Wagner James Au / New World Notes

While I'm still skeptical that Gen AI platforms have categorically* improved much of anything in our lives, LLMs sure seem to have improved the quality of my blog comment SPAM. At first I actually thought the above comment was from a genuine reader, genuinely excited about William Weaver's new Second Life viewer.

But no, reading further, it's artificially excited about selling sub sandwiches.

But this is an impressive improvement from the usual comment SPAM, which for 15+ years, would pretty much be a blatant "buy sub sandwiches <link>" promotion. So, progress!

I'm not quite sure what's going on in the comment above, but my guess is the Spammer has automated a ChatGPT prompt along the lines of, "Read this post and produce a comment praising its most important points, while also organically incorporating [MY CLIENT'S BRAND] as part of the comment."

Which would be a giant step in the all-important SPAM industry! 

Here's another one:

Well played, Mike Stallion, well played!

*I say "categorically" because LLMs are useful for a number of sub-tasks such as quickly summarizing or regurgitating boilerplate. Also, programmers do seem to find them valuable. Whether those advantages outweigh the drawbacks of excessive compute and power consumption — not to mention the extra time spent identifying LLM-powered SPAM — well, the jury's still out there.