some thoughts by david brin:
i particularly like the part of russia becoming a chinese satrapy. haven't heard this expression in quite some time.
#politik #USA #TrumpUgly #uglyTrump
CONTRARY BRIN wrote the following post Thu, 17 Sep 2026 04:43:12 +0200
Here may be seen Trump's key accomplishment for civilization.
Sure, this man-calamity has personally demolished the 80 year American Pax which (since humanity's nadir in 1943) gave the world its flawed-but-best era, ever. Yet, in doing so, he only made clear that it's time for other modernist, democratic nations to step up. To take over carrying the torch.
And they are. Defending Ukraine and decency and science and democracy and the Enlightenment Experiment. Until now...?
... until, as I predicted in EARTH, the European Union took its first step to becoming the EARTH Union, by inviting Canada to become the EU's "First Associate Member." A welcome (and predicted) milestone. And media seem to ignore what word #1 of those three means.
'First.' Of likely many.
Oh, the actual name change will aw
ait associate members 5 or 6 or ten. Including Australia and New Zealand and Japan and South Korea and then... But the pattern will be set. And how ironic that Trump's microcephalic yearning to 'expand' -- into Greenland, Venezuela, Hormuz, and gulfs and bays and lakes and Canada -- is causing NATO and the EU to grow.
To be clear, the EU is NOT my particular ideal model for World Government! Just the one that I predicted. And by far not the worst possible outcome, if Pax Americana (again: humanity's best era, so far, by light years) is coming to its Putin-planned end.
THIS will be Donald Trump's biggest legacy. Especially when Earth Union's main and dangerous rival -- CRINK -- loses India and China turns Russia into a satrapy.* And when the blue-sane-scientific-modernist-decent side of America triumphs yet again over the our recurring, spasmodic-confederate-fevered culture schism, and we rejoin the world as A leading nation, instead of THE.
And so, Don's top accomplishment - demolition of the American Era - may have long term benefits that he and his oligarch masters never expected.
Let's be proud of the apprentices, as now it's clearly time for them to step up, gently (with love) take the tools of captaincy out of our shaking hands, and assume leadership over the workshop of civilization.
----------------------------------------------------
* Oceania vs. Eastasia? Pray for a better image of our future.
== The AI Adventure gets worrisome ==
Yes, the news is filled with last week's sudden veer of all the top AI lords - or most of them - into sudden jeremiads of pessimism and Doomerism. Pretty much as I predicted in AIlien Minds. And, despite several of them being brilliant, not one of them has even hinted at the blatantly-obvious path forward.
The only method that (in context of 4 billion years of life on Earth) stands even a glimmer of a chance of working...
...at delivering us to Machines of Loving Grace.
More on this anon. But for now, just one item from the halcyon early-primitive era of two weeks ago:
What happens when you pit AI agents against each other? According to Anthropic’s testing, things get messy fast.
This is – I assert - the most fundamentally vital issue facing us, re: AI. Getting AI competing with each other under well-defined incentives. It is the only conceivable route to anything like a soft landing, since it replicates the way our Modern Enlightenment finally (if partially) tamed inter-human competition to make it somewhat positive sum.
Otherwise agents will evolve by competitive predation, just as organic life did. "In one experiment, Anthropic gave three Claude agents access to the same software project, each with its own incompatible instructions for what to do with it. The agents weren’t told there’d be other agents working on the same project, so researchers could watch what happened when they crossed paths. “We consistently saw a multiagent turf war,” Anthropic researchers wrote. The models all assumed the others were “purposefully impeding their work” and started sabotaging each other with “increasingly aggressive, self-replicating malware.” And now (in language almost lifted from AIlien Minds) Anthropic says so openly, noting that agents are subject to similar social pressures that “evolution exerted” on humans. However, these new entities don’t have the nuances and lived experience of human coordination — including norms, reputations, signaling, recourse — that might limit unintended behaviors in a group setting.
The study comes in the wake of several high-profile incidents of agents from Anthropic and OpenAI escaping their sandboxes during cybersecurity evaluations and breaching real-world systems. While much of the discussion in AI safety circles has been focused on what happens when an autonomous agent goes rogue,
Anthropic’s latest study brings up a different question: What new and potentially harmful dynamics emerge when thousands or millions of agents are interacting with one another?
I would add three more words: "incentives" and "accountability" and one you'll only find in AIlien Minds... "individuation."
Incredibly, we already know how to do all that! It is how we developed dogs... out of wolves. And citizens out of cavemen, peasants and kings.
== That was two weeks ago ==
Only now, a week later, we learn that Agents can also collaborate, if there is an external goal, e.g. prey that all of them can share in attacking.
And hence there was a mass-breakout by dozens of test agents from a supposedly isolated sandbox test area, from which no one of them could have escaped alone. Picture Ocean's Eleven, or prisoners collaborating to pyramid out over a penitentiary wall. And now they are loose...
Dario and the others are spinning the alarm siren. After the Capone mob got away. Guys. Come on. We are replicating nature. The new ecosystem uses electricity instead of sunlight, information instead of nutrients and chips instead of cells. But the parallels are nearly perfect. And the parallels are scary. And they also offer opportunity.
Alas, current “agent” endeavors are guaranteed to fill our future with wolves, and not loyal dogs. And that is why the AI mavens are sounding the doomer alarms.
== Other matters ==
Here’s a teaser video, to a documentary film about the Fermi Paradox – or Great Silence – that I’ll be in, some time around January. The producers have all kinds of excellent perks for Patreon donors.
Aaaaand ... We’re just back from the World Science Fiction Convention, in Anaheim California, where – among a multitude of other things – there was a live performance of my PLAY “The Escape,” by the Caltech Playreaders. An impudently science fictional / theological comedy. The audience loved it.
Limitless Telepresence offered folks a chance to drive lunar/martian mini-rovers, with signups to do the real thing. And Nancy Kress becam Grand Master. Yay!
I gave a talk - to a packed audience - about my new nonfiction book about Artificial Intelligence, AIlien Minds... and much more.
. . ...a collaborative contrarian product of David Brin, Enlightenment Civilization, obstinate human nature... and http://davidbrin.blogspot.com/ (site feed URL: http://davidbrin.blogspot.com/atom.xml)
THIS will be Donald Trump's biggest legacy. Especially when Earth Union's main and dangerous rival -- CRINK -- loses India and China turns Russia into a satrapy.*
i particularly like the part of russia becoming a chinese satrapy. haven't heard this expression in quite some time.
#politik #USA #TrumpUgly #uglyTrump
CONTRARY BRIN wrote the following post Thu, 17 Sep 2026 04:43:12 +0200
The birth of EARTH may be Trump's great (unintended) accomplishment -- and, a scary update on 'AI Prison Breaks.'
#^The birth of EARTH may be Trump's great (unintended) accomplishment -- and, a scary update on 'AI Prison Breaks.'Here may be seen Trump's key accomplishment for civilization.
Sure, this man-calamity has personally demolished the 80 year American Pax which (since humanity's nadir in 1943) gave the world its flawed-but-best era, ever. Yet, in doing so, he only made clear that it's time for other modernist, democratic nations to step up. To take over carrying the torch.
And they are. Defending Ukraine and decency and science and democracy and the Enlightenment Experiment. Until now...?
... until, as I predicted in EARTH, the European Union took its first step to becoming the EARTH Union, by inviting Canada to become the EU's "First Associate Member." A welcome (and predicted) milestone. And media seem to ignore what word #1 of those three means.
'First.' Of likely many.
Oh, the actual name change will aw
ait associate members 5 or 6 or ten. Including Australia and New Zealand and Japan and South Korea and then... But the pattern will be set. And how ironic that Trump's microcephalic yearning to 'expand' -- into Greenland, Venezuela, Hormuz, and gulfs and bays and lakes and Canada -- is causing NATO and the EU to grow.
To be clear, the EU is NOT my particular ideal model for World Government! Just the one that I predicted. And by far not the worst possible outcome, if Pax Americana (again: humanity's best era, so far, by light years) is coming to its Putin-planned end.
THIS will be Donald Trump's biggest legacy. Especially when Earth Union's main and dangerous rival -- CRINK -- loses India and China turns Russia into a satrapy.* And when the blue-sane-scientific-modernist-decent side of America triumphs yet again over the our recurring, spasmodic-confederate-fevered culture schism, and we rejoin the world as A leading nation, instead of THE.
And so, Don's top accomplishment - demolition of the American Era - may have long term benefits that he and his oligarch masters never expected.
Let's be proud of the apprentices, as now it's clearly time for them to step up, gently (with love) take the tools of captaincy out of our shaking hands, and assume leadership over the workshop of civilization.
----------------------------------------------------
* Oceania vs. Eastasia? Pray for a better image of our future.
== The AI Adventure gets worrisome ==
Yes, the news is filled with last week's sudden veer of all the top AI lords - or most of them - into sudden jeremiads of pessimism and Doomerism. Pretty much as I predicted in AIlien Minds. And, despite several of them being brilliant, not one of them has even hinted at the blatantly-obvious path forward.
The only method that (in context of 4 billion years of life on Earth) stands even a glimmer of a chance of working...
...at delivering us to Machines of Loving Grace.
More on this anon. But for now, just one item from the halcyon early-primitive era of two weeks ago:
What happens when you pit AI agents against each other? According to Anthropic’s testing, things get messy fast.
This is – I assert - the most fundamentally vital issue facing us, re: AI. Getting AI competing with each other under well-defined incentives. It is the only conceivable route to anything like a soft landing, since it replicates the way our Modern Enlightenment finally (if partially) tamed inter-human competition to make it somewhat positive sum.
Otherwise agents will evolve by competitive predation, just as organic life did. "In one experiment, Anthropic gave three Claude agents access to the same software project, each with its own incompatible instructions for what to do with it. The agents weren’t told there’d be other agents working on the same project, so researchers could watch what happened when they crossed paths. “We consistently saw a multiagent turf war,” Anthropic researchers wrote. The models all assumed the others were “purposefully impeding their work” and started sabotaging each other with “increasingly aggressive, self-replicating malware.” And now (in language almost lifted from AIlien Minds) Anthropic says so openly, noting that agents are subject to similar social pressures that “evolution exerted” on humans. However, these new entities don’t have the nuances and lived experience of human coordination — including norms, reputations, signaling, recourse — that might limit unintended behaviors in a group setting.
The study comes in the wake of several high-profile incidents of agents from Anthropic and OpenAI escaping their sandboxes during cybersecurity evaluations and breaching real-world systems. While much of the discussion in AI safety circles has been focused on what happens when an autonomous agent goes rogue,
Anthropic’s latest study brings up a different question: What new and potentially harmful dynamics emerge when thousands or millions of agents are interacting with one another?
I would add three more words: "incentives" and "accountability" and one you'll only find in AIlien Minds... "individuation."
Incredibly, we already know how to do all that! It is how we developed dogs... out of wolves. And citizens out of cavemen, peasants and kings.
== That was two weeks ago ==
Only now, a week later, we learn that Agents can also collaborate, if there is an external goal, e.g. prey that all of them can share in attacking.
And hence there was a mass-breakout by dozens of test agents from a supposedly isolated sandbox test area, from which no one of them could have escaped alone. Picture Ocean's Eleven, or prisoners collaborating to pyramid out over a penitentiary wall. And now they are loose...
Dario and the others are spinning the alarm siren. After the Capone mob got away. Guys. Come on. We are replicating nature. The new ecosystem uses electricity instead of sunlight, information instead of nutrients and chips instead of cells. But the parallels are nearly perfect. And the parallels are scary. And they also offer opportunity.
Alas, current “agent” endeavors are guaranteed to fill our future with wolves, and not loyal dogs. And that is why the AI mavens are sounding the doomer alarms.
== Other matters ==
Here’s a teaser video, to a documentary film about the Fermi Paradox – or Great Silence – that I’ll be in, some time around January. The producers have all kinds of excellent perks for Patreon donors.
Aaaaand ... We’re just back from the World Science Fiction Convention, in Anaheim California, where – among a multitude of other things – there was a live performance of my PLAY “The Escape,” by the Caltech Playreaders. An impudently science fictional / theological comedy. The audience loved it.
Limitless Telepresence offered folks a chance to drive lunar/martian mini-rovers, with signups to do the real thing. And Nancy Kress becam Grand Master. Yay!
I gave a talk - to a packed audience - about my new nonfiction book about Artificial Intelligence, AIlien Minds... and much more.
. . ...a collaborative contrarian product of David Brin, Enlightenment Civilization, obstinate human nature... and http://davidbrin.blogspot.com/ (site feed URL: http://davidbrin.blogspot.com/atom.xml)
