Knowledge engineering - Category Home
Knowledge engineering - Topics
Modern knowledge engineering is prompt design and fine-tuning: same bottleneck new costume
Mapping where the expertise actually lives in a 500-person company before the retirement wave hits
Ontologies promise machines that understand meaning and deliver spreadsheets nobody maintains
What the knowledge engineer actually did: sitting between the domain expert and the rule base
Expert systems rose on promises and collapsed on maintenance: the economics of encoded expertise
MYCIN outperformed junior physicians but was never deployed: what that says about knowledge systems and liability
Eliciting tacit expertise: why the best surgeon can't explain the wrist motion that saves the patient
Drawing the knowledge structure of a domain by hand: what card sorting and laddering interviews reveal that surveys don't
Cyc has been hand-coding common sense since 1984: what it got right that LLMs still haven't
The knowledge acquisition bottleneck: experts can't articulate what they know and interviews only scratch it
Text mining is not data mining with different punctuation: what changes when the object is language
KCS methodology says write the KB article while solving the ticket: why do support teams still resist
Repertory grids and card sorting: the elicitation techniques that survived the expert systems winter
Relation extraction from text has improved tenfold yet KG construction is still mostly manual curation
The tribal knowledge that never gets written down is precisely the knowledge the KB needs most
Polanyi's "we know more than we can tell" is the wall every formal representation hits
The brokers between teams hold disproportionate knowledge: how do you find them before they burn out
XCON saved DEC millions and then became unmaintainable: the original sin of rule-based knowledge systems