Jerome Leclanche

The through-line

Jerome Leclanche's posts return again and again to a single question: what does Europe need to actually build things, rather than just talk about building things. This shows up in how he runs Ingram Technologies (open-sourcing libraries because "my entire career was built on open source foundations" [14]), in how he designs Financica's accounting engine (deriving state from a ledger instead of bolting on ad hoc fields [7]), and in his running commentary on EU policy and investment culture, where he argues Europe keeps optimizing the wrong metric, whether that's model training spend [3], square metres of pavement [8], or "proven" KPIs over risk [15]. Over the period covered, the focus sharpens from general open-source advocacy toward more concrete interventions: releasing actual libraries [1][14], giving away product access to founders for feedback [4], and pushing specific EU policy asks like EU-INC [12].

On open source

Leclanche frames open source as foundational to his own trajectory and as a debt to be repaid through the company he runs. He notes his first contribution was to Wine "22 years ago" [14] and that "I wouldn't be where I am today without the open source community in general" [1]. That personal history becomes company policy: "a core principle at Ingram Technologies is to release the libraries we build as open source whenever possible" [14], concretized by releasing "8 financial libraries as open source" covering "#einvoicing and the UBL format, XBRL, ECB exchange rates, and other tools around #Peppol in Typescript" [14], and later expanding this into a dedicated "new open source hub" [1]. He frames a gap he's filling: "a lot of the Peppol ecosystem is in Java," so "there is a significant lack of frontend-friendly tools" [14], and he self-critiques that Ingram "has not been the model open source citizen I want us to be," making it "a priority to change this" [1].

On building Financica

His product philosophy for Financica is explicitly UX-first and architecture-first at once. He states the design goal directly: "we think UX-first: How to make the software a powerful accounting solution, without the kitchen-sink-complexity of standard accounting software" [2], asking "How do we make accounting more accessible?" [2]. Technically, this means treating the ledger as ground truth rather than syncing separate statuses to it: "we took a different approach: Use the ledger as the source of truth and derive everything that can be derived from that" [7], so that "there's no such thing as a 'Paid' status on an invoice" since it "can be derived from the ledger" [7]. He extends this to depreciation, "storing only the policy" and deriving the rest [7], and argues more broadly that "every engineer should learn the principles of double-entry" because "double entry databases are massively under-used and under-appreciated in the tech community" [7]. He also builds distribution features to close the gap between books and daily use, like letting users "talk to your books, directly from #Claude, #ChatGPT and even #Telegram" [10], and used direct founder access as a validation method, offering "1 year free subscriptions... to European founders willing to try it and give direct feedback" [4].

On Europe's tech sovereignty

Leclanche has a consistent, sharpened argument that Europe's AI and startup strategy misallocates money and risk appetite. On AI infrastructure: "the next innovations in the agentic AI space are not in the models, they're in the harnesses" [3], and "#Europe does not need to invest hundreds of billions training model, nor should it. The money is burned and it will not give us a leg up in the race" [3], because "the frontier labs are utility providers, competing against open weight models that are just as good given the right harness" [3]. On investment culture, he contrasts US and EU risk tolerance: "a huge difference between European and American investors is pricing on actual potential, rather than 'proven' potential," which "makes many EU investors nothing more than glorified bankers" [15], adding pointedly, "what's the point of risk capital without the risk?" [15]. On policy, he treats EU-INC as urgent and threatened by incumbent interests: "Europe cannot afford to have EU–INC fail to be adopted cleanly," warning against it being "watered down by lobbies such as... NOTARIES" [12], and calling it "our last shot at being relevant, autonomous, and actually regaining our sovereignty" [12].

On administrative absurdity

He treats small bureaucratic frictions as symptomatic of a larger digitalisation failure, drawing on his own experience running a company. He recounts that Belgian employers must "inform, BY REGISTERED PHYSICAL LETTER, the Social Inspectorate where certain payroll records are kept," per "a Royal Decree from 1980" with no digital replacement [5], calling it "the kind of administrative absurdity that is almost completely invisible unless you run a company" [5]. He generalizes: "Belgium's digitalisation problem is not limited to outdated websites or forms... It is embedded in thousands of small legal requirements... that still assume paper, printers, envelopes, stamps and manual processing" [5]. He applies the same lens to public infrastructure incentives, quoting a Goodhart's Law failure around a funding milestone "related to the number of square metres of completed public infrastructure" that made "completion of the paved surface the overriding priority" [8], concluding "Congratulations Brussels, you played yourself" [8]. This same instinct drives a concrete product fix: a financial plan required for Belgian incorporation that founders "generally don't even end up using... other than as paperwork" [9], which he replaces with a 10-minute AI-driven alternative [9].

Takeaways

• Ingram Technologies now open-sources core infrastructure by default, including 8 Peppol/UBL/XBRL/e-invoicing libraries in Typescript, because the Java-heavy ecosystem lacks frontend-friendly tools [14].
• Financica's accounting engine derives invoice status, depreciation, and reports from the ledger itself rather than storing synced state, an architecture he thinks more engineers should adopt via double-entry principles [7].
• He argues Europe should stop funding large-scale model training and instead invest in "harnesses," since open-weight models plus good harnesses can match frontier labs [3].
• He is pushing EU-INC adoption as a near-last-chance policy fight against lobbying (notably from notaries) that could water it down [12].
• He publicly diagnoses Belgian and Brussels bureaucratic and infrastructure failures (registered-letter payroll notification, Goodhart's-Law-driven public square construction) as systemic digitalisation gaps, not one-off quirks [5][8].
• He validates product decisions directly with users, for example giving founders free Financica access in exchange for structured feedback calls [4].