I'm starting a newsletter about practical uses of #AI. I hope to convince everyone that AI is useful, even if it isn't perfect. By learning how these models work and what they excel at, we can all free up some time in our busy lives.
christophermoravec.com/p/a763b209-6...
Posts by MoreHavoc
KM3NeT xkcd.com/3053
Radon xkcd.com/3037
Also if an AI could just admit it doesn’t know instead of making stuff up, that would be really helpful.
This is fascinating. Even as a human it can be hard to know if I’m right about something. My example is a recent debate with some friends about if eating snow will dehydrate you more than the water it provides. An AI that could fact check better than a human would be a powerful tool.
After further interrogations and suggestions (thank you for the thoughts everyone!)…
In the spirit of the Turing Test, I propose the Cripslock Test: No system can be called AGI if that system cannot reliably determine whether or not it is spewing absolute bullshit.
(Named for Sacharissa Cripslock)
2025's mission objectives:
• Launching a newsletter about AI that won't tell you to put glue on pizza (the newsletter won’t , AI still might) 📨
• Making more weird and wonderful things 🛠️
• Helping shape AI into something genuinely useful for humans
Sometimes a temporal landmark is a great excuse to reflect and think. 2024 was a .. year, but there were some good things:
• Gave my first keynote 🎤
• Validated my feeling that most Humans don't want AI hype 💯
• Built things, broke things, learned things (mostly in that order) 🥽
During my holiday break, I made a button to disable our Pi-Hole and I just got the video posted, check it out: youtu.be/h51sABc1pn0
Just got my latest project video posted! For Halloween this year, I made a giant snake (Named Señor Serpent) to dispense candy, and it was by far my best one yet! youtu.be/yx-XmNHf8fQ
#maker #halloween #youtube