Mathias Celis

Mathias Celis returns again and again to one idea: the inbox is not a productivity problem, it is a design failure, and most of the tools built to fix it are treating the wrong disease. His clearest statement of this is blunt. Speeding up email replies, he argues, is "a faster belt on the same treadmill. You are still running. You have just been handed nicer shoes" . The distinction he keeps drawing is between processing more email faster and needing to be in the inbox less at all, so that "your real work gets the good hours instead of the leftovers" . This is not a minor product positioning point for him, it is the founding premise of Alkmist and its assistant Pidgy, and he restates it from different angles across months of posts rather than treating it as a one-off pitch.

That premise is built on psychology, not just frustration. He traces the felt weight of an inbox to Bluma Zeigarnik's 1920s observation that unfinished tasks stick in memory far more than finished ones, concluding that "every unread is an open loop" and that "responsibility lives in your memory instead of in the system" . He pairs this with a second research thread on why to-do lists fail: a study showing that people who reliably finish tasks are the ones who decide in advance when and where they will do the work, leading him to the line "a to-do list makes you decide what to do every time you open it. A calendar already decided. You just show up" . He is explicit that this changed his own planning and is "part of what we're building into Pidgy: turning the emails that pile up into tasks that actually have a time next to them" . The throughline across both citations is that structure, not willpower or discipline, is what closes loops and returns attention to people .

He extends the same logic to how firms relate to clients, arguing that the coming wave of AI adoption in professional services is missing the actual complaint. After surveying more than 100 clients, he reports that "not one named the quality of the work" as the problem; what came up instead was the experience of the process itself, sending the same file three times, never knowing where things stood, starting from zero every year, even as three out of four firms plan heavy AI investment nobody asked for . His conclusion is stated as a flat correction: "Your client is not asking for AI. Your client is asking for attention" . This is consistent with his broader claim that once AI tools are commoditized, "the tools stop being the difference" and the client's experience of the work becomes "the whole game" .

Celis is also unusually candid about wanting Alkmist to be loud in a category he sees as visually and rhetorically identical booth to booth. He describes an accounting trade show as "rows of booths that blend into one beige wall of logos, everyone saying roughly the same thing in roughly the same font" , and the company's response was to build a two-meter box of fake painkillers, Paracetamail, for what he calls "the headache every accountant knows. The inbox" . He frames this explicitly as risk-taking he enjoys as a founder, recalling the moment the oversized prop came off the printer: "That is genuinely enormous. What have we done" . The same instinct shows up in a run of posts built around a staged arrest at Canary Wharf and a joking reference to being "bailed out of London prison" , and in directing followers to find the giant box at events for a hidden payoff . Whether discussing the stunt or the product, his stated logic is the same: differentiation now lives in how the message and the experience are delivered, not in the underlying tool.

On the audit profession itself, he uses a food metaphor to make an argument about perception versus value: "Audit is the kiwi of professions... Brown, hairy, a little off-putting from the outside... for what you pay, almost nothing else gives you that much nutritional value" . His point is that the work happens out of sight, so young people fill the gap with a cliché, and "you can't aspire to something you never see" . This concern with visibility and access for young professionals is presented as a driving reason behind Alkmist's involvement in mentoring, discussed in the same appearance where he examines why the profession matters to society and struggles to explain itself to newcomers .

His view of how to build, rather than just what to build, was sharpened by watching AI-industry figures on stage in Brussels. Observing OpenAI, Anthropic, and Google DeepMind representatives alongside Belgian founders, he concludes: "none of them waited for the conditions to be right. They started rough and fixed it on the way. Most teams do the opposite. They keep planning because it feels safer than starting. It rarely is" . He places Alkmist deliberately in that camp, describing its approach as to "push the boundary first, refine as you go" . Taken together with his account of managing a PhD, prior startups, and current ventures by relying on periodic "cleanup rage," self-care rituals, and people who challenge him and keep him accountable ¹⁶, the portrait is of someone who treats both product design and his own working life as systems to be re-engineered rather than endured: identify the open loop, whether it is an unread email or an unstarted project, and build the structure that closes it instead of asking for more willpower.