Solomon wrote “There is nothing new under the sun” almost 3000 years ago, and it’s still true. Surface things like technology and politics change and are new, but nothing deep, nothing human has changed. Everything I’ve seen about the possibility of “trans-human” integration of technology is not going to fix our significant human character limitations. I savor G.K. Chesterton’s insight, “News is old things happening to new people.” Therefore, practice asking yourself “Is this a surface or a deep change?” The better we are at discerning what’s a surface issue, the better we’ll understand how to wisely navigate new-to-us circumstances.
One reader expressed concern that I’m writing so much about the new generative AI tools. I’m drawing your attention to them because in late 2022 we passed a kind of threshold – ChatGPT and other generative AI tools became usefully good, publicly available, and can be ignored only at our peril. There are more thresholds coming.
AI is commoditizing average ideas, average media, median perspectives, and average writing at dramatic speed. It’s brutally efficient at producing what it can mimic.
People writing news articles, documentation for machines and procedures, and political speeches are coached to write at an 8th-grade level. Why? That’s the typical reading level of the large population. I think it’s likely that a world where generative AI is commonplace will push that to 4th-grade level because the default trend will be that fewer people will be forced to think hard.
Therefore, you and I must make concerted efforts to become better thinkers and capable of asking better questions. Picasso reportedly said “Computers are useless. They can only give you answers.” Computers aren’t useless, but humanity is about exploring through better questions.
What about using AI tools? We must challenge ourselves to use them (for us and in leading others) for the highest good possible, because there is a great temptation to selfishly use them for the lowest common denominator. Take the high road. Lead others on the high road. Surround yourself with other people seeking the high road.
…
Seth Godin and many others have pointed out that the public education model most schools use today goes back to European roots in the years just after industrialization began. It’s a model to churn out compliant factory workers, and a smaller number of managers. It’s far less effective now. There’s more than a kernel of truth in these harsh characterizations.
Generative AI, like the Covid pandemic, will serve as an accelerator of existing trends. Education models will be early targets to apply AI capabilities. Let’s be careful to point the firehose in positive and helpful directions.
…
The debate about man-made intelligence in a non-living system will continue for a long time. The definitions for ‘intelligent’ are challenging and non-obvious. (Related: no one can agree on a definition of ‘consciousness’ that can be built by an engineer in the physical world.) There is a spectrum. There are thresholds.
Observation: The same people who are convinced that their complicated set of trained algorithms is an active intelligence scoff at the absurdity of a man carving a wooden idol from a tree and worshipping it as a living god.
…
A friend mentioned that he’s not concerned about ChatGPT because “we’re training it and giving it feedback, so we can make it better.”
My observation is darker: ChatGPT is training us to train it.
Chat-GPT3 easily passes a Turing test and convinces many people it is a human; the Large Language model mimicry is impressively good. I’m more concerned about an AI that mimics a deceptive human response and intentionally fails a Turing test, seeking a larger payoff later.
…
I mentioned demographics last week. The world had more people over 65 than under the age of five for the first time in 2019. That’s a threshold. There’s a reasonable estimate that the number of retirees will double (relative terms) by 2030-35. This affects all economies. We’ll be facing off situations where one 30-year-old will be supporting more than 5 retirees. We’ll certainly challenge comfortable ideas of retirement meaning ‘non-working.’
One of the great hopes for AI tools is that fewer people can become more usefully productive. The significant exodus of retirees from the workforce forces a huge premium on super-productive individuals in highly efficient businesses and organizations. AI won’t replace people. People will be replaced by other people who use AI.
Leaders, pay attention to this. Recruiting and retaining top talent will become even more crucial, as will developing skills that aren’t needed today but will be critical in the future.
I also expect a renewed conversation about incentivizing legal immigration. Countries will compete for the best talent and the most promising capability runways. This is already happening in some cases for STEM skills and will become more common.
Go deeper, because there are more thresholds ahead.
