Harold Kinet

The through-line

Everything Harold Kinet says circles back to a single question: where does trust actually sit in a transaction, and can it be moved out of an institution and into code without losing legal force. He frames this as an unfinished job for the whole industry, arguing that "the payments industry has automated almost everything. Except the moment that matters most: conditional settlement between parties who don't fully trust each other" [12]. The answer he builds toward is escrow that no party controls, settlement in stablecoins, and arbitration handled by a named human with legal accountability rather than a platform.

The second thread is the passage of blockchain from fringe to institution, which he narrates from personal vantage. He has been in the ecosystem since 2017, when "la blockchain c'était encore un sujet de niche, souvent regardé avec naïveté et ensuite avec méfiance par la finance traditionnelle" [1]. Nine years later he opens a seminar on digital assets in front of actuaries, and reads that as the meaningful signal: traditional finance is reclaiming the technology, but on its own terms, "avec la rigueur qu'elle exige toujours : modélisation du risque, cadre réglementaire, standards actuariels" [1].

On the moment payments never automated

The gap he keeps returning to is conditional settlement. His evidence is loss data: the FBI recorded $2.77B in losses from business email compromise alone in 2024, most of it wire fraud on high-value transactions [12]. His framing of the problem is blunt, "Why do we still wire millions on trust?" [12]. The existing alternatives fail on availability rather than concept: bank escrow is "slow, expensive, often unavailable cross-border or below 5M", which leaves everything under that threshold running on pure trust [12].

The mechanism he proposes is deliberately simple to state. The buyer locks funds in a neutral smart contract escrow, the seller ships, and funds release when all parties, up to ten partners, validate delivery [2][8]. The constraint that makes it work is negative rather than positive: "No party can move the money unilaterally" [2][10], and "No intermediary ever holds the funds. No escrow license needed" [12]. Paycifi is positioned as "programmable trust infrastructure" [2], built by a Belgian company and a Circle Alliance partner [2][12].

On who moves first

The same architecture is applied to bilateral crypto trades, where he identifies the deadlock precisely: "The biggest risk is always the same: who moves first?" [13]. The buyer will not send USDC before receiving BTC, the seller will not send BTC first, and "that's where most OTC deals either stall or get exploited" [13]. The escrow resolves it by making the lock visible: the buyer locks USDC in a non-custodial smart contract, the seller sees the funds are secured and sends the BTC, both confirm, and the USDC releases automatically [13].

On arbitration by a human, not a platform

He is explicit that automation stops at the dispute. When something goes wrong, "a specialized lawyer steps in as arbitrator", and he draws the distinction himself: "No platform deciding who's right. A human with legal accountability" [13]. This is paired with ICC arbitration built into the product, described as fast and enforceable in 170+ countries [2][12], which is what makes the code-based settlement usable in real cross-border trade rather than only among counterparties who already trust each other.

On settlement rails and pricing

Settlement is offered in USDC/USDT or EURC, same day and cross-border [2][9], with fiat in and out available for companies not using stablecoins [2][6]. The pricing is stated openly and scales with how much traditional plumbing is involved: from 0.2% for stablecoin settlement to 1.5% where fiat and the on/off ramp are included [2][12], with OTC deals quoted at 0.5% to 1.5% [13]. Commercial focus is on African trade corridors, where deals are currently being structured [2][5], and the standing invitation is to test it against a real transaction rather than argue in the abstract [2][13].

On MiCA splitting the stablecoin market

His reading of European regulation is that it has produced two separate markets: "MiCA just split the stablecoin market in two. A regulated European bubble, and everything outside it" [11]. On the euro side, EURC broke every record it had in Europe, with 1,760 active addresses per day and 713 new wallets created daily according to Circle's data since the stablecoin provisions took effect [11]. On the dollar side the story reverses. USDT is not MiCA-compliant because Tether never applied for the e-money-token authorization, and its CEO was explicit that the requirement to hold 60% of reserves in European bank deposits is incompatible with Tether's reserve model [11]. The market data follows the rule: Kaiko found USDT volumes on EU-regulated venues fell over 70% between Q4 2024 and Q2 2025 while USDC volumes on the same venues nearly doubled, order-book depth for USDT/EUR pairs collapsed, and Coinbase, Kraken and Crypto.com all removed USDT for EEA users [11].

On security failures being human

His account of the Bybit hack refuses the easy conclusion that the technology broke. Some 400,000 Ethereum were stolen by the Lazarus Group, but the entry point was phishing emails that handed the attackers control of operator computers, and the transfer to a pirate account was authorized through human error [16]. He is direct that the platform's own security was not compromised, and that the failure was operator negligence [16]. On recovery he is realistic: roughly 40 million dollars have been recovered, but Lazarus's laundering operations mean most of the stolen funds will disappear into North Korea's military program [16].

On defining the technology precisely

He does definitional work rather than hype work. A metaverse, taken etymologically, is meta as "beyond" and verse as "universe", and what distinguishes it in practice is persistence: shared virtual worlds where modifications remain permanent, unlike video games where the environment resets [15]. He separates this from augmented reality, which overlays virtual elements onto the real world, and stresses that there is no single metaverse, with Facebook's version sitting alongside decentralized ones [15]. The same appetite for structure shows in how he describes BE Blockchain as a consulting and development firm in blockchain and Web3 with two strategic activities, consulting services and building its own products such as its sandbox [14].

Takeaways

• The unautomated part of payments is conditional settlement between parties who do not fully trust each other, and bank escrow does not fill it below roughly 5M or across borders [12].
• Non-custodial design is the whole point: no intermediary holds the funds, no party can move the money unilaterally, and no escrow license is required [2][12].
• Automation should stop at the dispute, where "a specialized lawyer steps in as arbitrator" with ICC arbitration enforceable in 170+ countries [13][12].
• In OTC trades the binding constraint is sequencing, not price: locking the buyer's USDC visibly is what unblocks the seller [13].
• MiCA has bifurcated the market, with USDT volumes on EU-regulated venues down over 70% from Q4 2024 to Q2 2025 while USDC volumes nearly doubled and EURC hit record European usage [11].
• The Bybit loss came from phishing and operator negligence rather than any breach of the platform's security, which is where defensive attention belongs [16].
• Blockchain's move from niche curiosity in 2017 to a seminar topic for actuaries reflects traditional finance absorbing it on its own terms: risk modelling, regulatory framework, actuarial standards [1].