Jo Lernout

The through-line

Everything Jo Lernout says circles back to one conviction: that a small Flemish company had a real, structural advantage in speech technology, and that the advantage was taken from it rather than lost. The advantage was language. American companies built speech tech only in English, which left an opening for a Belgian company to build every other language version and sell it back to the Americans [1]. That arbitrage founded Lernout & Hauspie, carried it into deals with Ray Kurzweil and Bill Gates, and remains the premise of the multilingual AI avatar business he is building at 76 [1]. The second half of the through-line is the loss, which he describes not as a business failure but as a theft: "Die 16 miljard dollar, ik ga heel luid en duidelijk articuleren Bert, die hoorden toe aan al die mensen die hun geld kwijtgespeeld zijn, inclusief Pol en ik. Het werd ons gestolen, letterlijk bestolen op een schandelijke manier" [1].

What has shifted is the addition of self-accusation. He now names his own naivety as a form of guilt: as the person ultimately responsible, failing to sense the game being played against him is itself a failure [1]. That conclusion has hardened into his single defensive lesson, and it sits alongside an unusually cheerful posture toward the present: "Ik heb mij nog nooit zo geamuseerd als voorheen, ook nog nooit zoveel gewerkt" [1].

On language as the competitive edge

The founding insight was an arbitrage, not an invention. American firms were building speech technology in English only, so the other languages were unclaimed ground, and a Belgian company could build them and sell them back into the American market [1]. He extended the same argument into the sales pitch itself. Against Microsoft's 500 monolingual engineers he set Belgian engineers who had spent years in school with 20 hours of language classes a week, and framed his people as "first linguists, then engineers" [1]. The claim was not that L&H worked harder but that its engineers were built differently by their education, and that this was a structural advantage no amount of Redmond hiring could copy quickly [1].

On holding your ground

The Microsoft negotiation is his set piece on nerve. When Gates tried to halve the deal at signing, Paul Hauspie walked out with the line "Do you think we are a bunch of stupid farmers from the West Vlaanders?" [1]. The next morning the answer arrived: "Bill Gates komt binnen en zegt: guys, I apologize, it's a deal. Het was 6 miljoen dollar" [1]. The lesson he draws is simple and unsentimental, that the party trying to reprice you at the table is testing whether you will stay [1].

On sharing the passion and the wealth

He treats entrepreneurship as three universal requirements plus one of his own. The universals are offering added value, knowing your trade, and being able to count [1]. The fourth was culture: "Onze leuze was share the passion maar share the wealth. De laatste was de echte reden waarom dat de club die bij ons werkte, mensen die ik vandaag nog altijd tegenkom, die zeiden: dat was een zalige tijd bij jullie" [1]. He insists the wealth half did the real work of retention, and he backs it with the hardest case: when L&H could not make payroll, the founders paid employees with mortgages first and put themselves last in line, at a moment when both founders were nominally worth around 500 million each [1]. The anti-founder-privilege stance was not a slogan applied in good weather.

On how it was taken

His account of the collapse is deliberate rather than accidental. He describes the Dictaphone acquisition as a "reverse acquisition" trap: buying it loaded L&H with $450M of debt, and Dictaphone's Pentagon-linked business made a foreign owner of surveillance-relevant speech technology intolerable to US intelligence interests [1]. The domestic side of the story is political cowardice. A 50 million bridge loan proposed by the Flemish science minister was overruled by Verhofstadt's circle, where he argues a French company with government backing would have been rescued [1]. What he says Belgium threw away is scale, not sentiment: "Dan was er hier een halve Google in België en dan gingen wij met de belasting alleen die we erop betalen heel die put hier vullen die ze nu hebben" [1].

On foxes and fences

The defensive lesson is the one he states in his own idiom: "Als je een mooi kiekenkot hebt en er zijn veel vette kiekens, zorgt dat een omheining bouwt, want van al die vette kiekens komen de vossen van over heel de wereld" [1]. Success attracts predators in proportion to how visible the value is, and the fence means two things at once, structural and legal protection around the asset, and people around you who genuinely have your back [1]. This is where his self-diagnosis lands as practical advice rather than regret: naivety at the top is not innocence, it is a failure of the job [1].

On avatars and why he is still building

The current business rests on an observation L&H already made decades ago: people freeze when they talk to a blank screen, and converse naturally with a lifelike avatar [1]. The face is what unlocks the speech technology, and combined with the multilingual premise it is the same bet as the first one, now with the interface problem solved [1]. He is running it at 76 and reports both more enjoyment and more work than at any earlier point [1].

On what counts as success

He defines success in a way that survives losing the company: "Succes is als je uw visie kunt omzetten in werkelijkheid. Zelfs al hebt het dan niet meer zelf gedaan, dat is voor mij de maatstaf van succes" [1]. The vision realised counts even if someone else finished it. His summary instruction follows from that: "Nieuwsgierig blijven en kijken wat er mogelijk is en wordt in de maatschappij, toepassen en eru mee amuseren, en de opbrengst delen met uw mensen. Dat is mijn boodschap" [1].

Takeaways

• Look for the market the incumbents have not bothered to serve: American speech companies built English only, which left every other language available to a Belgian company to build and sell back to them [1].
• Sell the advantage you cannot be copied on. Against 500 monolingual Microsoft engineers, the pitch was Belgian engineers with 20 hours of language class a week, "first linguists, then engineers" [1].
• When a counterparty tries to halve the deal at signing, walk. Paul Hauspie did, and Gates returned the next morning at the full $6M [1].
• Three universals for any founder, added value, knowing your trade, and being able to count, plus a fourth that drives loyalty: share the passion and share the wealth [1].
• Pay your people before yourself, especially in a crisis. In the payroll failure the founders paid employees with mortgages first and put themselves last [1].
• Build the fence before you need it: "Als je een mooi kiekenkot hebt en er zijn veel vette kiekens, zorgt dat een omheining bouwt, want van al die vette kiekens komen de vossen van over heel de wereld" [1].
• Treat naivety at the top as culpability, not bad luck. Failing to sense the game being played against you is itself a failure of the person responsible [1].
• Give speech technology a face. People freeze in front of a blank screen and speak naturally to a lifelike avatar [1].