A demonstration of three simple ways to add invisible text to a PDF that could be used to prompt AI how to review it. And a method of detection.
AI
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?
What was the key to successfully learning to drive, ride a bike, or speak a foreign language? Confidence, and it’s every bit as important in computing, and in macOS.
Log entries mentioning errors are common and can mislead. Those marking the consequences of errors may not mention the word error. Turn detective to work out where the error really is with these tips.
What happens when you can’t solve a problem, so you get a log extract and ask Claude to diagnose it for you? Here’s a full worked example.
If Visual Look Up is so easy and low-power for Apple silicon Macs, maybe Tahoe’s new Foundation Models will prove more challenging, and wake up the neural engine.
We’re in the midst of a drought in the UK. We’ve all been recommended to delete old emails and photos to economise of water use. Is that wise?
Dire warnings of the consequences of Google’s answer engine resulting in over-exploitation of the web and its exhaustion. It’s time to stand up and declare what’s written by a human, not an AI.
A mystery volume mounts from nowhere, and its name starts with Creedence. This is a cryptex, also favoured in Apple’s Private Cloud Compute. Is it a ghost of the past, or harbinger of the future?
The Eclectic Light Company 