FSC: 2027 Outlook

Any time a system summarizes data, it loses detail. The part “discarded” is of a specific quantity, and it can be computed. That's the whole of what Final Stop Consulting does: find where a system replaced reality with a summary, and measure what the summary dropped.

I spent the last several years developing that methodology in physics, using math to build precise models of complex behavior. What I took away from that is that real discovery happens when you use conflicting viewpoints to test assumptions, rather than just defending settled positions.

With that said, here is where FSC is headed in 2027 as we explore new areas of productive collaboration:

Soil and ecosystems

Long-term environmental data relies heavily on net numbers, which may obscure actual mechanics. If a soil's net gas absorption changes, a net number can't tell you which underlying processes shifted. It’s a measurement design idea, and it's worth a thought.

AI systems

AI models work by forcing new inputs into categories they already recognize, rounding off anything that doesn't fit. When these models sit inside hiring, benefits, or credit decisions, that rounding quietly alters people's lives. We need a way to measure those invisible gaps, and there may be a mathematical way of doing so.

Science of Science

Every manuscript rejection I've received that came with substantive feedback produced a better paper, a better understanding, and a better argument. Granted, as an independent author, those are the minority.

Most institutional filtering doesn't start with the research; it pattern-matches the author. When credentials and institutional ties dictate what gets read, solid work gets filtered out before anyone evaluates the argument.

But “papers” are the least of it. A paper is the record of a question someone was permitted to ask. The filtering happens earlier, on the question itself, and it happens hardest to the people inside. A graduate student with three years of runway learns fast which questions are fundable and which are career-ending, and narrows to fit (because that's what the pipeline selects for). A lab renews on results, so it proposes what it already knows how to find. A field's boundaries harden until a question that belongs to two of them belongs to neither, and nobody asks it. None of that produces a rejection. Nothing was submitted. I can count my rejections, but nobody can count those.

That's the loss that truly hurts, and science has no instrument pointed at it. The published record measures what survived the filter, which tells you nothing about what the filter removed (the same way a net number tells you nothing about the processes underneath it). Publication rates go up. The questions being asked narrow down. Both are true at once and the record can't distinguish them.

I'm exploring plain-language evaluation frameworks that strip away status signals, forcing work to be judged strictly on its methodology, underlying basis, and results. Not to make it “easier to publish.” To make it possible to see what isn't even being attempted.

None of this is about validating independent authors or turning “cranks” into scientists. It's about whether a system that filters on credentials before it reads the argument has any way of knowing what it discarded. That's the same failure as every other item on this list: a summary standing in for the thing itself, and no one checking the difference.

Enterprise

This is where I started and where I still spend most of my time.

The same structural problem exists in business: headcount, engagement scores, and competency models are just summaries. Over time, companies start managing to the dashboard and stock market rather than the reality of its employees.

Right now, capital is tight and companies are making heavy cuts in the name of “discipline.” But efficiency isn't real if you're just transferring the cost outside the company's boundaries (onto employees' households, public systems, or nowhere at all). A firm can post great quarterly numbers while quietly destroying the operational capabilities that made those numbers possible in the first place.

I don't think leaders are doing this out of malice. I think they gradually replaced reality with a summary and didn't notice the swap. That’s a solvable problem, and it’s the one I’m most interested in fixing.

If you’re dealing with the measurement side of any of these issues, I'd like to hear from you.