Situations

Learning a Hard Skill as an Adult

Why adult skill-learning stalls differently than it did in school, and three documented approaches, from Oakley, Ericsson, and Bloom, for making real progress.

By Gareth Hoyle·6 October 2026·7 min read

The course was finished months ago, the book sits half-read, and the skill still isn't there in any way that holds up under real pressure. It's not that the learning stopped from lack of effort. It's that the effort went into comfortable repetition rather than the specific, uncomfortable practice that would have actually moved the needle. Unlike childhood learning, which often came with a built-in structure, a class schedule, a curriculum, graded feedback, adult skill-building has to supply all of that structure for itself, and most people never quite do.

Why does generic skill-building advice fail here?

"Just put in the hours" treats learning as a simple function of time spent, when the actual evidence points to something more specific: not all hours of practice are equal, and many adults spend their learning time on activities that feel productive, rewatching tutorials, re-reading notes, without the structure that produces real skill gain.

Generic AI prompting tends to reinforce this gap. Ask an assistant "how do I get better at this skill" and it will often return a generic study plan, more hours, more repetition, without the specificity Bloom's and Ericsson's frameworks both insist on: which specific sub-skill is the actual bottleneck, and what kind of practice targets it directly.

The deeper issue is that adult learners often default to the methods that worked for passing exams in school, reading, re-reading, passive review, which build recognition and recall but don't reliably build the deeper application and synthesis that most real-world skills actually require.

Those methods worked reasonably well for school because exams often tested recognition and recall directly. Most adult skills, a language used in real conversation, a craft practiced under real conditions, a technical skill applied to a new problem, test application and synthesis instead, which passive review was never actually designed to build.

What do three documented thinkers say to do instead?

Barbara Oakley: alternate between focused and diffuse thinking deliberately

Oakley's work on learning how to learn documents two distinct cognitive modes: focused, concentrated attention on a specific problem, and diffuse, the relaxed background mode where the brain continues processing without deliberate effort. Skill consolidation depends on both, not on focused effort alone.

Applied: Structure practice sessions with deliberate breaks, a walk, a shower, sleep, between focused study blocks, rather than cramming for as long as possible in one sitting. The diffuse-mode gap isn't wasted time; it's when the brain does meaningful work on consolidating what was just practiced, which is why a short walk after a frustrating session often produces a breakthrough that more grinding in the same session wouldn't have.

Oakley's research also documents chunking, building small, well-practiced units that eventually combine into fluent, complex performance. Early practice should isolate and repeat a small chunk until it's genuinely automatic, rather than attempting the full complex skill before any sub-component is solid. Attempting the whole performance too early tends to produce the frustrating experience of feeling incompetent at everything simultaneously, rather than the more productive experience of mastering one piece at a time.

Anders Ericsson: deliberate practice targets a specific weakness with fast feedback

Ericsson's research, synthesized for a general audience with science writer Robert Pool in Peak (2016), distinguishes deliberate practice, targeted work on a specific, identified weakness with immediate feedback and a mental model of expert performance to compare against, from naive practice, simple repetition of what's already comfortable.

Applied: Before a practice session, identify the single specific sub-skill that's currently the weakest link, not the skill in general, and design a focused drill that isolates and repeats it, with feedback fast enough to correct errors within the same session rather than days later.

The feedback speed matters more than most learners assume. A week between attempting something and finding out whether it was done correctly allows errors to become habitual before they're ever corrected, which is far harder to undo than catching the same error within the same session. Ericsson's documented research across domains found that the specific combination of a clear performance target, immediate feedback, and focused repetition at the edge of current ability, not raw hours, is what reliably separates improving performers from plateaued ones.

Benjamin Bloom: track which cognitive level you're actually operating at

Bloom's taxonomy distinguishes levels of mastery: recall, comprehension, application, analysis, synthesis, and evaluation. Many self-directed learners mistake recall, being able to recognize or restate a concept, for a level of mastery much higher than it actually represents. Bloom's original work was built for classroom assessment design, but the underlying distinction between levels applies just as directly to anyone trying to honestly gauge their own progress outside a classroom.

Applied: For any sub-skill, honestly test which level you're actually at: can you merely recognize the concept when shown it, or can you apply it unprompted to a genuinely new problem you haven't seen before? Most stalled progress comes from practicing at the recall level while believing the practice is building application-level skill. Re-reading a chapter feels like learning because recognition improves quickly and noticeably; it just isn't the same cognitive operation as applying the material under new, unfamiliar conditions.

Which approach fits which kind of plateau?

ApproachFromWhen it fitsWatch out for
Focused and diffuse alternationBarbara Oakley, learning-how-to-learn researchPractice feels effortful but progress has stalled despite consistent hoursDiffuse mode requires genuinely stepping away, not just a different kind of screen time
Deliberate practice on a specific weaknessAnders Ericsson and Robert Pool, PeakYou can't identify what's actually holding your overall performance backRequires honest, specific diagnosis of the weak link, not just more general repetition
Taxonomy-level self-checkBenjamin Bloom, taxonomy of educational objectivesYou suspect you're mistaking recognition for genuine masteryNeeds a genuinely novel test case; testing with familiar material will overstate your actual level

What can you run to actually start?

1. The weak-link diagnosis
"Here's the skill I'm trying to build and what my practice has looked like so far: [describe it]. Help me identify the single specific sub-skill that's most likely the actual bottleneck holding back overall progress, based on what I've described."

Why it works: naming one specific weak link turns a vague sense of "I need to get better at this" into a focused, practicable target.

2. The mastery-level test
"Here's a concept or sub-skill I think I've learned: [describe it]. Design a genuinely novel test scenario, one I haven't seen before, that would reveal whether I can actually apply it, not just recognize or recall it."

Why it works: a novel test case exposes the gap between recognition and real application that familiar review material tends to hide.

3. The deliberate practice drill
"Here's the specific weakness I've identified: [describe it]. Design a focused practice drill that isolates just this weak point, with a way to get fast feedback on whether each attempt improved or not."

Why it works: isolating the weak point and building in fast feedback is what separates deliberate practice from the naive repetition that produces hours of effort without proportional improvement.

What won't a better study method fix?

These approaches sharpen how practice time gets used and how progress gets honestly assessed. None of them can manufacture the time itself; a genuinely time-constrained adult life limits how much deliberate practice is possible regardless of how well-targeted it is. The frameworks improve the return on the hours available. They don't create more hours, and a realistic timeline for most hard skills still has to account for the genuinely limited time an adult life leaves for deliberate practice.

Barbara Oakley sits in the Performance & Mindset category, and Anders Ericsson and Benjamin Bloom both sit in the Educator category alongside Edward de Bono and the rest of the expanded bench there.

All the copy-paste prompts from this guide, and the rest of the Situations series, live at /prompts too, free to copy without reading the guide first.

FAQ

Frequently asked questions

Is it actually harder to learn a hard skill as an adult than as a child?

The research is mixed and depends heavily on the skill, but the more consistent finding is that adults have less structured time, more competing obligations, and less tolerance for visible incompetence, all of which slow progress more than any difference in raw learning capacity.

How much deliberate practice does Ericsson's framework actually require to see results?

The famous '10,000 hours' figure is widely misquoted and refers to a specific study of elite violinists reaching world-class levels, not a general threshold for basic competence. Meaningful, visible improvement in most skills happens on a much shorter timeline if the practice is genuinely targeted at a specific weakness with fast feedback.

What's the difference between Oakley's focused and diffuse modes in practice?

Focused mode is concentrated, effortful attention on a specific problem. Diffuse mode is the relaxed, unfocused state, a walk, a shower, sleep, where the brain continues consolidating what was practiced without deliberate effort. Both are necessary; skipping diffuse mode entirely, by cramming only in focused sessions, actually slows consolidation.

How do I apply Bloom's taxonomy if I'm learning alone, without a teacher assigning levels?

Self-assess honestly which level you're actually operating at for a given sub-skill: can you merely recall a concept, or can you apply it to a new problem you haven't seen before? Most self-directed learners overestimate their level because recall feels like mastery until it's tested against a genuinely novel application.

Can AI actually help me learn a physical or hands-on skill?

It can help structure a practice plan, identify what to focus on next based on specific weaknesses you describe, and generate novel practice scenarios, but it can't observe your actual physical technique or hands-on execution. For skills with a physical component, a human coach or structured video review still fills a gap AI alone can't.

What's the single most common mistake adult learners make?

Staying in comfortable repetition of what's already been mastered rather than deliberately practicing at the edge of current ability, where the real discomfort and the real growth both happen. Repetition feels like progress; it's often just maintenance.

Written by Gareth Hoyle. Last updated 6 October 2026. Part of the authority.md guides library.

Keep reading

More guides.