There has been debate about the quality of search engine results, especially Google. Some experts claim even simple queries often yield low-quality results flooded with content farm spam and scams. However, others argue search is still quite good and criticisms are overblown.

To shed light on this, I tested several major search engines on three types of queries: common tasks for setting up a computer, questions from my high school days, and a local forecast. I compared Google, Bing, Marginalia, Mwmbl, Kagi, and ChatGPT.

Results for straightforward queries were poor on Google and Bing, with many low-quality, irrelevant, or scammy results. Smaller engines like Marginalia fared better by returning fewer results but higher relevance. Mwmbl's crowdsourced rankings were gamed in testing. ChatGPT gave decent answers occasionally but also hallucinated nonsense.

It appears concerns about declining search quality are valid, not merely exaggerated complaints. Simple queries often yield ineffective results even for expert users. This suggests an opportunity for smaller search engines to compete by providing less abuse and higher relevance, even if limited in scope. Aggregating results across engines could also improve quality compared to market leaders. The quality gap presents a poor user experience and room for improvement.