Fluid and Crystallized Intelligence: Why Smart People Keep Solving the Same Problems

Fluid and crystallized intelligence helps explain why smart people often end up solving the same problems twice. Sarah had just cracked it: three hours of deep work on a Tuesday afternoon, and she had finally figured out why her team's quarterly planning kept derailing. The insight was clean, the solution elegant, and she felt that buzz that comes when something important clicks into place.

Psychologists call this insight problem solving.

Six weeks later, she was staring at the same mess again. Different project, same underlying problem. The previous solution was somewhere in a notebook, or maybe buried in a Slack thread. She would have to start from scratch again.

That pattern, problem solving followed by amnesia, is one of the quieter ways intelligence fails to compound. We solve problems, feel good about it, and move on. But the solution path, or the insight itself, evaporates. The next time a similar challenge appears, we are back to square one.

I have built my cognitive training programmes around what I see as a crucial weakness in how many people think about thinking. A lot of cognitive training focuses on getting better at solving novel problems. But that is only one third of the story. Intelligence is not just solving. It is a loop.

The Gf-Gc Loop

The so-called G Loop maps onto something psychologists have studied for decades: the relationship between fluid intelligence (Gf), your raw problem-solving power, and crystallized intelligence (Gc), your accumulated knowledge, skills, and patterns. But where most people see these as separate abilities, I think it is more useful to see them as a self-reinforcing cycle.

The sequence is simple: solve a problem using fluid intelligence, crystallize what you learned into retrievable knowledge, then redeploy that knowledge the next time you face novelty.

Gf -> Gc -> Gf. Solve, capture, reuse.

The middle step is where things usually fall apart. We are culturally obsessed with problem solving: hackathons, brainstorms, breakthroughs, moments of insight. But we are often poor at crystallization. That insight you had in the shower, that clever decision you made under pressure, that mental move that saved you three hours of confusion, it often stays trapped in episodic memory, filed under "that one time", and impossible to retrieve when you need it.

The default mode is solve-only. We rely on being fresh, on having time to think, on remembering what worked last time. And when you are tired, rushed, or the context has shifted slightly, you fall back on your oldest habits. You can be genuinely smart and still feel as though you are living the same week on repeat.

Why habits are not enough

The natural response to this problem is usually some version of: make it a habit. Practise the same move enough times and it becomes automatic. But for genuine intelligence, that misses something important about what makes crystallized intelligence actually useful.

Habits are powerful, but they are also rigid. They work by binding a specific response to a specific cue: when you walk into the kitchen, you make coffee; when you sit at your desk, you check email. The automaticity is real, but it is locked to particular contexts. A habit optimised for one situation often fails to fire, or fires incorrectly, when the surface features change.

What we are after is not just automatic response. It is portable automation: tools that you can recognise you need and deploy flexibly across different contexts, even when the cues are not identical.

That is a more demanding target than ordinary habit formation. It requires encoding solutions at the right level of abstraction: specific enough to be actionable, but general enough to recognise when they apply in situations that look different on the surface. The "when to use it" part has to capture the deeper pattern, not just the surface features of the first situation in which it worked.

This applies to rules as much as skills. It is the difference between "check assumptions before I optimise marketing copy", which is too specific to travel well, and "when everyone has opinions but results are flat, test the crux assumption first." The second version is portable. It could apply to hiring decisions, relationship conflicts, or product roadmaps.

What the crystallization of intelligence actually looks like

The IQMindware approach is about compressing insights into what cognitive scientists call mindware: not motivational slogans or vague principles, but cognitive tools with clear conditions of use. Each mindware operator captures five things: when to use it, what to do, how to tell whether it worked, what the tempting wrong move is, and where the tool does not apply.

Take the crux test. A marketing team is stuck because its onboarding emails are not converting. The natural response is to argue about tone, rewrite the copy, and split-test subject lines. Some of that might be necessary eventually, but it is often premature optimisation.

The crux test says: before you polish anything, identify the one assumption that must be true for any of this to matter. Then run the quickest possible test that could prove it wrong. Maybe the assumption is that people are actually opening the emails. Test that first. If they are not, subject-line testing is a waste of time.

This can sound obvious in retrospect. You might say it is just good scientific thinking. But that is exactly the point. Most people can reason this way occasionally when they are fresh and focused. They just do not capture the move in a form that lets them deploy it reliably the next time, especially when the surface features have changed.

The redeploy problem

The third part of the loop is where my approach diverges from much of the productivity world as well as much of cognitive training. It is not enough to document your solutions and insights. You have to prove they transfer.

After creating a mindware operator, the programme requires you to redeploy it under novelty: use the same solution path in a different context, with at least one key constraint changed. If the crux test worked for email marketing, try it on a personal decision: choosing a gym plan, hiring someone, or selecting a course to take. The move is the same. Identify the assumption that has to be true, test it cheaply, and only then optimise the details.

This is deliberately effortful. It would be easier to write the insight down and call it finished. But without testing whether a move survives a different domain, different stakes, and different surface features, you do not know whether you captured the right level of abstraction. You may have memorised "test email assumptions first" when the real skill is "identify and test crux assumptions." One is trivia. The other is transferable.

Far transfer's secret

Most brain training can improve performance inside a specific task. Get better at a memory game, and you will get better at the memory game. The more valuable question is far transfer: whether training in one domain makes you better in genuinely different and meaningful contexts. Historically, the evidence there has been mixed at best and often bleak.

My argument is that one reason far transfer is hard is that most people never do the work required to make it possible. They do not encode the invariant, the part that actually generalises. They do not make retrieval easy through practice. And they do not practise correct deployment under novelty.

We are trying to build skills that travel, and we only count a breakthrough as real when it survives a new context with different constraints and can still be deployed in the moment, flexibly cued by the situation.

The loop itself, solve, crystallize, redeploy, is the mechanism. Not magic. Just the basic engineering of turning experience into skill.

The compounding that usually does not happen

What makes this absence feel invisible is that most people are learning from experience, just inefficiently. You do get somewhat better at navigating office politics, debugging code, or making decisions under uncertainty. But it happens slowly, messily, and with enormous amounts of redundant effort.

The solve-crystallize-reuse G Loop is about making that process deliberate. Not solving faster, but keeping what you solve. Not having more insights, but losing fewer of them. Not being smarter in the moment, but having better tools available when the moment arrives and you are tired.

It is the difference between intelligence as performance, something you can do when conditions are right, and intelligence as infrastructure: a system that quietly accumulates, compounds, and stays available even when you are not at your best. What was once a problem to solve becomes an operator you can draw on in a later fluid reasoning challenge.

Most brain training programmes ask you to stay in permanent problem-solving mode. We are trying to build a system where effort turns into something you keep.

Whether that works is still an empirical question. But the diagnosis, I think, will feel familiar. We solve problems all the time. We just rarely finish the loop and extract the transferable skill and knowledge that compounds.

Being in the G Zone

One final detail matters in the IQMindware framework, because this loop is not just a "write it down" problem. It is also a state problem. The Gf -> Gc -> Gf cycle seems to work best when you are in what I call a near-critical zone: not under-stimulated, not overloaded, but right on the edge where the mind is alert, flexible, and engaged.

Too damped, and you do not generate anything genuinely new. You stay on autopilot, collecting information without changing how you think. Too over-amped, and you may still work through solutions, but they will not crystallize well. You are too reactive, too tense, or too over-generative.

Near-critical is the sweet spot where three things can happen at once: you hold enough focus to solve the problem, stay flexible enough to notice the underlying pattern, and remain calm enough to encode what worked rather than simply escape the moment.

In other words, the best insights do not just happen in the G Zone. They consolidate there. That is where fluid intelligence can become portable mindware, and where crystallized intelligence becomes a usable complement to fluid intelligence rather than just a pile of disconnected memories.