Finding the Way to AI: The Speed of AI
By Ben Dyer
ChatGPT was released on November 30, 2022. It has entered the public consciousness so rapidly that it’s hard now to recall the “early times.” Alan Turing began working on the concept of machine intelligence in WW II and gave his first lecture on this topic in 1947. The Turing Test came along in 1950. In the intervening decades since the 50’s, considerable research and experimentation advanced this field of computation to what has now become something available to the masses.
This month of November has already been momentous for AI. I have often joked that if you have an AI application in mind, you should act on that before dark or risk seeing someone else own it. It’s evolving at a pace that requires great genius to be a leader in the realm. And, that genius has to attract matching human and capital resources.
The OpenAI user meeting in San Francisco on November 6 brought a string of surprises that reshaped many business models. The leader of the AI pack announced an ability for users with no special training to create their own chatbots, and they’ll even have an App-like distribution plan for those who want to make theirs public. This was hailed by many as killing the businesses of droves of startups and developers working on interesting new use cases. With no warning those cases became ready-for-the-market features of OpenAI’s core offerings. To top that, OpenAI said this is just scratching the surface of what they have in their pipeline.
One prominent VC quickly noted that their AI investments would henceforth be de-risked by steering well clear of anything that might be potentially subsumed by OpenAI. I’m actually wondering how one defines safe bets in the face of a few overpowering companies in this market. Competition is fierce. Moats are becoming harder and harder to claim.
There is one catch. If you create your own chatbot, you are responsible for how it behaves. Data needs to be protected, and actions must be according to your blueprint. The world doesn’t need another unfettered social media tsunami.
***
A couple of days prior to November the White House issued an Executive Order to set the stage for proper regulation of AI. You can read that at the link below:
This laudable effort begs many questions. Who is qualified to be in charge of such an executive order? Can any governmental institution keep up with the pace of innovation in AI? Regulating something that happened a year ago is not going to be very fruitful. This order was prepared in collaboration with the private sector Oligarchs of AI and computation; determining exactly how this firepower will be put fairly to good use is a monumental task.
***
A highly interesting treatise by Bill Gates on November 9 addressed the nearly unlimited potential of AI “agents” to make our lives much easier and more productive. He explores a range of areas from medicine to education to business and more and paints a rosy picture of where AI will take us in 5 years. Our ways of interfacing with AI capabilities will be quite different when these agents take over.
Look at this newly revealed Humane for a wearable that essentially is a way to talk to your hand for everything you might otherwise do on your iPhone. It’s pointing the way for the agent society as envisioned by former Apple engineers.
***
Finally, for today, here’s a video clip from Elon Musk from Rishi Sunak’s November 2 AI Safety Conference at Bletchley Park in the UK.
https://youtu.be/A2IrvS9qsQg?si=BrxTZlhoORA-a1ff
Musk is not alone in predicting that a world that has far less need to human workers and provides a stipend for everyone to do what they please for the enrichment of their lives. For him this is a prediction dating back several years, but it got renewed exposure at Bletchley Park. Would you like never to have to work? (I would not.) Who would calculate the distribution of incomes? What about the human labor of many service jobs that may be augmented by AI but not completely replaced? I’d say this one is a hot potato, but AI can be self-correcting on the fly, and who knows what will ultimately be possible?
Now, what will December bring?
Let’s chat more, message me in the Finding the Way Community
Ben Dyer is best known as the founding CEO of Peachtree Software and has been responsible for numerous startups in both Atlanta and Austin ranging from technology to financial services. He is currently an Entrepreneur-in-Residence at the ATDC at Georgia Tech and spent 7 years in starting in 2011 in similar roles at the University of Texas at Austin. He is a prolific writer and has written many hundreds of blog posts on entrepreneurial topics. Among numerous honors, he is a member of the Georgia Technology Hall of Fame.
