Trust but Verify
Ronald Reagan
As you know I’ve had some fun with AI helping it expand my Hogan’s Heroes/Rat Patrol crossover and playing 2000 year old man with Christianity and Bacon but a lot of people are using AI for serious stuff.
Today I saw the following tweet in X concerning AI and the COVID vaccines: (Here is the text in case the tweet gets zonked:
Brubes Brubes793 Jul 27
@grok says there’s nothing to see here. The vax was wonderful. Any bad outcomes were outweighed by the positives… Like getting C19 anyway and spreading it anyway. Well worth the risks. Consensus science says so. Replying to @Brubes793 @USMortality and @grok
I was about to reply when it hit me that said reply is a basic fact that is one of the most important things I’m likely ever going to tweet so it belongs in the blog first.
Presenting DaTechGuy’s 1st Law on Using AI (Must Credit DaTechGuy)
“Unless you’re doing something that’s pure whimsy, When you ask any AI any question always tell it to show its work & links to the sources supporting their answers to your questions”
Until you know what the sources for this information you have no idea WHY an AI is saying what it says. For example if you asked an AI if Joe Biden was healthy enough to be president 15 month ago odds are it would have told you of course and the GOP stuff is just deep flakes. If you asked it about the Russian Hoax six months ago it would have supported it because the MSM pushed that narrative.
Once the cost of creating propaganda became negligible (internet age) the source for any information had to be questioned.
So follow this law when you use AI for anything more serious than asking which person from the Bible are you: (Grok says I’m Baruch the scribe of Jeremiah).
While I remain the first member of new media to question Donald Trump in a press conference setting odds are If I get remembered for anything this is likely it.
Update: Here is the tweet sent after the blog post:
Da Techguy's (no longer backup) Blog 