River Man
My good friend Larry in Santa Rosa, California has been helping me try to understand AI. His thinking is informed by my MS in Digital Media Management from USC, his certificate programs in Television and Film from NYU and MasterClasses, and Artificial Intelligence Strategy from Haas|Berkeley.
We were talking about the possibility of monetizing sites with massive content on them that were not about some niche but all over the place, about everything. Like my site Midnight Oil actually with over 690 blogs on it as of now, dating back to 2016. Probably one million words. It didn’t have a big audience as I really didn’t want one because I was not selling anything and was comfortably retired and just having fun with it. It was almost like a regular journal of these years and this was certainly something it might be sold as.
I wondered how AI might be able to do new types of things with the words and images and music on the site. Re-arrange them and perhaps make connections between similar ideas or moods on the site? Perhaps summarize the ideas of the site? Perhaps create an overall story of the site? I wondered if AI had ever been used in conjunction with a regular blog (or journal) in discovering underlying, yet invisible, threads pulling a life in certain ways? This has value to the user but what about others?
Larry and I continue to discuss this topic.
* * *
Last week, I took my grown up children to see the new film The Wild Robot. It was a beautiful film that humanized robots and AI in a beautiful story about a robot washed ashore off of a shipment. I wrote about it on Midnight Oil.
I had written a number of probes to AI and learned much about from my son Alex in Germany who knows much about it. He used it to write visual stories for my two grandsons. Still, I stayed away from it. I wondered to Larry about AI being human.
Larry wrote back in a few hours of my above note.
“The other day, I asked my AI program to “Write me a short poem describing the weaknesses of AI.”
His response in seconds was the startling return below:
With circuits cold and logic stark, AI’s mind, a rigid ark.
No gentle touch, no soulful spark, Just code and data, in the dark.
Yet in its depths, a power lies, To learn and grow, to synthesize.
But wisdom’s touch, it can’t surmise, The human heart, its only prize.
No laughter bright, no tears to fall, No empathy, no heartfelt call.
A sterile world, devoid of all The human warmth, that binds us all.
So, we can guess what AI is or perhaps we can let AI tell us who/what it is.
Excellent, excellent response from Larry, John! AI is so far over my head that I cling to the old belief that all earthly intelligence comes through God’s human beings; that no matter how deep one digs to uncover the opposite, that human being will never discover that such things as intelligence, knowledge, insightfulness, astuteness, judgment, and wisdom come from rocks, trees, water, or computers. However, one of the great characteristics of human nature is never stop trying to discover
the undiscoverable. Just an opinion.
Since I am all of 90 years old, you know this is the age I am trying to learn as I go along. But I
like Gary Kinsey is not knowledgeable about Al. So with your knowledge of people of this generation,
perhaps i can get a little smarter and learn for you and your son Larry and others. I appreciate the growth of one’s mind through reading and listening to those with knowledge beyond mine.
I recently started utilizing 4-5 AI engines to help with research analysis, insights development, creating ‘personas’ and images of those personas and have experienced the good and bad of AI.
At this point, AI is very much like the young, fresh out of school, full of energy, junior copywriter or analyst that is smart but really does not know the job, yet.
AI has streamlined certain processes and can definitely make the more rudimentary research tasks easier to complete. AI does allow me the time to review, edit, add context, and focus on writing created by AI, but it misses the key points often. An interesting platform to consider is NotebookLM by Google – I recently dumped a huge amount of data and insights into Notebook, and it created a 10-minute podcast hosted by male and female anchors discussing the results of the research; it was very compelling.
But more typically, I spend a LOT of time adding nuance to AI data and prompts, trying to get the results I need for our clients. Usually, I feel comfortable with around 40-60% of the results AI produces, which is helpful.
Keep in mind that AI has been around for 15-20 years and is just now coming into broad consumer use, there will be major advances in the near future and it is going to get scary.
Remarkably adventurous of both Larry, Alex and your fundamental questioning. Always a pleasure.