AI is unlike technology advances that have come before.
Yep. Sounding the alarms.
Like most tech professionals my age, we lived through every remarkable step-change and wonderful advancement in tech over many decades. I literally learned computer programming on keypunch card decks we had to feed to the County mainframe overnight. (Fortran IV, thank you.) As a side note, I was one of the only female students in my high school class to take the inaugural class in “computer programming.” That was 1977. It was a milestone in our high school’s history.
From there, I lived through the IBM-dominated mainframe computing era to the client/server revolution, to the more disruptive PC revolution. Tech advances continued to the shift to mobile, SaaS, and both Internet revs 1.0 and 2.0. Everyone my age in tech lived through these transitions.
Although these shifts always move chess pieces on the game board, they were not viewed as dangerous or lethal. They simply realigned roles and positions. People employed in old skills, retooled and learned new skills.
But, dear friend of the ITSinsider blog, I’m telling you…as a friend, AI is different. It’s not the same as what’s come before.
This blog has been always been subtitled, “What’s Next in Tech.” I stopped writing here a few years ago. I’m rethinking that decision now with the pace and scale of AI advancements.
According to the research firm, CBInsights, there are 117 AI startups right here in Austin in various stages of growth. Yesterday’s “duty to warn” letter from leading tech, political, and academic luminaries probably won’t stop progress, but it shot an important red flare distress signal out to the world.
I’ve been checking in with friends around the techosphere, to take their temperature on how AI is sitting with them.
Yesterday, for instance, I caught up with my old friend, colleague, and thought leader on Future of Work, Dion Hinchcliffe. I’ve always known Dion to be a techno-optimist like I am. Dion has been experimenting with GPT-4 on his own and has been unsettled with the results. “I tested it using knowledge it can’t possibly know, and it does what I asked.” he said. “I will say to myself, ‘It will never pass this [test],’ and it does– in a couple of seconds.”
He told me there’s no putting AI back in the can. He shared that he’s most concerned about the Python runtime AI that is fully firewalled. The AI can not only write the code to answer questions, but it can execute that code as well. The fact that the AI can reason with fully working knowledge of a subject is what’s most alarming, he said.
The era of AI is upon us. As someone who’s been openly optimistic about using technology to improve life on the planet, this has my full attention. I will be checking in with more friends around the globe, and looking into some of the companies making moves in this space. Feel free to tell me your thoughts here or on social channels where we are connected.