A.I. mystery solved by border collie
A thought experiment in whether (and what it would take for) a robotic canine could be taught to herd sheep — 45 minute comedy video introduction to the new Planet Waves FM.
This email is from Planet Waves on Substack, open to all readers reglardless of subscription status. All materials included here are also open access. You know that we are sustained by your subscriptions. Thank you my customers, astrology clients, subscribers and the many who spread the word and help out in ways large and small.
There are numerous people in the audience who have followed me for many years because you value my persistent, consistent rational presence in these chaotic, isolated times. You read me because you trust me and the integrity of my work in all its forms. You can do your part to make sure that I’m here for you and can cover the costs of the services I provide.
Subscribing to this Substack grants access to my Thursday horoscope and article. There are also several options to sign up directly through Planet Waves.
And now for the greatest question ever asked about A.I. technology:
A simple thought experiment breaks the concept of “artificial general intelligence” — could you train a perfect border collie robot to herd sheep? This is one task, and the one thing besides wag their tails that border collies are born knowing how to do. What would it take? How many million lines of code?
We think A.I. can replace human customer service reps, fly jet planes, solve climate change or design new useless drugs. But can it do the work of a six-month old puppy?
Below is a link to the full program — one of my best ever — the second half which I devote to the Woodstock Festival anniversary, which is this weekend…I tell many stories from the different festivals and the people I met covering them as a young reporter for Woodstock Times. For fans or scholars of the 1960s, here is my tribute to festival founder Mike Lang.
Below again is my recent letter associated with Ice Nine News, our daily effort to collect the latest about the A.I. train wreck. Yesterday I was emailing with my old friend Doug Harper, author of the Online Etymology Dictionary, about how “large language models” like ChatGPT have devoured his life work in one bite without compensation. Every last thing in an A.I. system was stolen from its authors.
It has taken Doug decades to write this dictionary — the best I’ve ever seen and the only one we can attribute to one author — and now Google has replaced it with its “A.I. search result,” pushing it lower, stealing Doug’s work, and choking off his advertising revenue. And this is only one part of the problem — in the end, A.I. is a total fail except for making money on fake girlfriends and cheating on school papers.



