Of course some of us expected Microsoft Bing and ChatGPT to be rubbish—and we knew ChatGPT would make stuff up. Because Bing makes stuff up.
 

 

If you have a normal, functioning web crawler (or spider), there’s no way you would ever wind up with pages that have never existed. Nothing about this is normal.

The latest contributions from Microsoft’s Wayback Machine for site:lucire.com are these. On my phone, I noticed it had ranked in third place, after two framesets from the early 2000s, a page we had for Plucker for the Palm Pilot! That gives you an idea of how old Bing’s index must be.
 

 

On the desktop, meanwhile, a site:lucire.com search now includes sites that aren’t lucire.com. I guess if your index is that small now, you need to pad it out not just with repetition, but other domains. One is related to us—it’s our Dailymotion channel—but the other is totally random with no connection whatsoever. Bit like ChatGPT.
 

 

My friend Robin Capper has discovered the same, when enquiring with the new Bing about himself. It claimed to have sourced from his Linkedin—but fed him back facts that are nowhere to be found. Here’s his blog post. I like how he put artificial intelligence in quotes, since there’s nothing intelligent about this. It’s a simple text processor, but it sure gets a lot of things wrong.