A useful prediction gives you something to check. It identifies a change, the conditions needed for it to spread, and evidence that would strengthen or weaken the claim. You can use that approach to follow a technology, a changing occupation, or a decision in your community.
Start with a question that affects you: “Could this help with a task I do?”, “What would make this service practical here?”, or “Which skill would remain useful across several outcomes?” The goal is a better next step, with room to revise it as you learn.
Give each claim the right name
On a small screen, scroll the table sideways. Keyboard users can focus the table region and use the arrow keys.
| Term | What it tells you | What to ask next |
|---|---|---|
| Signal | A specific observation that may indicate a developing change. | Is this an isolated example? Who else sees it? |
| Trend | A pattern across observations over a defined period. | What was measured, where, and against which baseline? |
| Forecast | An estimate of an outcome, often conditional on assumptions. | What is the time horizon, method and uncertainty? |
| Scenario | A coherent account of a possible future used to explore choices. | Which assumptions make it possible, and how would my decision perform? |
The Government Office for Science Futures Toolkit combines scanning for change with methods for exploring possibilities and testing decisions. The distinctions above are a practical starting point for everyday use.
Keep technical maturity separate from timing. A working prototype demonstrates a capability under particular conditions. A paid service demonstrates some willingness to use it. Reliable operation across many locations, with support and acceptable cost, answers a broader adoption question.
Follow the chain from observation to consequence
Rewrite a headline as three short statements: what was observed; what someone infers from it; and what would have to happen for the claimed consequence to follow. This makes hidden assumptions visible.
- Find the original evidence. Open the paper, official data release, product documentation or public project record. Several articles repeating one announcement are one underlying source.
- Check the comparison. Identify the previous method, baseline year or control group. An improvement on one benchmark can coexist with weaknesses in other tasks.
- Read the conditions. Record the tested population, environment, version, geography and period. Ask which of these match the situation you care about.
- Identify the adoption steps. Look for manufacturing, cost, training, infrastructure, permissions, maintenance and user trust. A missing step can become the main constraint.
- Choose a next observation. Name a result that would change your view: independent replication, a completed deployment, repeat customers, or a measured reduction in failures.
A company is a useful primary source for its specifications and announced plans. Its own claims about universal usefulness require additional evidence. A study can measure an effect well while leaving cost, access or long-term outcomes unsettled.
Work through one claim
Describe uncertainty in useful terms
Separate uncertainty about performance from uncertainty about adoption. A device may work reliably but remain expensive to install. A service may attract customers but still depend on scarce staff. Write each uncertainty in a form that points toward evidence.
Use phrases such as “demonstrated on this task,” “announced for this market,” and “depends on this condition.” Reserve numerical probabilities for sources that explain how they were produced. A precise-looking percentage adds little when the event and deadline are undefined.
Look for evidence that could contradict your preferred explanation. A trial that stalls, a product with high return rates, or a system whose review work exceeds its time savings can be more informative than another successful demonstration. Record the observation before deciding what it means.
Build a small, repeatable watching routine
Choose three questions you care about and one primary source for each. Review them monthly or when a named milestone occurs. Keep the original source date separate from the date you checked it.
For each entry, write an observation, a tentative interpretation, the unresolved condition and your next action. Keep earlier entries so you can see why your judgment changed. When a forecast deadline passes, record the outcome instead of silently moving the date forward.
AI can help compare documents or draft questions. Check quoted passages, dates and source links yourself; ask it to distinguish source statements from its own inference. Keep sensitive workplace and personal information out of public tools.
When a decision has several plausible outcomes, use scenario planning to find actions that remain useful across them. You can be prepared without choosing a single future in advance.
Use the worksheet
Download the editable signal log. Save the Markdown file and edit it in a text editor or Obsidian. It includes instructions and space for your own notes.
Continue exploring
- Plan for several possible futures
- Assess emerging consumer products
- See how scientists test AI-assisted discoveries
- Explore the Future library
Evidence and links reviewed October 17, 2026. Future developments remain conditional on the evidence and circumstances described.