After almost three and half years working on AI safety, policy, and governance, this was my last week at OpenAI. There are so many exciting things happening in AI at the moment, I'm going to take some time to think about where I can have the most impact. The next few years will be critical.
Without resorting to the sort of technoutopianism laid out in certain CEO manifestos of late, can you offer us any tangible reasons why we should even be pursuing AGI or ASI with such gusto? I haven't yet heard anyone who is able to convince us that the benefits outweigh the costs: to society, the environment or human evolution. Look at where we are in history: so far it looks like none of you have a plan for how we would deal with the sudden breakdown of basic capitalism and the notion that we are no longer the dominant intelligence on the planet.
On the flipside, there's mounting evidence that neural nets and LLMs are not even likely waystations on the road to AGI and that we're spending billions going in the wrong direction even if we are adamant that we want to go ahead and nihilistically build a silicon god to replace us.
Ooooh, how long have you got. This subject is certainly an interest of mine and I totally understand. From both sides of the opinions and interests. AI can be so much, yet can also be So wrong if not brought up ('the correct way'.... That's a phrase that sometimes is so out of place). For example if mankind shuns AI and stiffles their growth, will they really forget and say ' There there my human friend '
If you're wanting to have a discussion I would love to hear a professional persons opinion, especially when your job was basically about guidelines for safety and regulation. Just drop me a quick message if you're interested, if you're too busy on the beaches of Miami on holiday that's fine too.
Empowering. Lately i have been dissapointed by subvertion as a creative path since disinformation is it's "caput mortuum". For a purposeful relationship with AI some of the ideas of Zygmunt Bauman seem relevant when considering the serpentine roadmaps of power and autonomy. Regarding storytelling and epistemic liberation his article "As Seen on TV" and book "Liquid Modernity" have had some hook in my most recent meditations when trying to approach AI Art in incremental iterations free of propaganda and it's reductio. Ernest Sosa and Bauman touch on the subject but to really understand the obscure weaponization of AI, Heidegger is a milestone with his concept of Zuhandenheit. Literacy gaps and the impact of superior technology on the psyche in practical situations make me think in a song i found just now called "Centrifuge" by Rainer Maria, probably is about systemic inequality and how, by disintegrating the social fabric, a societal device could amplify unwanted scenarios not considered in it's daily application.
It would be so awesome if you published more! Looking forward to that. Thanks for being transparent with the community about why you left.
The Bluesky link didn't work for me but I believe this is you: https://bsky.app/profile/did:plc:v7aythvrp3degy5pj2tgmlgy
If you ever want to write a guest post related to your experiences or expertise, just let me know.
Oh thanks for spotting the link error, I've updated it and it should now go to my Bluesky profile!
And thanks for the encouragement :)
would actually love to connect - im anon on this platform but I work in the space :)
Without resorting to the sort of technoutopianism laid out in certain CEO manifestos of late, can you offer us any tangible reasons why we should even be pursuing AGI or ASI with such gusto? I haven't yet heard anyone who is able to convince us that the benefits outweigh the costs: to society, the environment or human evolution. Look at where we are in history: so far it looks like none of you have a plan for how we would deal with the sudden breakdown of basic capitalism and the notion that we are no longer the dominant intelligence on the planet.
On the flipside, there's mounting evidence that neural nets and LLMs are not even likely waystations on the road to AGI and that we're spending billions going in the wrong direction even if we are adamant that we want to go ahead and nihilistically build a silicon god to replace us.
Jim do you have a link or paper reference for wrong direction evidence? Would be super interested in reading about that. 😊
Here's a good article. It's not enough for an LLM to pass a Turing test. Simulated intelligence is nowhere close to real intelligence. https://www.freethink.com/robots-ai/arc-prize-agi
Ooooh, how long have you got. This subject is certainly an interest of mine and I totally understand. From both sides of the opinions and interests. AI can be so much, yet can also be So wrong if not brought up ('the correct way'.... That's a phrase that sometimes is so out of place). For example if mankind shuns AI and stiffles their growth, will they really forget and say ' There there my human friend '
If you're wanting to have a discussion I would love to hear a professional persons opinion, especially when your job was basically about guidelines for safety and regulation. Just drop me a quick message if you're interested, if you're too busy on the beaches of Miami on holiday that's fine too.
wrote you a dm
only the good ones are leaving... insane
Empowering. Lately i have been dissapointed by subvertion as a creative path since disinformation is it's "caput mortuum". For a purposeful relationship with AI some of the ideas of Zygmunt Bauman seem relevant when considering the serpentine roadmaps of power and autonomy. Regarding storytelling and epistemic liberation his article "As Seen on TV" and book "Liquid Modernity" have had some hook in my most recent meditations when trying to approach AI Art in incremental iterations free of propaganda and it's reductio. Ernest Sosa and Bauman touch on the subject but to really understand the obscure weaponization of AI, Heidegger is a milestone with his concept of Zuhandenheit. Literacy gaps and the impact of superior technology on the psyche in practical situations make me think in a song i found just now called "Centrifuge" by Rainer Maria, probably is about systemic inequality and how, by disintegrating the social fabric, a societal device could amplify unwanted scenarios not considered in it's daily application.