Greg Dorfman of Budget Blinds looks over an AI-generated program he has been working on to help answer typical
If you want to stay a skeptic about the various technologies misleadingly called “artificial intelligence,” you probably shouldn’t chat with Greg Dorfman during an AI presentation.
“It’s a real tool and I’m hoping to make it a better tool,” Dorfman told me in no uncertain terms. “It has become an everyday thing for me.”
Dorfman’s enthusiasm for AI carries weight because he’s neither a software guy nor somebody trying to goose his Anthropic stock. He is the general manager of Budget Blinds in Concord, a midsize business with $3 million-ish annual sales and a dozen employees, and he works with customers to design and sell window covers of various types, sending out field crews to install them.
It’s the sort of business where I’d expect to see a fax number on the business cards rather than AI on the computers.
Dorfman said he started experimenting with the technology around the start of the year after his wife began using it at home. He began with simple ChatGPT uses, such as turning his ideas into an agenda for morning staff meetings, but quickly moved up the complexity ladder.
His latest use is an executive coaching module that helps determine the key points of each job in a business and resulting metrics for employees, which he said should speed up the most important task of an office manager, dealing with staff, so he can do the most important task of a business manager, finding customers.
“It lets me spend time working on the business, instead of in the business,” is how he put it. His goal is to train a specific AI agent, or self-contained unit, to help with or handle new customers when they call in for an initial consultation.
I met Dorfman at a conference hosted by the Concord Area Chamber of Commerce on AI for small businesses. The presenter was Ryan Robinson, chief service officer for Mainstay, a company that sells AI tools and training to individuals and businesses.
Robinson is, as you would expect, gung-ho on the technology. He argued that AI modules for businesses, usually known as agents, are already the equivalent of human interns: You need to hold their hand and triple-check their work, but a good one can be helpful.
Soon, he said, they’ll be the equivalent of employees who can get a job done and do it well with a minimum of oversight. And eventually, they’ll be specialists: Give them a goal and they’ll create the job needed to get there and will do it, over and over for as long as you want, even interacting with other agents to expand the scope of their work.
At that point, he said, the technology will allow a re-imagining of business practices just as the adoption of electricity a century ago allowed a re-imagining of manufacturing processes. After that, he said, the sky’s the limit.
As I indicated at the start, I am skeptical about large-language models and other types of machine learning that get the “artificial intelligence” branding. I’m an old guy who has seen so many technologies explode through the hype cycle before settling down to ordinariness or even disappearing altogether that I discount every earthshaking new tech.
My guess, for what it’s worth, is that AI will end up more useful than NFTs but less important than HTTP. A lot of smart, experienced people disagree, however.
There are good reasons to think that AI is different than past hyped technologies and will have sweeping consequences for business and governance and daily life, just as the internet has. In that case, you’d be a fool not to throw all the money and time you can spare at the technology because you’ll be left in the dust otherwise.
Maybe. But in the meantime, the industry has a whole bunch of downsides.
It is upending our feeble efforts to make the electric grid less toxic, giving corporations an excuse to crack down on or just plain fire workers, larding every piece of software with annoying pop-ups begging you to use their new AI and potentially creating an investment bubble that will clobber our already-Trump-rattled economy.
So, that’s not good. Is it worth it? Should we start reining in AI before it’s out of control or give it all the room we can so it can reach its glorious potential?
I realize there’s nothing lamer than ending a column with “time will tell,” but I’m afraid this is a case where time will, indeed, tell.
Note: No large language models were used in writing this column. Honest.
Read more on SentinelSource.com
