The Jacob Coxon tweet that broke the internet… When asked about in a roundtable debate, the answers were divided. “ The world is finally starting to take the risk seriously” or “There still isn’t enough evidence to justify claims that extreme” Is this a risk we should genuinely be... worried about, or is it becoming a distraction from the problems AI is causing right now?show more

Super distraction. And not only that you did not actually show are any researchers at all. These folks are researchers in the same way that North Korea is a democractic republic. All these folks are blog and book writers and have not done a single lick of actual ML researcher. They take fake titles to look more official. If your point requires you to fake your credentials then you should be asking yourself why they need to fake their identity to make their point and what other frauds are they committing? A strong argument exists on its own merits and does not require astroturfing public opinion with fake credentials.

@hilbertspaess I usually love ur show but the fact that you probably didn’t even try to put a woman on that panel. There are amazing female researchers working on both sides of this

@hilbertspaess Honestly we're spending so much time talking about the negative and not enough time talking about the positive.

@hilbertspaess * I'm concerned about cyber attacks on Banks and infrastructure and being blamed on a 'conscious' AI. * I'm concerned about a Digital Dystopia like has been rolling out in China that I'm sure AI will be needed for. * I'm concerned that Convenience will mean Communism.

@hilbertspaess I think we should take both seriously. The risks we can see today deserve attention, but uncertainty about future AI risks shouldn't mean we ignore them. The real challenge is learning how to control increasingly capable systems before we understand the consequences too late. 🧠

@hilbertspaess The positives about AI don't matter if the negative is so absolute

I’ve watched the whole thing and it’s obvious that you’re monetizing this wave of fear mongers. While you pull it off well, this panel is “experts” are speaking in platitudes not because they want to communicate to lay people, but because they’re just saying bs based on their own personal feelings. Entertaining nonetheless.

@hilbertspaess

@hilbertspaess Feels like we’re debating future risks while some very real AI risks are already here.

@hilbertspaess It's not a distraction, it's a smokescreen. Focusing purely on apocalyptic risks lets tech leaders and politicians avoid answering the most pressing question: what happens when hyperautomation renders labour obsolete? A new economic policy framework is what we should be debating.

@hilbertspaess ❤❤

@hilbertspaess There’s a possibility we are sprinting full speed in the dark towards a cliff, or it’s nothing to worry about. The truth is we don’t know and most likely nobody does. We will only know when it happens. So I think we should get comfortable not knowing lol

Humanity First: What To Do. The message is simple: prove control before building systems that may become impossible to control. This means: → Pause the dangerous combination of generality, broad agency, long-running action & self-improvement → Keep useful, narrow AI moving → Require independent testing of shutdown, auditing, self-change & covert communication → Give inspectors real authority to halt runs → Disclose swarm incidents and track frontier compute We’re not saying “no” to AI. We’re saying: prove it’s safe first. #AI #AISafety #ArtificialIntelligence #AIAlignment #AIControl #ResponsibleAI #HumanityFirst

@hilbertspaess I think Nate did exceptionally well here

@hilbertspaess Has AI produced an original thought or idea ?

@hilbertspaess Yeah through a bigger psyop than Covid hahaha, the real question is why? Why my boomer mother who does not have a computer, and a housewife, texting me asking "Is AI going to kill us all?" Now that's distrubition.

@hilbertspaess The voices of reason usually don't talk much. The voices of scared and confused people, grifters and mentally ill people with a messaging need scream very loud. Do not forget that "mentally ill" can be high functioning - just for other goals.

@hilbertspaess It is 100% a distraction. We don’t have superintelligent gods plotting our extinction; we have corporate balance sheets that need existential drama to justify insane valuations and capture the market before anyone else can compete.

@hilbertspaess Not either/or. Dr. Yampolskiy (one of your guests here) argues we can't verify AI is safe before we build it — like grading a huge exam by reading one page. We turned his case into a plain-English one-pager, no jargon:

@hilbertspaess Over 2 hours of psychosis.

@hilbertspaess It's definitely a big question with no easy answers. Lots of factors to consider.

@hilbertspaess However you look at it, governments need to set restrictions.