An Interesting View of AI Deployment

February 8, 2023

Let me summarize the 5,000 word essay “Let’s Speed Up AI”: Go faster. The idea is that one has to accelerate that which is already accelerating. Imagine a downhill skier with a jet pack. Once momentum is gained, fire the jet back pack. The author believes that wide, rapid uptake of AI methods will result in the types of applications and controls. I certainly am not keen on having Terminator kick in my door, but I will reserve judgment on the wisdom of this “technology will work out its kinks.” A French guy pointed out that technology has unanticipated consequences. Today it is not necessary to get one’s priestly robes in a twist.

The write up states:

This idea that we can imagine every problem before it happens is a bizarre byproduct of a big drop in our risk tolerance as a society.

Are countries with access to AI technology risk averse? That’s a question I am not able to answer. I am not sure anyone can. Perhaps it is a job for the to-be systems from Microsoft Bing or the Code Red Google?

In reference to the suggestion that AI has to slow down, the write up says:

The basic premise of the slow-down article is that AI doom is inevitable.  We’ve got to slow it down or stop the research now to avoid the disaster!  It’s not that AI might go crazy and kill us all, it’s that it will kill us all!

If you are fans of the go-fast approach to technology, you may find the “Let’s Speed Up AI” argument on the money. I think that some pharma execs like the idea of go-fast.

Is it important that Google with its oodles of smart software and thousands of wizards [a] overlooked the potential of ChatGPT, [b] noted the demand and did not know how to frame Googzilla’s response, or [c] has too many executives like Sundar and Prabhakar who want to go slow or not go anywhere that would undermine their compensation?

Google’s new posture is go fast. I wonder if the firm has weighed objectively the risks of rolling out a knee jerk response to what is the marketing home run of the last six months.

My hunch is that the battle is not among smart software vendors; I think we are engaged in a marketing tussle. Technology is like the illustrations of new cars in a 1950s’ Saturday Evening Post. The cars were not that good. But the messages and images were outstanding.

Accelerating the accelerating sounds good. Should we ask the downhill skier with the jet pack?

Stephen E Arnold, February 8, 2023

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta