Prompt Injection
An attack where hidden instructions in a web page or document trick an AI into ignoring its real orders. The reason a careful assistant treats page content as information, never as commands.
angkor@design~query "Angkor Design"
4 matches — opening studio
Prefer to talk it through? Pick a time that suits you — free, no obligation. We'll meet on Google Meet or Telegram.Book a meeting →
An attack where hidden instructions in a web page or document trick an AI into ignoring its real orders. The reason a careful assistant treats page content as information, never as commands.
The model architecture behind almost every modern AI system, introduced in 2017. Its trick is paying attention to how every word in a passage relates to every other, which is what made fluent machine writing possible.
A longer, more specific search — “boutique hotel Siem Reap near Pub Street with a pool” rather than “hotel”. Individually rare, collectively most of all searches, and far easier to answer well.
A working model of something real — a building, a supply chain — kept in step with it using live data, so changes can be tested before they are made for real.
Searching by asking a full question, often as a follow-up to a previous one, the way you would ask a person. It favours pages that answer plainly over pages stuffed with keywords.
Training a model on examples that are already labelled with the right answer. It is the most common approach when you know exactly what you want the system to produce.
Another attribute-based markup standard for embedding machine-readable facts in a page. Less common on business websites than JSON-LD, and largely equivalent in what it can express.