On August 22, 2026, Anthropic engineer Thariq shared a four-line tool called /eli5: type in any topic, and the AI generates an HTML explainer page that breaks down complex concepts with big images and few words. It racked up 260,000 views within hours. We think the real takeaway here is not how clever the tool itself is, but the underappreciated trend it points to—"how you ask an AI" is becoming a distributable, reusable product form.
What It Is
/eli5 is a community plugin inside Claude Code (Anthropic's AI coding assistant), short for "Explain Like I'm 5." At its core is a four-line prompt instruction: treat any topic as if the other person understands nothing, and output an HTML page rather than plain text, with large images paired with very little text on the page.
Three typical uses: before refactoring unfamiliar code, look at a "map" first (/eli5 how does this module work); quickly get the core logic behind a team's historical decision (/eli5 why did we make this tradeoff); build an intuition framework before reading an incident report (/eli5 what caused this incident). The common thread: it does not replace deep analysis, but serves as a "warm-up step" before deep analysis—picture first, text second.
Industry Reactions
Supporters argue that it standardizes the frequent need to "have AI explain things simply"—you no longer have to think about phrasing every time; just type a command. For engineers who face unfamiliar code every day, these friction-reducing micro-tools have real value.
The skeptical voices deserve attention too. The first critique comes from a cognitive angle: forced analogies can sometimes confuse people more. One example: Claude compared an API call to "driving to a friend's house (the database)," which made the explanation more tangled than necessary. Simply saying "explain simply but keep technical accuracy" actually worked better.
The second critique is sharper: since /eli5 makes the AI produce a verbose answer first and then simplify it, why not just train the AI to output concisely by default? That is a classic case of "paying for the process twice." Thariq's response was candid—reducing verbosity and visual simplification are two different things; the former is an output style, the latter is a way of understanding. That response acknowledges the problem, but also draws a clear line around the use case.
Impact on Everyday People
For enterprise IT: future internal knowledge bases and cross-team handoff documents will almost certainly sprout many /eli5-like "skills"—not new systems, but best prompting practices packaged as one-click commands.
For individual careers: when you inherit a new project, face unfamiliar code, or encounter a new decision, "asking for a picture first" will save considerable cognitive load compared to "reading the full documentation first." This habit will spread from programmers to everyone who must deal with unfamiliar information.
For the consumer market: education and learning apps are most likely to adopt this first, packaging the "explain-quiz-teach-back" learning loop into a standardized product. Someone is already doing it—explainx.ai's Melo built in a similar mode 11 days before Thariq's post, and added a follow-up learning loop on top, going one step further than /eli5.