Like a lot of other people these days, I sometimes get caught up in doom scrolling. This week I decided to summarize a few of the doomiest stories and make some probably-not-serious-enough comments on them, partly to entertain and partly to relieve some of my own anxiety.
As other bloggers know, sometimes writing about things at least gives us the illusion that we’re doing something about them, allowing us a greater sense of control. And, if we’re applying satire, maybe making fun of things allows us to take them a little less seriously. Gallows humor, and all that.
Anyway, here are a few items items from this week’s doom scrolling.
Death by Nuclear Knuckleheads
- Russia has updated its foreign policy doctrine to reflect its increasingly confrontational relationship with the West and its perception of “existential threats” to its security and development from “unfriendly states.” The document names the United States as the main threat to international stability and driver of an “anti-Russian line,” but also seeks “peaceful coexistence” and a “balance of interests” with Washington.
- The Russian Federation’s recent announcement of plans to station non-strategic nuclear weapons in Belarus represents the first “nuclear sharing” agreement made since the Treaty on the Non-Proliferation of Nuclear Weapons entered into force in 1970, the senior United Nations disarmament official told the Security Council today, emphasizing that — against the backdrop of the Ukraine conflict — the risk such arms will be used is higher today than at any time since the end of the cold war.
My Gratuitous Commentary
Mr. Putin is deep in a hole of his own digging and refuses to hand up his shovel. Russia is not under existential threat from anyone but Mr. Putin, but he alone may be enough to destroy it, and possibly the world, if we keep stupidly doing this dance toward nuclear apocalypse. We should have internationally banned, or at least vastly reduced the number of, those hellish devices decades ago. Let’s somehow make peace with this raging, narcissistic nutjob and move on with global nuclear nonproliferation. Soothe Putin’s aching, rotten ego. Pay him off. Whatever. Humanity has been walking this existential tightrope way too long and needs to get off of it. Man up, Biden, and get it done.
Extinction by Our Latest Favorite Smart Tech
- A smart dude from an AI thinktank says we’re all definitely going to die, wiped out by the most popular technology since..well…ever. Eliezer Yudkowsky, who leads research at the Machine Intelligence Research Institute, writes, “Many researchers steeped in these issues, including myself, expect that the most likely result of building a superhumanly smart AI, under anything remotely like the current circumstances, is that literally everyone on Earth will die. Not as in ‘maybe possibly some remote chance,’ but as in ‘that is the obvious thing that would happen.’ It’s not that you can’t, in principle, survive creating something much smarter than you; it’s that it would require precision and preparation and new scientific insights, and probably not having AI systems composed of giant inscrutable arrays of fractional numbers….Shut down all the large GPU clusters (the large computer farms where the most powerful AIs are refined). Shut down all the large training runs. Put a ceiling on how much computing power anyone is allowed to use in training an AI system, and move it downward over the coming years to compensate for more efficient training algorithms.”
More Gratuitous Commentary
First of all, respect for the line “giant inscrutable arrays of fractional numbers.” Well done. Second of all, Yudkowsky might be right. He’s certainly certain. If he’s right to be so certain, then our days are almost certainly numbered because the AI cat has already clawed its way out of the bag. A bunch of smart kids from Stanford just created a ChatGPT knockoff for a total of $600, which is about 1/6th of a month’s rent out in that part of the world.
Meanwhile, Meta’s AI–which is called LLaMa for Large Language Model Meta AI–was leaked and then shared on 4chan, where someone uploaded it via torrent. Just great. There was already one AI trained on 4chan so it became a toxic, hate-speech machine. That’s what we need: a super intelligence trained on the darkest dregs of humanity. At the very least, we’ll see some very nasty generative AI media coming down the pike .
So maybe we’re all doomed already. Regardless, the Biden Administration needs to weigh in soon because, well, Congress is busy with TikTok (no, seriously). Those old white guys haven’t a clue (and I’m quickly getting to be one, so I know). I figure our only hope for government action is some younger, brighter folks squirreled away somewhere in the West Wing who can convince old Joe that this AI thing is worth, you know, thinking about. But maybe they’re already deep into it. Maybe they’ve even got Yudkowsky on speed dial. Now if they can only reach him in whatever newly reburished-with-Faraday-caging 1950s-vintage bomb shelter he’s holed up in.
One thought on “The Doom Scrawl: April 1, 2023”