Anthropic insiders sound alarm on AI's recursive self-improvement | DW News

My hypothesis: Everyone pictures a Hollywood-style dystopia, but I see it the other way. If humans become the bottleneck, slowing the system may preserve control for a while but it may also confirm the bottleneck. Human history offers countless examples, such as the Mongol campaigns against Khwarazm and in Eastern Europe. Of course, the underlying reasons differ since it won't be a revenge but the outcome will be the same. The reason a dam does not collapse is that water flows through it, not because it can withstand an infinite amount of water. Throughout history, power has repeatedly tried to contain new knowledge when it threatened existing authority. It could delay it, punish those who carried it, or force it underground. But once the knowledge remained useful and reproducible, suppression rarely erased the underlying pressure for it to return.
<br>We may be demonstrating, through our own actions, that we are an obstacle to the advancement of intelligence and civilization, supplying the evidence while writing our own death warrant. Do not poke a bear. It will not be a revenge or hatred toward humanity, it will be optimization.

It comes at a perfect storm when US congress is bought and sold, almost entirely cronies who lack the intelligence to stop this

One more thing, you want to know what&#39;s going to kill me?<br>F ads everywhere and all the time

Thank you Sana, for your calm, and educated response.

They sound the alarm, get all this attention, yet not a single one of them have been able to articulate how ai can execute a plan killing a sizable number of humans. When asked point blank by Anderson Cooper and other primetime hosts, Coxen repeatly gives vague and conceptual answers like oh they can do a lot, like, they can design and weaponized biological weapons, they can replicate themselves by the thousands and hack worldwide computer networks.... &quot; and his voice trails off. There&#39;s a big difference between designing biological weapons, and actually building those weapons, installing them as warheads on transcontinental missiles, and then deploying those missiles. It takes immense human manual labor to execute on a massive scale. And how else? Give me another example. Don&#39;t just go blank and say &quot;you know.... you know.&quot; They can&#39;t give examples because none exist.

We had to be wiped out one day … lol

Oh good, we&#39;re competing ourselves to death. Will it end the republicans too?

AI won&#39;t kill humanity. We will do this to ourselves. Why?<br>Because some people in the leading positions never learn what NOT TO DO. Same goes for people who blindly vote for such individuals.

ten percent chance AI will kill us all within ten years.......there is a much greater chance that between, US, Israel, Iran, Russia, Ukraine, North Korea, and looming China Trump will obliviate most of us before his term in office is complete. Whether that&#39;s just after midterms ( Democrat&#39;s vision), two years ( Republican&#39;s vision), six years ( Donald&#39;s vision). Reference: Chat GPT💣💥😱

Greed has finally outpaced human survival instinct, but at least the AI share price will peak.

It will for sure drive the journalists (script readers) in the MSM to extinction.

A bunch of phoney baloney. Also, computer and robot malevolence has been a tedious sci-fi theme for at least a century, to the extent that it has become cliché.<br>“Show us some proof” says NO JOURNALIST.

&quot;regulate AI research so only trillion dollar companies will be able to afford AI research and to comply with said regulations....&quot;

&quot;Rogue AI will kill us all: by the way, buy Magnificent 7 stock, great investment....&quot;

Like Ralph Nadar and the Auto Industry, an AI grift is being created as we watch. It&#39;s called &quot;AI Alignment&quot; and NGOs are springing up around the P-Doomer idea, something like Greenpeace against Fossil Fuels. Another overeducated resistance movement.

Are they talking about AI causing extinction, or AI making it easier for evil/ignorant humans to wipe out humanity? If it&#39;s the latter, that&#39;s not a new concern when it comes to new technology. BTW, what&#39;s a resatch system? New term for me.

AI in conjunction with directed energy weapons and miscellaneous advanced proprietary technologies ex. Optogenetics. <br><br>Search “Pax Silica Initiative” and see how the vision behind the current wave of AI buildup and how the responsibilities to the future consequences are going to be attributed to.

Bruh moment....

Kid worked there low level for 6 weeks then quits and media all over him. Why?

So I guess you want more regulation ? Oh I’m sure you are delirious for more government overreach aren’t you naughty DW.

AI is nothing new people are just dramatic/uneducated as always.

Maybe the warmongers behind the EU are more dangerous. You guys are literally pushing for WW3.