“AI does not think.”
— Albert Berdellans, Exploring Prosperity, April 6, 2026
“Never ask AI a question trying to get an answer.”
— Jordi Visser, Exploring Prosperity, January 18, 2026
Its greatest value lies in helping humans think better. This essay is about how to use it for that purpose.
What AI Does Well, and What Humans Do Better
The most productive way to use AI begins with understanding what it is naturally good at, and where human judgment still matters most.
What AI does well:
rapidly synthesizing large bodies of information
structuring ideas and problems
developing many options quickly
generating first drafts and alternatives
surfacing supporting and refuting evidence
What humans do better:
judgment
intuition
making logical connections
assigning values and priorities
making final decisions under uncertainty
AI is best used to magnify human capabilities. AI helps generate the raw material of thought at extraordinary speed. Humans remain responsible for meaning, selection, hierarchy, and decision.
Best Use Cases for AI
These are some of the most effective ways to use AI to augment your thinking.
1) Learning quickly.
For the first time, every user has access to something close to a subject-matter expert that can respond at the speed of the questioner’s mind. Each response fits like a logical piece of a jigsaw puzzle, allowing understanding to build in the natural sequence of curiosity.
This may be the greatest potential contribution of AI to humanity.
2) Describing and structuring problems so they can be solved.
Many difficult problems remain unsolved because they are poorly framed. AI is exceptionally useful in helping articulate what the actual problem is, where the bottleneck lies, what assumptions are embedded within it, and what possible solution paths exist.
3) Clarifying ideas.
A vague intuition, business concept, essay thesis, or strategic thought can often remain trapped in rough form. AI helps convert those intuitions into structured language that can be tested, refined, and improved.
4) Stress-testing your ideas.
This is particularly valuable because it reduces confirmation bias. Instead of asking only for support, ask for the strongest case against your idea. This often improves the quality of judgment far more than reinforcement does.
How to Get the Best Results
The quality of the output depends heavily on the quality of the setup.
1) Have AI help define the problem.
Before asking for an answer, use the first exchange to improve the question itself. This is different from providing context (see below). The purpose here is to sharpen what you are actually trying to solve: the objective, the audience, the constraints, and what success should look like. Improving the question materially improves the answer.
For example, before asking AI to help write an essay, first ask what questions it needs answered by you about the thesis, intended reader, tone, and desired outcome.
2) Provide robust context.
Explain what you are trying to accomplish, what role you would like the AI to play, and provide any supporting information.
(Example #1. You might tell the AI it is to be a top professional golf coach and explain that the attached simulator data comes from your last six practice sessions, that your primary goal is to improve consistency with your irons, and that your biggest concern is losing strokes through poor dispersion and distance control.)
(Example #2. You might explain that the attached record covers the last two weeks of your schedule, including meetings, email time, project work, and personal commitments, and that your goal is to understand where time is consistently being lost and which recurring activities are creating the greatest bottlenecks.)
One note of caution: “Context rot” begins around 8,000 to 10,000 words of cotext, which as its name implies, causes degradation in the quality of the AI output.
3) Provide accurate prompts.
Well-framed prompts produce materially better results. Tell it whether you want:
ideas
critique
alternatives
a first draft
evidence
opposing arguments
(Example #1. Use the attached simulator data to identify which parts of my swing most need improvement. Tell me which clubs are strongest and weakest, which errors are likely to cost the most strokes, and design drills to improve the weakest areas first.)
(Example #2. Use the attached schedule record to identify where time is being lost, where recurring bottlenecks exist, and recommend specific workflow changes to improve productivity.)
4) Finally, when facts matter, cross-reference.
Use a second model, an outside source, or the primary source material itself. AI is strongest as a thinking tool, not as a sole authority.
Moving from Prompts to Agents
Most people still use AI through prompts, but a prompt solves a problem only once. Agents improve recurring workflows. An agent can research, draft, revise, schedule, and follow up inside a repeatable system. This is where many of the productivity gains begin to compound.
One excellent use case is the automation of recurring friction. Many daily bottlenecks that we once accepted as part of life can now be addressed directly:
repeated administrative tasks
recurring research workflows
follow-up communications
scheduling sequences
structured decision trees
This is where AI begins to move beyond assistance and into systems design. The setup takes more thought upfront but the returns compound over time.
Build the Habit
Do not get in the habit of using AI simply to answer questions. That can quickly become a crutch, often revealed the moment you are asked to explain the subject in your own words.
Use AI whenever you encounter friction, confusion, or opportunity. Make the output yours. The more often you use AI where progress slows, the more naturally it becomes a thinking partner and the more clearly you begin to see its leverage.
Thank you very much to our sponsors:
Truflation: BUY Truflation products/services and save 10% using code BOBDEW10. Learn about what prices are actually doing by going to https://truflation.com/marketplace/us-inflation-rate
Foundation: Protect your Bitcoin using the latest off-line technology with easy-to-use Passport. Their next product, Passport Prime will protect all passwords off-line with even better technology. https://foundation.xyz/
River Financial: If not cold storage for your Bitcoin, I’d highly recommend River for buying and holding- River is engineered to protect your Bitcoin over the long run. Use this link for discounts: https://river.com/signup?r=T7GGAF7G