Topic

AI may erode human control

3m

Rapid advances in artificial intelligence are fueling robotics research and intensifying debate over cybersecurity, corporate power and whether increasingly autonomous systems could erode human control.

51%
Can't imagine how mathematicians feel right now. AI companies aren't passionate about pure math; they're using math problems to show how powerful their models are. It's as if an arms dealer kept bombing houses in your neighborhood just so people would buy their bombs to use on more valuable targets.
November 20, 2026 at 10:39 PM
148%
113%
Agreed :)
November 21, 2026 at 2:31 AM
117%
148%
40%
my dear bluesky friends: AI is not a fad. i feel like the predictable (and satisfying!) crypto/NFT crash gave folks a false sense that they can just wait this out and things will go back to normal. there is almost certainly some degree of AI bubble right now but when it pops AI will still be here.
November 20, 2026 at 4:44 PM
44%
7%
16%
This is where I depart most strongly from deflationary accounts of AI. This is indeed extremely interesting. It's the point that the Sparks of AGI paper was trying to make, but failed to do so persuasively.
November 21, 2026 at 5:29 AM
37%
I actually think there is a lot of concrete value in researchers talking to researchers about what better AI governance looks like, and encouraging them to get involved in ensuring that AI companies don't get to self-regulate, in a personal/political capacity drawing on their technical expertise
November 21, 2026 at 2:17 AM
5%
AI corporations have nearly completely unregulated finances.

Much of their debt is off the books or financed with private credit.

Now they're pulling insurance companies into their orbit to insure against default.

Who will be stuck with the tab when the bubble bursts?
November 20, 2026 at 9:00 PM
3%
8%
Now that is a gorgeous analogy (about a hideous phenomenon).
Can't imagine how mathematicians feel right now. AI companies aren't passionate about pure math; they're using math problems to show how powerful their models are. It's as if an arms dealer kept bombing houses in your neighborhood just so people would buy their bombs to use on more valuable targets.
November 20, 2026 at 10:50 PM
48%
14%

Reposted by Virginia Dignum

24%
Oooh, finally translated into English too!

> Olivia Guest calls warnings about rogue AI agents acting “behind the backs” of humans “nonsense to drum up impressive seeming results and avoid accountability”. Dignum [ @vdignum.bsky.social ] also sees such claims as total garbage.
Doomerism does not dominate the AI debate by coincidence
By focusing on the hypothetical risks of AI, we lose sight of the actual harm it causes.
apache.be
November 20, 2026 at 10:32 AM
37%
60%
8%
37%
I still don't like attributing assisting or facilitation to non-moral-agents. Maybe we just need to ask for scales of autonomy for cybersecurity too, in fact all AI applications, not just "driverless" cars.
November 20, 2026 at 1:23 PM