Explaining What Is New in Frontier Tech
Writing · 10 min read ·
How to write about new technology for people who have not met it: start from the problem, define terms once, show an example and mark what is unproven.
New technology has a communication problem. By the time you understand something well enough to build with it, you have also lost the memory of not understanding it. The words feel obvious, the problem it solves feels obvious and the example you reach for is one only an insider would recognise. The people you want to reach, customers, journalists, investors and colleagues in other fields, are left outside.
This guide is about writing for them. It sets out a method for explaining genuinely new technology so that someone who has never met it can follow, evaluate and, if it suits them, try it.
Remember what readers do not know
Three gaps are common.
The problem gap. Readers may not see why the problem matters, or even that it exists. The technology only makes sense once the problem does.
The vocabulary gap. New fields coin terms quickly. Readers may know none of them, or may know a different meaning of the same word.
The evidence gap. Readers cannot tell what has been demonstrated and what is hoped for. Emerging technologies, those that are still being developed or whose practical applications are not yet fully realised, mix real results with speculation, and outsiders find it hard to separate them.
Good explanation closes each gap in turn.
Start from the problem
Open with the situation the reader already knows. "Anyone who has tried to keep a team's documents in sync knows..." Then show why existing approaches are not enough. Only then introduce the new thing.
This order is the opposite of how technologists often write, which is to start with the technology and hope the audience will find a use for it. The problem-first order gives readers a reason to continue and a yardstick: they can judge whether the new thing solves the problem.
Keep the problem concrete. A specific person doing a specific job beats an abstract description of a market.
Use one example, all the way through
A single carried example does more than five scattered ones. Choose a case that is:
- Ordinary. Something the reader could imagine doing.
- Specific. Real inputs and outputs.
- Revealing. It shows what is new, not just what is useful.
Return to it in each section: here is the situation, here is how it works on the example, here is what happens when something goes wrong, here is what you would do next. By the end the reader has seen the technology at work from several angles without having to hold abstractions in their head.
Define terms once, in plain words
When you must use a term, define it the first time, in a short clause, and use it consistently afterwards.
- Explain, do not just name. "An embedding is a list of numbers that represents the meaning of a piece of text so that similar texts have similar lists."
- Prefer existing words to new coinages. If a plain word works, use it.
- Do not change terms midway. A reader who sees three words for one thing will think they are three things.
- Keep a glossary for longer pieces, but do not rely on it; readers should not have to leave the text.
- Spell out acronyms on first use.
Plain English advice applies even at the frontier: short sentences, common words and the active voice help expert and novice alike.
Show what changes
For each claim about the technology, finish the thought with its consequence for the reader. "It runs on the device" becomes "so your data does not leave your phone". "It supports longer inputs" becomes "so you can paste an entire contract rather than a page".
The test is the "so what" question. If a sentence does not answer it, either add the consequence or cut the sentence.
Mark maturity honestly
Technologies range from ideas to proven infrastructure. Readers need to know where yours sits.
The technology readiness level scale, originally developed for aerospace, grades technologies from basic principles observed through to systems proven in real operations. You do not need to use the formal scale, but borrowing the idea helps: say plainly whether the technology is a research result, a prototype, in limited use or in wide production.
- Demonstrated: "In our tests of three hundred documents, it did X."
- Observed: "Early users report Y; we have not measured it."
- Expected: "We expect Z once the next version ships."
- Unknown: "We do not yet know how it behaves with W."
Use these categories explicitly. Mixing them is how readers end up believing hopes.
Beware the hype cycle
The hype cycle is a model describing how public enthusiasm for a new technology typically rises to a peak, falls into disillusionment and then recovers to a more realistic plateau. It is a model, not a law, but it is a useful reminder that early excitement overshoots. If you are writing at the peak, give your readers the grounding that the trough will later demand. Describe what the technology can do now, what it cannot and what would need to be true for the bigger claims to hold.
Avoid the traps of novelty
The solution in search of a problem. If you cannot state a problem that a real person has, rethink the piece.
Analogies that mislead. "It is like a brain" illuminates nothing and invites false beliefs. Use analogies carefully, and say where they break down.
Futurism as evidence. Predictions are not facts. Label them.
Name-dropping. Mentioning famous organisations that have not endorsed you is a form of false authority.
Complexity as a signal. Complicated explanations do not make a technology seem more advanced; they make it seem less understood.
Skipping the downsides. New technology has costs and risks. Mention them, and say what you are doing about them.
Structure for different readers
One page rarely serves everyone. A layered structure helps.
- A two-sentence summary at the top: what it is and why it matters.
- A short explainer for newcomers, with the carried example.
- A deeper section for specialists: architecture, benchmarks, limits, with links to documentation and code.
- A "what is unproven" section for sceptics.
Readers can stop where their interest or knowledge ends, and each layer is consistent with the one above.
Visuals
A diagram can explain what paragraphs struggle with. Keep diagrams simple: few boxes, labelled in words, with arrows showing flow. Provide a text description so that readers using assistive technology get the same information, and make sure any text in the image is also real text on the page. Avoid decorative visuals that imply capabilities the product does not have.
A worked example
A team has built a tool that lets a user search their own documents by meaning rather than by exact words. Their first draft opens: "Our semantic retrieval engine leverages dense vector embeddings to unlock context-aware knowledge discovery." They rewrite.
"You are looking for the contract clause about ending an agreement early. You search for 'cancel'. The document says 'terminate'. A normal search finds nothing. Our tool finds it, because it compares what the words mean, not the letters. Here is how it works on that example." Then: "Each passage is turned into a list of numbers, called an embedding, so that passages with similar meaning end up with similar numbers. When you search, your query is turned into numbers the same way and the closest passages are returned."
A maturity box follows: "Demonstrated: on two hundred sample contracts, the right clause appeared in the top three results in most searches, and the test method is below. Observed: early users say it saves them time; we have not measured it. Unproven: how it behaves on documents in languages other than English." The page includes a link to a working demo.
Readers who have never heard of embeddings follow the page to the end; specialists find the details they need, and sceptics find the limits stated.
Questions writers ask
What if the technology really is too complex for a short explanation? Then explain a layer of it. Say plainly what you are leaving out and where to find it. Partial, honest explanations are better than complete, confusing ones.
Should I avoid all jargon? No. Specialists often prefer the precise term. Use it, define it once and link to a fuller explanation for those who want it.
How do I choose the example? Ask three people who have never seen the technology to describe a problem in their own work that it might help. The best example is usually one of theirs, not yours.
What if I cannot show a demo? Show a recording of a real run, with the inputs and outputs, and say it is a recording. Never present a staged result as live.
How do I handle questions about safety or ethics? Address them directly in the piece, not in a footnote. State the risks you know of, what you are doing and what remains open.
A simple test before publishing
Give the draft to someone outside your field and ask them to underline every word they do not understand and every sentence where they ask "so what?". Fix each mark. Then ask them to explain the technology back to you in two sentences. If they cannot, the explanation has not yet worked. The test takes half an hour and is the quickest way to find the curse of knowledge in your own writing.
Summary
Explain new technology by starting from a problem the reader knows, carrying one concrete example throughout, defining each term once in plain words and showing what changes for the reader. Mark clearly what is demonstrated, observed, expected and unknown, remember the hype cycle and avoid misleading analogies and name-dropping. Layer the content for different readers and keep diagrams simple and accessible. Clear, honest explanation is how new technology earns its first users.
Questions and answers
- How do I explain something nobody has heard of?
- Start from the problem it solves, use one concrete example, introduce each new term once with a plain definition and show what changes for the reader.
- How much jargon is acceptable?
- Only as much as the audience needs. Define each term the first time and avoid inventing new names where existing ones work.
- How do I avoid overselling a new technology?
- Separate what is demonstrated from what is expected, give the maturity of the technology and say what remains unproven.
- What is a technology readiness level?
- A scale used to describe how mature a technology is, from basic research to proven use in real operation.