The same data read as text slips away. Seen as a moving picture, it stays. Now that AI can render interactive visuals inside the conversation itself, the speed of understanding has changed. Text wasn't the end of knowledge — it was the entrance.
Ask an AI to explain something, and you almost always get a long piece of text. A table, maybe a code block, but text at the core. For a long time I took that for granted. Then Claude shipped a feature that renders interactive visualizations inside the chat itself. A stock dashboard appears mid-conversation. A compound-interest calculator responds to my slider. Hover the mouse and numbers move. The first thing I felt was, "the speed of understanding is different."
Today let's unpack that difference. This is not a story about Claude. The story is this: explanation and experience enter the brain at different speeds. Slowly now.
Humans understand through two channels.
The first is the explanation channel. Read letters, hear words, translate them into mental images. Read "at 10% compound interest, principal becomes 2.6x in ten years," and your brain converts that into numbers, places them on a timeline, and imagines the outcome. Mid-translation is required. It takes time, and precision varies.
The second is the experience channel. See something move, touch it, adjust it. Move the slider to 10% on a compound-interest calculator and the graph jumps up. Nudge to 11% and it steepens. You translate nothing. Eyes and hands understand directly.
Both channels work. But their speed and staying power differ. Explanations are thorough, and they drift. Experiences are brief, and they stay. The long news article you read yesterday is fogged already. The app you poked at yesterday is still in your fingertips.
Until now AI leaned almost entirely on the explanation channel. Because chat windows couldn't hold moving things. Even the best AI translated everything to text. This is where the illusion was born — that text is the end-form of knowledge.
Concretely. Claude's new visualization works like this.
I wrote, "I want to see compound interest intuitively." Claude wrote some text, then placed an interactive compound-interest calculator inside the chat itself. Slider. Adjust the annual rate. Change the period. Graph reacts live.
I dropped the rate to 5%. Curve mellowed. Bumped it to 12%. Curve steepened. Stretched the period to thirty years. The top shot so high the principal almost vanished.
I read zero text in that moment. Yet the principle — "compound grows more terrifying the longer it runs" — landed in my fingertips. Reading "by the rule of 72, at 10% it doubles every 7.2 years" as a sentence cannot match the intensity.
And this isn't off in a separate panel. Old AIs said, "open the side window." This feature appears inside the flow of the chat. Right between the question I asked and the answer it wrote, the moving picture slots in. Conversation and experience aren't split apart. That is the core.
Picture a bicycle. Try to explain riding one in words. "Grip the handlebar, press one pedal with your foot, kick off with the other, when you build speed bring the second foot up, keep balance by shifting weight forward…" Pages of it. Doesn't work.
When you first learned to ride a bike, did you read a manual? No. Someone held you up, you climbed on, fell, got back on, and learned with your body. Thirty minutes and you had it. Spend the same thirty minutes reading a manual and you still can't ride.
The harder the subject, the bigger this gap. Try explaining stock charts in pure prose — dry and confusing. Hover over a live chart and ask "what happened here?" — five minutes and the intuition arrives. Experience is the shortcut of explanation.
AI being able to render visualizations inside the chat means this shortcut is open on every subject. Old flow: write code, host on a server, share a link. New flow: one message, and you are already in experience.
Same concept (10% compound, 20 years), two modes. My felt experience.
| Item | Long text explanation | Interactive visualization |
|---|---|---|
| Time to understand | ~8 minutes | ~40 seconds |
| Accuracy of grasp | ~70% (numbers confuse) | ~95% (confirmed by touch) |
| Memory after one week | Mostly gone | Big picture still holds |
| "Let me try a new value" | Needs another query | Just move a slider |
| Emotional afterimage | Low | High (remember the surprised moment) |
The point isn't "faster." It is that the brain keeps it differently. Eight minutes of reading fades in a week. Forty seconds of sliding still lingers a year later.
The aha.
Understanding isn't created by explanation. It is created by a picture you can operate.
Not every topic needs a visualization. Some topics read better as prose. One question sorts them.
"To get this, do I need to change values, or chew on sentences?"
Two answers.
When you need to feel relationships, trends, sensitivity. Compound interest. Price reactions. Price-volume relationships. Demographic age distribution. Design comparisons. Sliders and graphs race ahead of words.
When you need meaning, philosophy, interpretation. Legal clauses, poetry, product strategy backstories, historical context. Visualizing these leaves the substance on the floor. Long-form prose wins.
Three words: Values picture. Meaning prose. Border is operation.
Here is a real case from my own stock analysis.
Scenario 1 — a buy-timing dashboard for KOSPI. I asked Claude, "show me the moments over the last three years where timing was favorable." Claude rendered an interactive dashboard inside the chat. Red dots marked buy windows. A range slider. An adjustable threshold for "undervalued." As I moved the threshold, I felt it immediately — "loosen this a bit and the number of buy windows doubles." Intuition in three minutes.
Scenario 2 — the philosophical weight of market swings. Same day I asked, "explain how the volatility since 2008 has shaped Korean retail investor psychology." That came back as long-form prose. Because this topic isn't about varying values — it asks you to read context. Some meanings cannot be drawn on an axis.
Same subject (Korean stocks), received as experience and as explanation. Using both channels made the subject three-dimensional. Neither channel is superior. They do different jobs.
Let me pull it together.
The human brain has an explanation channel and an experience channel. Explanation is thorough and drifts. Experience is brief and stays. Until recently AI lived almost only on the explanation channel. Now that visualization runs inside the chat itself, the experience channel is open. Text was never the final form of knowledge.
The name Claude may change. The visualization feature may be renamed. The principle holds. AI will keep moving from "explain" toward "let you experience." The skill is choosing which channel for which topic. Value relationships go to pictures. Meaning context goes to prose. Tools change. The two channels don't.
Three words to close.
Values picture. Meaning prose. Understanding operates.