I really enjoyed this Jim. Great analogy. I used to be this person too until I woke up.
I think the deeper layer here is that the industry has defined intelligence simply as knowledge. Yet the word is about the application of knowledge. And these machines cannot do this.
This is not a miscommunication because AI companies want to be regulated like the nuclear industry. They genuinely believe that the technology they are working on has the potential to become dangerous and powerful quickly. In straightforward terms, they believe they are sitting on an existential threat to humanity more severe than a thermonuclear exchange.
The “existential risk” view predates LLM technology and was well known amongst AI researchers. The trajectory of LLM development vindicated many of the technical arguments of this view and elevated the non-profits founded with this view to prominence.
So a lot of what you’re hearing is activists and non-profits advocating for government regulation of a dangerous technology, not company executives talking to customers. An important thing to keep in mind when listening to public statements from AI organizations is that OpenAI was one of these non-profits until Altman took it over (in dramatic fashion). Altman had to conform to the “existential risk” view before the takeover, and no transition is ever clean, so that’s why there’s “humanity” language mixed in with regular product marketing.
Note that since the takeover OpenAI has been fighting against the regulations supported by all other US AI organizations.
Great piece but I’m left wondering, how does the current media landscape inform this metaphor. In the days of TMI there was no 24/7 cable news, social media and a million influencers all screaming their arm chair subject mater expertise louder into the void. Grappling with the facts of Ai disruption at the human level has to compete in an attention economy with no attention left to give.
It might be more similar than you think; The China Syndrome was released right before Three Mile Island happened and a lot of attention was being paid to how much of the fictional movie was coming true which took attention away from the facts on the ground, especially since the movie inspired fears of a coverup.
In hindsight we know that it’s true that no one was hurt and no environmental contamination occurred, but at the time I don’t know how anyone could have convinced the public of that. Some things don’t leave the mind once mentioned; in the modern day the main defense against social media rage/conspiracy bait is to not see it.
Very cogent read on the current state of AI communications. Thanks. Aside: if I could I would buy up all of the IP that Moore has disowned (includes The Watchmen) and hand it over to him. He deserves it.
From the perspective of an outsider for which all of this is glaringly obvious, the situation seems like the Hindenburg/Titanic/Titan/nuclear folly all rolled into one. That these people fail to see it despite having the greatest supposed intelligence ever at their fingertips (and people literally telling them to their face and online) seems mind boggling.
I really enjoyed this Jim. Great analogy. I used to be this person too until I woke up.
I think the deeper layer here is that the industry has defined intelligence simply as knowledge. Yet the word is about the application of knowledge. And these machines cannot do this.
https://smudge.com/journal/a-history-of-ai-as-manufactured-expert-judgement/
This is not a miscommunication because AI companies want to be regulated like the nuclear industry. They genuinely believe that the technology they are working on has the potential to become dangerous and powerful quickly. In straightforward terms, they believe they are sitting on an existential threat to humanity more severe than a thermonuclear exchange.
The “existential risk” view predates LLM technology and was well known amongst AI researchers. The trajectory of LLM development vindicated many of the technical arguments of this view and elevated the non-profits founded with this view to prominence.
So a lot of what you’re hearing is activists and non-profits advocating for government regulation of a dangerous technology, not company executives talking to customers. An important thing to keep in mind when listening to public statements from AI organizations is that OpenAI was one of these non-profits until Altman took it over (in dramatic fashion). Altman had to conform to the “existential risk” view before the takeover, and no transition is ever clean, so that’s why there’s “humanity” language mixed in with regular product marketing.
Note that since the takeover OpenAI has been fighting against the regulations supported by all other US AI organizations.
Great piece but I’m left wondering, how does the current media landscape inform this metaphor. In the days of TMI there was no 24/7 cable news, social media and a million influencers all screaming their arm chair subject mater expertise louder into the void. Grappling with the facts of Ai disruption at the human level has to compete in an attention economy with no attention left to give.
It might be more similar than you think; The China Syndrome was released right before Three Mile Island happened and a lot of attention was being paid to how much of the fictional movie was coming true which took attention away from the facts on the ground, especially since the movie inspired fears of a coverup.
In hindsight we know that it’s true that no one was hurt and no environmental contamination occurred, but at the time I don’t know how anyone could have convinced the public of that. Some things don’t leave the mind once mentioned; in the modern day the main defense against social media rage/conspiracy bait is to not see it.
Very cogent read on the current state of AI communications. Thanks. Aside: if I could I would buy up all of the IP that Moore has disowned (includes The Watchmen) and hand it over to him. He deserves it.
From the perspective of an outsider for which all of this is glaringly obvious, the situation seems like the Hindenburg/Titanic/Titan/nuclear folly all rolled into one. That these people fail to see it despite having the greatest supposed intelligence ever at their fingertips (and people literally telling them to their face and online) seems mind boggling.
Brilliant, brilliant piece.
So you're saying .... the industry needs an anti-antichrist awareness campaign?
Great article!