Five test questions put to Google’s AI-enhanced search. Answers analysed to understand its weaknesses rather than strengths. And why every answer is a bit random.
I asked Google how I could clear disk space on macOS, and it couldn’t have been more helpful. But what I was told to do downloaded malware to my Mac. How come?
A meta-search for Apple’s patents on search, and turning the incessant log chatter of RunningBoard to our advantage in following uninstrumented apps.
Ranking search results by popularity is deeply flawed. Combined with new AI overviews, it’s already resulting in reduced publication, and will leave AI summarising AI and amplifying its errors.
As the web exploded in size, early curated directories were replaced by search engines. These were revolutionised by Google’s PageRank for popularity, and are now being replaced by AI overviews.
The most popular interface to Spotlight, through the menu bar, dates back 27 years to Sherlock, when mixing web and local search seemed a good idea. A great deal has changed since, but not Spotlight.
What should we learn from the “mysterious data corruption issue” which struck down Mac Pros used in editing studios?
Search is getting worse: only 2 good hits in the first 15. The time has come for independent benchmarks to provide Google and others with consumer-oriented targets.
Can you translate Selbstporträt als Fahnenträger into English, please, Google?
How can a blogger (or website author) ensure that they do not infringe copyright when using images of artworks? Here are some suggestions.
