The Shred Shark Thesis

The internet made publishing free, and the value moved to knowing what was worth reading. SaaS made software rentable by the month, and the value moved to knowing which system would hold. Apps put a build in everyone’s pocket, and the value moved to knowing what actually worked once real people used it.

Each wave was called a fad, each one removed a layer of cost, and each one raised what it cost to be wrong. The tool got cheaper; the checking got dearer. That is the pattern the board is built on.

AI generates work that looks finished. Code that compiles, a drawing that renders, a report that reads well, a control sequence that runs once. The gap between plausible and correct is now the expensive part of every engineering job, and it is closed by someone with enough experience to know what correct looks like.

Robotics pushes it further, because the checking happens in the physical world: a crane, a rotation, a process line, a commissioning test. As more of that work is done by machines, the rigour goes up rather than down, and the person accountable for the result becomes more valuable, not less.

So we do not treat AI as the end of technical careers. We treat it as the reason verification becomes the discipline — human in the loop, in unison with the tool, from software through the trades.

Capital is funding the automation of judgement: tools that screen, review, write and decide. We are funding the supply of judgement instead — the experienced engineers, technicians and operators who are trusted with the result.

That is unpopular for a reason: it is slower to build, it needs real people with real evidence behind them, and it cannot be faked by a demo. It is also where the bottleneck sits, and value accrues where the bottleneck is.

The second half of the bet is flexibility. Experienced technical people are scarce enough to ask for the arrangement they want, and companies that refuse to offer one lose them. A board that carries the arrangement on every listing — remote, compressed, rotational, hybrid — is negotiating on the side of the scarce person, which is how the supply stays.

Job seekers pay nothing: every listing, every alert, the matches from their profile, and messages from companies.

Companies claim their profile free — the page already exists before anyone claims it — and pay for the two things they cannot get for free: a featured post placed above the standard listings, and messaging the talent database. Attention and access, priced plainly, with no lock-in and no per-seat games.

The asset that compounds is the supply of verified technical people. Employers pay for reach into it; job seekers never pay to be part of it, because a board that charges its own supply has nothing to sell.

Profiles built from evidence rather than adjectives: the years, the stack, the seniority, the fields worked in, and the arrangements someone will actually take.

Listings that state the arrangement they offer, so a four-day week is a fact on the card rather than an implication in the description.

Breadth on purpose: remote engineering roles are live today — software, data, cloud, platform, security and QA — and the same standard widens into the other technical fields and the other ways of working as listings arrive. Aviation, lifting and heavy plant, marine, controls and automation, and the skilled trades are the long-term point, not a slogan.

If verification becomes fully automated and trusted without a human — if a model can certify its own work and be believed — the premium we are building on moves somewhere else.

If employers stop paying for attention because every board has every role, the two paid products lose their price.

If flexible technical work becomes common enough that no board needs to argue for it, the positioning stops being a differentiator and becomes the floor.

We would also be wrong if the supply never arrives: a board for experienced technical people is worthless if experienced technical people do not believe it is on their side. That is why job seekers never pay here, and why the copy on this site keeps saying so.