AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic.
-
@geolaw exactly - this alone is enough reason NOT to use LLMs.
@ntnsndr @emilymbender -
AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic.
Maybe I've missed it, but I haven't seen any statements from CEOs and other execs who are more on the AI Booster side of the ledger.
🧵>>
@emilymbender someone on HN commented that this Jacob Coxon guy worked there for less than 6 weeks which if true is ... special. No reason to trust that guy, he barely knows how to even request his PTO -
I want everyone, especially journalists, to notice the silence of the tech execs about this. This includes people like Bill Gates and Mark Zuckerberg whose recent blog posts talk about balancing the benefits and harms of "AI", and don't include Doomer scenarios among the harms.
They are complicit in the misinforming of the public. They are complicit in stoking everyone's fear. They reap the benefits of the Doomer narrative because it is their narrative.
/fin
@emilymbender Honestly I am kinda hoping the efforts to confuse legislators land at "Okay, we have to regulate it as if it's going to kill us all", because if the narrative (bullshit as it may be) is 'this has a 10% chance to kill everyone in the next 10 years" - then it's more dangerous than a nuke, and they just, like... give it to random people? to literal children?
Legislators who actually *believe* this narrative & regulate it as if it were really that dangerous would be a fitting outcome. -
@emilymbender Honestly I am kinda hoping the efforts to confuse legislators land at "Okay, we have to regulate it as if it's going to kill us all", because if the narrative (bullshit as it may be) is 'this has a 10% chance to kill everyone in the next 10 years" - then it's more dangerous than a nuke, and they just, like... give it to random people? to literal children?
Legislators who actually *believe* this narrative & regulate it as if it were really that dangerous would be a fitting outcome.@miss_rodent @emilymbender it's a nice hope, but that sort of approach leads to the kind of legislation we're seeing right now, where AI companies are only liable of they commit mass murder, and not at all if they tell a kid to commit suicide.
edit: oh and i forgot even the liability for mass murder is shit
chatgpt's license to kill you | Caelen Conrad
https://www.youtube.com/watch?v=34bMvXOQfEM(tried finding a text version of the same criticism but search engines are impossible nowadays)
-
@miss_rodent @emilymbender it's a nice hope, but that sort of approach leads to the kind of legislation we're seeing right now, where AI companies are only liable of they commit mass murder, and not at all if they tell a kid to commit suicide.
edit: oh and i forgot even the liability for mass murder is shit
chatgpt's license to kill you | Caelen Conrad
https://www.youtube.com/watch?v=34bMvXOQfEM(tried finding a text version of the same criticism but search engines are impossible nowadays)
@Yuvalne @emilymbender It's not being regulated as a threat at all currently, and I doubt it will be, because I don't think anyone in power actually believes the skynet doomsday bullshit, but if legislators acted seriously as if it was even a tenth as dangerous as the AI companies themselves say it is, the only reasonable response is to regulate it so heavily that no civilian can even get close to using one without a security clearance.
-
They can't, because the Doomer narrative *is* the Booster narrative. To debunk the Doomer stuff (and it *is* bunk), you have to point out that it's based on pure fantasy that the mathy maths are going to combust into consciousness ...
... or if not consciousness, then some thinking, reasoning, self-improving entity.
>>@emilymbender I agree consciousness is unlikely, but I don't think that's technically necessary; all it takes is something that looks like it enough to make humans treat it as such. Then we could get laws treating them like a legal person with rights! A bit like the (feels old saying now!) AI can't do your job but it can convince your boss enough to sack you.
It might not have to be real, just convincing to a largely uneducated population.
(Read your book, BTW, brilliant, thanks!)
-
@Yuvalne @emilymbender It's not being regulated as a threat at all currently, and I doubt it will be, because I don't think anyone in power actually believes the skynet doomsday bullshit, but if legislators acted seriously as if it was even a tenth as dangerous as the AI companies themselves say it is, the only reasonable response is to regulate it so heavily that no civilian can even get close to using one without a security clearance.
@miss_rodent @emilymbender i mean, it happened for a sec with Anthropic, but the state quickly backed down from attacking corporate profits.
i think we'll much more likely see some sort of US federal legislation (or let's be honest, and executive order) saying the extent of these companies' liability is they shouldn't let people develop nukes or malware "without supervision". and the people donutting suicide can go to hell because they don't matter. -
@Yuvalne @emilymbender It's not being regulated as a threat at all currently, and I doubt it will be, because I don't think anyone in power actually believes the skynet doomsday bullshit, but if legislators acted seriously as if it was even a tenth as dangerous as the AI companies themselves say it is, the only reasonable response is to regulate it so heavily that no civilian can even get close to using one without a security clearance.
@Yuvalne @emilymbender This is.. not really a *reasonable* way to regulate them - but it would mean all the money spent on infiltrating civilian companies, contexts, schools, devices, every aspect of day-to-day life is rendered useless.
Not an ideal case (and tbh worse than just an outright ban, or various other options to regulate it), but, would at least funny to have their own narrative undermine all their efforts. -
AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic.
Maybe I've missed it, but I haven't seen any statements from CEOs and other execs who are more on the AI Booster side of the ledger.
🧵>>
Part of me thinks that guy is still on the payroll, anything to hype the awesome power of a failing stats model.
-
@miss_rodent @emilymbender i mean, it happened for a sec with Anthropic, but the state quickly backed down from attacking corporate profits.
i think we'll much more likely see some sort of US federal legislation (or let's be honest, and executive order) saying the extent of these companies' liability is they shouldn't let people develop nukes or malware "without supervision". and the people donutting suicide can go to hell because they don't matter.@Yuvalne @emilymbender The US currently is run by an openly fascist death-cult, with a large number of it's base actively seeking to bring about the end times so Jesus comes back - there is no path to meaningful regulation of *anything* in the US that doesn't involve putting down the fascist movement.
-
@emilymbender All the while, human culture erodes and oceans boil.
@jordan @emilymbender I first saw the comment about boosters vs doomers and a moderate stance about the harms and was really confused because the harms will plausibly do the job the doomers are afraid of, so I wondered what was the difference between moderate and doomer.
Reading the other comments made it make more sense. The theft, the energy use, poisoning water supplies, destroying homes with eminent domain, income inequality... These are the problems.
-
@Yuvalne @emilymbender The US currently is run by an openly fascist death-cult, with a large number of it's base actively seeking to bring about the end times so Jesus comes back - there is no path to meaningful regulation of *anything* in the US that doesn't involve putting down the fascist movement.
@miss_rodent @emilymbender
and do we think the rest of the US colonies that are the west will do anything about it? especially when the US does not? -
AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic.
Maybe I've missed it, but I haven't seen any statements from CEOs and other execs who are more on the AI Booster side of the ledger.
🧵>>
@emilymbender yeah no this basically nails it. they're perfectly happy to let everyone believe that AI is some inevitable doom product, just like how they want to keep the conversation around easily obscurable water-usage statistics and not irreplaceable resources and energy grid failure.
-
@soulsource @emilymbender @dotsie If it is as dangerous as the doomer narrative is trying to frame it as, then regulating it appropriately means no one without a security clearance can use it.
You can not rent a nuke, which is far less existentially-risky than the doomer narrative frames the slopbots as.
Regulatory capture is the path we are on *now* in the US. -
@jordan @emilymbender I first saw the comment about boosters vs doomers and a moderate stance about the harms and was really confused because the harms will plausibly do the job the doomers are afraid of, so I wondered what was the difference between moderate and doomer.
Reading the other comments made it make more sense. The theft, the energy use, poisoning water supplies, destroying homes with eminent domain, income inequality... These are the problems.
@fullywoolly @jordan @emilymbender
Which hits on another purpose behind the doomerism narrative. It muddies the water.
When folk who aren't read into this hear "LLMs are an existential threat", the person saying it generally gets dumped in with "those crackpots who think it's going to become superhuman and enslave us all".
When what is actually meant by that is, "these same billionaires pulling the strings are using it as an excuse for concentrating power and wealth by destroying jobs, careers, livelihoods, communities, governments, economies, oceans, the atmosphere, and society itself."
Which in turn forces opponents to become overly specific in their discussion of it in order to be understood, which wastes time, effort, energy, and increases the likelihood of being ignored or burning out.
Also see: their deliberate conflation of both ML and LLMs as "AI".
-
@ligaturerecords not exactly
@emilymbender @ligaturerecords wasn't Elon one of the original doomers re AI so he founded the nonprofit OpenAI to have irresponsible profit hunting out of the picture?
How come he's now pro AI? It's profits, isn't it?
-
I want everyone, especially journalists, to notice the silence of the tech execs about this. This includes people like Bill Gates and Mark Zuckerberg whose recent blog posts talk about balancing the benefits and harms of "AI", and don't include Doomer scenarios among the harms.
They are complicit in the misinforming of the public. They are complicit in stoking everyone's fear. They reap the benefits of the Doomer narrative because it is their narrative.
/fin
@emilymbender "oh nooo the chatbot we're making is soooo smart it's gonna take over the world if we're not caaaareful, you'd better give us more money so we can figure this out!"
imagine having control over that much wealth and power and choosing to spread fear over a fake scenario that isn't happening just so you can hoard a bit more wealth and power, instead of, I dunno, focusing on actual problems like the ongoing collapse of our entire biosphere
the sad part is some of them are so far gone and full of their own opinion that they might honestly actually believe in roko's basilisk
-
AI Doomerism has been in the news lately, especially with the loud, viral resignation of Jacob Coxon from Anthropic.
Maybe I've missed it, but I haven't seen any statements from CEOs and other execs who are more on the AI Booster side of the ledger.
🧵>>
@emilymbender
If humanity were exterminated by something it created itself, It will be a well-deserved ending. -
@miss_rodent @emilymbender one advantage we have when it comes to the climate crisis is that the green transition makes pure economic sense not only in the long term (as has always been the case), but also in the short term (as has only been the case these past few years).
for AI it definitely makes sense in the long term (betting all of the world's economy on this vaporware doesn't make much sense), but in the short term? stocks are riding high. -
@emilymbender
If humanity were exterminated by something it created itself, It will be a well-deserved ending.I’ve always said aliens don’t need to invade this planet to conquer it — humans will kill ourselves and they’ll be able to get it for free, along with a hell of a show to watch from orbit.
