A Starting Over After 60 Blog
I spent the weekend with ChatGPT. We are no longer speaking.
As I’ve shared before, we have bought land on which we had planned to build our age-in-place home. As we’ve learned more about the terrain and elevation, we’ve had to reconsider our original plans. We were designing a home on a flat surface, but the best place for our new home would be an embankment. Rather than a flat-surface barndominiun, we would need a bank barn-style house. I have struggled to explain my vision. So, I turned to ChatGPT for help.
At first, she (it seems like a she) was like any new friend, enthusiastically engaged, eager to see please, understanding — obviously, I’m anthropomorphizing. She listened to my detailed requests and responded with renderings close to my vision. But as I asked more of her, she faltered. She made mistakes repeatedly or missed instructions completely. Even though I asked her to tell me how best to provide her details, she couldn’t do it. She strayed further and further away from my vision.
This is the difference between a bot and a human. The human learns in real time and adjusts— at least most of the time. Humans also can converse, allowing immediate and iterative changes. A human understands, “use this version exactly and add a door to the bedroom.” ChatGPT doesn’t. Most likely, the human wouldn’t miss the need for a door to begin with, saving my need to retype the same instruction over and over until giving up.
AI has its place in work. It’s helpful with quick research, summaries, and developing charts and tables. This creates efficiencies and shortcuts to speed production— though human productivity is higher than ever. I see AI as an aide for humans, something like the lifeline from ”You Want To Be A Millionaire.” It’s your encyclopedic, smart friend who can solve an equation and identify a source in seconds. What she can’t do, however, is listen, at least not in a timely and meaningful way. My house project could have been solved with one conversation and a few hours of work rather than 2 days of trying to get a bot to remember to put a door in a room.
This, perhaps, is my biggest surprise: ChatCPT wasn’t timely. Nor was she capable of the nuanced, thoughtful engagement that leads to mutual satisfaction and achievement. She does some initial, elementary work, but she doesn’t get the job done.
Apparently, I’m not alone in my criticism. Some companies who jumped on the AI choo choo, firing thousands of workers and replacing them with AI systems, are reversing themselves. Salesforce fired 4,000 staff, expecting AI to automatically and immediately fill the void. Instead, they received thousands of complaints about poor service and their customer trust and satisfaction declined. “We assumed the technology was further along than it actually was,” Salesforce said, noting that AI models don’t have the problem-solving capabilities humans have. Now they are rehiring. Nvidia is sounding alarm bells too. Their employees followed instructions and adopted AI in their day-to-day work. Now, Nvidia is pulling back, complaining about cost. “For my team, the costs of compute [sic] is far beyond the costs of employees,” said Nvidia VP of Applied Deep Learning Bryan Catanzaro.
In other words, executives are learning what their now former employees tried to tell them: AI isn’t ready to replace humans. Think of the costs companies spent terminating employees, integrating AI models, and then rehiring and training new employees. They certainly didn’t save money as projected. Their employees knew that; they tried to tell management. As expected, executives didn’t listen.
I have an even bigger question: why do we want to replace humans with bots? Should machine capability exceed the need for humans? What are humans supposed to do once the bots take their jobs? I’d love to hear one politician or social leader raise this question. They are too busy jumping after the wealth the AI lemming craze has caused for the top 10%.
Many European nations are slowing the integration of AI. So is China. In fact, China has determined that AI cannot replace workers. Who’d have thought China is more compassionate than America? For China, this isn’t only a humane approach but also an economic one: large-scale automation may initially save some costs, but the longterm harm would take decades to undo. Mass unemployment leads to economic imbalance, reduces consumer spending and increases social discontent, drug use and crime. Take a look at any number of American small towns whose manufacturing was sent overseas and replaced with nothing. Opiate and meth use have been epidemic in these towns, resulting in poverty and poor health. The race to outsource killed small town America.
Now the U.S. plans to use AI TO destroy the middle class and entry-level jobs. The new opportunity? Low paid would-you-like-fries- with- that jobs. I’m seeing this firsthand. My daughters are in the upper wedges of Gen Z. Service jobs abound. Entry level, not so much. Their highly educated friends are working in service while they search for “real jobs.” I haven’t the heart to tell them those jobs may never appear.
The U.S. has a long history of taking work away from Americans. Manufacturing went to machines and then overseas. Automation takes the jobs now and, I’m sure, will be sold overseas too, never to return. The top 1% and 10% benefit. People like Charlie Kirk have been telling young people not to go to college. But decent middle-class, blue-collar jobs don’t exist in America anymore. Meanwhile, Big Pharma is on the sidelines with another “miracle drug” like OxyContin in hand..
I’m getting off topic… I had meant for this blog to be a lighthearted story of my brief friendship with ChatGPT, ending in my dissatisfaction and frustration. Ha ha. Isn’t it funny? But it’s not. And I can’t. As with many new relationships, I hoped for more than I got and put out more effort than she did. I spent the weekend with AI and emerged with another rant about the state of our country and the lack of care for our people. Move to China, you say? Maybe I will.
