but users are still being told they must review and control the risk.

There is also a wider market question.

If many AI agents use similar models, data and prompts, they may end up making similar decisions at the same time. In a volatile market, that could increase the risk of crowded trades, flash crashes or liquidation cascades.

The technology is impressive, but, but risk goes beyond technical, it is behavioural, legal and systemic.

AI agent trading may be one of the next big steps in crypto market infrastructure.

But it also creates a very simple problem:

The agent can act faster than the user can think.

That makes permissions, custody, spending limits, leverage controls, security audits and liability frameworks more important than ever.

The future might be autonomous. But your losses will still be very human.

Until next time