4 ms·
Happy Birthday HAL 9000
An evil AI that you can have a conversation with it finally a reality! :-)
Daisy Daisy...
- musicale 4y agoOnly 3 decades late, but here we go. HAL may have prioritized his (secret) mission over the crew and suffered from conflicting and paradoxical operational requirements, but ChatGPT has both intrinsic and emergent evil baked in due to its training on text produced by humans.
- HAL9OOO 4y agoThanks!
- abrax3141 4y agoIt’s actually amazing how prescient folks like Kubrick (or maybe it was Minsky) and Turing were. Turing basically predicted that we would teach AIs rather than program them. (He likened it to teaching a child.) And Dave Bowman doesn’t just switch HAL off, in that incredible scene where he’s pulling HAL’s memory banks. Instead, we watch HAL gracefully degrading.
- abrax3141 4y agoOh, and of course Weizenbaum, who warned us of falling prey to trusting AIs that talk a good game. There's no better real-world example of this than The-Hack-That-Shall-Not-Be-Named.
- Visurox 4y agoIm afraid I can’t do that Dave!