A verifiable minimum theoretical foundation. Exposure to structured working methods. The ability to learn within an institutional framework. These are useful things to verify, and the diploma does it reasonably well for profiles who followed that path.

Competence acquired outside formal education — through self-teaching, personal projects, years of production without a safety net. The capacity for diagnosis under pressure, which is not learned in a lecture theatre. Ingenuity under real constraints, which the Blue developed because his grandmother had no unlimited budget. These competences leave no trace in a diploma system — they leave a trace in the systems that keep running.

A recruiter who uses the diploma as a hard filter acts rationally at the individual level — they reduce their sorting cost. But at sector level, they collectively produce a market that under-uses its pool of real competence and overpays for titles that do not always correspond to operational capability. The aggregated result of individually rational decisions is collectively suboptimal. This is a classic system-level effect.

The intern does not yet carry the constraints of production. They do not have a system to keep alive at 3am. They do not have incident tickets to clear before they can think about anything else. This absence of constraint, which looks like a deficit, is precisely what allows them to spend time on research, experimentation, and exploring technologies that do not yet have an established use case. They explore while the senior maintains. Both are necessary. Neither replaces the other.

The implicit model of an internship is unidirectional: the company transmits its knowledge to the intern, the intern learns. This model under-exploits the relationship. A well-designed internship explicitly formalises the flow in both directions — the intern learns the production system, the team learns from the intern about what they practise outside production. This is not goodwill. It is knowledge management.

The Blue is the archetype. Real technical competence, sometimes surpassing the average graduate, acquired through practice, curiosity and exposure to real problems. His obstacle: he does not pass the automatic filters of ATS (Applicant Tracking Systems) that screen out CVs without a diploma before a human eye reads them. He does not exist for the formal market — even if he has been running production systems for years. His only current lever: finding a recruiter or instructor willing to look at what he can do rather than what he has on paper.

The apprentice faces the opposite situation: institutional legitimacy in the process of being acquired, but treated by the market throughout their training period as a "not yet ready" profile. They are on the ground, genuinely contributing to production, but their "still a student" status creates friction at the point of conversion to permanent employment — some employers prefer to keep the apprentice as an apprentice rather than converting them, for cost and flexibility reasons. The entry pathway recommended by the sector paradoxically produces its own extraction loops.

The intern is the profile most exposed to extraction without return. They make a real contribution — often on tasks the teams do not have time for — and leave without any guarantee of conversion, without formal recognition of their contribution, and sometimes without a usable reference on the job market. The infrastructure internship is structurally imbalanced: the company gets working capacity at marginal cost, the learner gets an experience that does not always translate into market access.

An AWS-certified candidate who has never seen a physical server room, never cabled a switch, never diagnosed a network fault in production at 3am is considered more qualified by many recruiters than a Blue with ten years of bare-metal field experience. This paradox is the direct symptom of the hard limits dissolution described in "Anatomy of the Loss": when physical competence disappears from evaluation criteria, what remains is mastery of service interfaces — and the certifications that validate them.

A cloud or vendor certification costs between €200 and €400 per exam, not counting preparation course costs. For a self-taught person without stable employment or an apprentice in training, this cost is prohibitive. The market therefore requires, to validate a competence often already acquired, a financial investment that excludes precisely the profiles who acquired that competence through their own effort rather than through a funded curriculum. It is an economic filter as much as a credential filter.

The problem with the IDPE is not its design — that is sound. The problem is its access schedule: the opening of applications is announced in the Journal Officiel once a year, on an unpredictable date. For a professional who discovers this mechanism outside the annual window, the wait can last up to twelve months. For a self-taught person actively seeking employment, this administrative friction is often disqualifying. The Blue cannot wait for the next JO to be recognised.

Processing an IDPE application represents a time cost for the school (file analysis, jury constitution and convening). This cost can be covered by processing fees, as is the case for VAE in other curricula. The school finds a complementary revenue model alongside its core activity, and a grounding in the local professional fabric. This is not a free service — it is a fee-based service that benefits both parties.

The IDPE title is already awarded by a jury of graduate engineers, under the same standards as the CTI title. It is not a lightweight title — it is an alternative access pathway to the same title. Devaluation would come from lowering jury standards, not from opening the calendar. Making access continuous does not change the criteria — it removes an administrative friction that has no technical justification.

The Validation des Acquis de l'Expérience (VAE — Recognition of Prior Learning) already allows degrees and titles to be obtained through recognition of professional experience. A permanent IDPE would be its specific version for the engineering title — with the additional rigour of the CTI jury. These two mechanisms should be explicitly articulated, to create a continuum of recognition for non-formal competences that runs from RNCP titles up to the state-certified engineer title.