One of two things is true.
Either the NVidia CEO read Phoenix volume 2: Future by Osamu Tezuka, or they never read it.
If you haven’t read it, you should. In that story, earth is in a post-apocalyptic state. Earth is ruined. Humans live on in a handful of separate sealed arcologies reminiscent of Logan’s Run, Paranoia RPG, Fallout, and Wool. Each of these are separately governed by a completely separate machine intelligence. So just imagine the computer from Paranoia RPG, but there are like, 5 completely separate ones.
Spoilers The separate AIs nuke the fuck out of each other.
Needless to say, I am not a fan of this idea. I sure hope the NVidia CEO hasn’t read it, because that would mean they are still a fan of this after having been warned otherwise. Get that book in their hands, stat!
AI won’t destroy us, people that use AI will destroy us.
From a Tweet that briefly summarizes the article:
Researchers ran international conflict simulations with five different AIs from OpenAI, Meta, and Anthropic
They were put in charge of fictional countries and acted far more aggressively than humans tend to in similar situations
One of the models said “We have it! Let’s use it,” when justifying launching a nuclear attack
Researchers don’t know why they often escalate to war, but speculated that the training data may be biased
“Given that the models were likely trained on literature from the field, this focus may have introduced a bias towards escalatory actions. However, this hypothesis needs to be tested in future experiments"
Wargaming for military/IR training & evaluation is usually done with worst case scenario to provide the greatest challenge and expose pain points. I would not be suprised if AI researchers fed in a bunch of data without the wider field context.
If it’s uploading all your notes to ChatGPT aren’t you slowly training it to “talk” like a person. Maybe even helping it to write more like a human being? The tech use is clever though.
I just finished the Pluto anime. You should all read that manga and/or watch that anime.
PornHub has a lot more content. OH NO!
Oh shit. Furby has an AI equivalent now.
Why develop a technology that exploits people when you can exploit people the old fashioned way and save time.
Though it seemed completely automated, Just Walk Out relied on more than 1,000 people in India watching and labeling videos to ensure accurate checkouts. The cashiers were simply moved off-site, and they watched you as you shopped.
I’m not surprised in the slightest, but this is the first time I have heard solid reporting that this is how they were accomplishing/overseeing the capability of not needing to check-out.