Dang! When I saw this was by Stephen Fry, and was a transcript of a recent lecture he gave at King's College, I went hunting to try to find a video of it... since not only is he brilliant, his...
Dang! When I saw this was by Stephen Fry, and was a transcript of a recent lecture he gave at King's College, I went hunting to try to find a video of it... since not only is he brilliant, his speaking ability and voice are also absolutely wonderful. But sadly it doesn't appear they have released a video of it (yet?). Hopefully they recorded it and plan on releasing it to the public later, because what a profoundly wasted opportunity that would be if they didn't, or don't plan to release it if they did. :(
Stephen Fry makes some comparisons between artificial intelligence and people, but I think a more relevant comparison might be to machines. We have pretty high standards for the predictability and...
Stephen Fry makes some comparisons between artificial intelligence and people, but I think a more relevant comparison might be to machines. We have pretty high standards for the predictability and reliability of machines, particularly when safety is on the line. Based on that, we tend to assume that programs that use an LLM as a component are or should be reliable, but they usually aren’t.
It’s going to take a lot of work to figure out how this messy and expensive technology can be put to good use. As someone on Twitter said recently, “there are no AI-shaped holes lying around,” places where you can just plug an AI in and turn it on. AI Chat lacks instructions for specific uses. You’re just supposed to figure it out.
I think this means it’s going to take time before we see big disruptions. Like driverless cars, you can see it coming for quite a long time. People are going to get bored (already are getting bored), so they move on.
When we read history or watch movies, we often start thinking of long periods of time as if they were short, because we don’t have to sit through all the boring bits. A montage stands in for time passing.
Dang! When I saw this was by Stephen Fry, and was a transcript of a recent lecture he gave at King's College, I went hunting to try to find a video of it... since not only is he brilliant, his speaking ability and voice are also absolutely wonderful. But sadly it doesn't appear they have released a video of it (yet?). Hopefully they recorded it and plan on releasing it to the public later, because what a profoundly wasted opportunity that would be if they didn't, or don't plan to release it if they did. :(
Someone wrote in a comment that the video should show up on Linkedin:
https://www.linkedin.com/showcase/kingsdigitalfutures/
Ah. Cool. Thanks for the heads up. I'll try to keep an eye out for that post.
I did the same, but the talk was only a few days ago. Maybe it'll come out in the coming weeks.
🤞
Stephen Fry makes some comparisons between artificial intelligence and people, but I think a more relevant comparison might be to machines. We have pretty high standards for the predictability and reliability of machines, particularly when safety is on the line. Based on that, we tend to assume that programs that use an LLM as a component are or should be reliable, but they usually aren’t.
It’s going to take a lot of work to figure out how this messy and expensive technology can be put to good use. As someone on Twitter said recently, “there are no AI-shaped holes lying around,” places where you can just plug an AI in and turn it on. AI Chat lacks instructions for specific uses. You’re just supposed to figure it out.
I think this means it’s going to take time before we see big disruptions. Like driverless cars, you can see it coming for quite a long time. People are going to get bored (already are getting bored), so they move on.
When we read history or watch movies, we often start thinking of long periods of time as if they were short, because we don’t have to sit through all the boring bits. A montage stands in for time passing.
Full video of the lecture was posted to YouTube:
AI: A means to an end or a means to the end? with Stephen Fry
cc: @teaearlgraycold, @skybrian