Inspired by Steve Blank’s recent essay “No Science, No Startups: The Innovation Engine We’re Switching Off.”
His observations about the erosion of scientific capacity in the U.S. resonate deeply with what we now witness in Europe.
In my own field — digital recruitment and HR automation — I see the same pattern emerging: science is the silent engine behind every genuine startup, yet its role is being underestimated precisely when AI depends on it most.
Science as the Foundation of Innovation
Steve Blank rightly argues that startups are not born from business plans but from scientific progress — discoveries that find their way to market.
Every breakthrough — from semiconductors to mRNA vaccines to machine learning — began with fundamental research, often publicly funded, long before commercialization was imaginable.
In Europe, we have strong institutions — IMEC, Fraunhofer, CEA, TNO — bridging pure and applied science.
But as public budgets shrink and private investment shifts toward short-term development, the foundation weakens.
And when science weakens, startups follow.
A Decade of Data — and the Scientific Method in Business
Ten years ago, the HR sector was still manual and intuition-driven.
Recruitment relied on human sales teams, not data.
At that time, my company began automating job distribution and applicant tracking — ideas that sounded radical in 2015 but were based on clear, testable hypotheses:
Can data predict job performance? Can algorithms replace repetitive sales?
We collected everything — millions of job listings, campaign results, and user interactions.
That data has now matured into a scientific dataset.
With today’s AI and autonomous agent frameworks, we can finally activate it — transforming static information into living intelligence.
Each campaign becomes a learning experiment; each employer dataset refines the model.
What once looked like marketing is now applied behavioral science powered by machine learning.
It’s exactly what Steve Blank described: science, engineering, and entrepreneurship merging into one continuous feedback loop.
From SaaS to Science-as-a-Service
Traditional SaaS models are collapsing because AI scales per workflow, not per user.
In our field, this means moving from “per-seat” subscriptions or CPC advertising to outcome-based systems that sell results — faster hiring, lower cost-per-hire, or predictive talent analytics.
The next step is AI-native, one-tier SaaS: each client gets a custom, semi-open installation trained on their own data but powered by a shared core model.
Thanks to AI-assisted code generation, we can now deliver this level of customization efficiently, even for public-sector tenders — a market previously inaccessible to small innovators.
This evolution transforms how science reaches society: open, adaptive, and measurable.
AI Depends on Science — It Cannot Replace It
Some policymakers argue that AI makes traditional research obsolete.
Why experiment when models can predict?
Because AI itself is the result of decades of scientific curiosity — mathematics, linguistics, cognitive psychology, and computer science.
Without new hypotheses, datasets, and experiments, AI stagnates.
It cannot generate new knowledge; it can only recombine existing patterns.
Scientific method — not compute power — remains the source of progress.
The Chain of Progress
Science → Engineering → Entrepreneurship → Investment.
If one link weakens, the chain collapses.
Europe still holds a powerful position — strong universities, deep research capacity, and global talent — but the pressure is visible:
administrative overload, declining R&D budgets, and fewer young researchers entering the pipeline.
As Steve Blank warned, societies that neglect science don’t just slow down — they silently surrender their future competitiveness.
Closing Reflections
Every startup begins with a question, a hypothesis, and an experiment — just like every scientific project.
The overlap is not accidental; it’s structural.
Science is not a luxury for startups — it is the startup mindset.
Curiosity, iteration, data, and learning from failure: that’s the scientific method, not just entrepreneurship.
So when we underfund research, we don’t just harm universities — we strangle the seedbed of every future company.
The lesson is clear:
Without science, there are no startups. Without startups, no renewal. Without renewal, no future.
Reacties
Een reactie posten