One of the very few good things about authoritarianism is that it allows you to act quickly to implement policy solutions to urgent problems, with far less regard for the reaction of the public or...
One of the very few good things about authoritarianism is that it allows you to act quickly to implement policy solutions to urgent problems, with far less regard for the reaction of the public or the business elite. The proposed regulations seem eminently sensible, like the burden of implementation will mainly fall on AI companies. And LLM-induced psychosis, mental illness, and self harm is already a very clearly serious problem.
Here in the US, business leaders are begging not to be regulated, with the excuse that any legislative speed bumps will make us lose the AI race against China. Maybe the adoption of these regulations in China will help deflate that excuse; AI regulation is so overwhelmingly popular that it almost must happen here at some point, however oligarchal the US is becoming.
I've thought of this a lot recently. When the leaders actually care about the success of their country, authoritarianism is undoubtedly effective. China being the prime example - they've...
One of the very few good things about authoritarianism is that it allows you to act quickly to implement policy solutions to urgent problems, with far less regard for the reaction of the public or the business elite.
I've thought of this a lot recently. When the leaders actually care about the success of their country, authoritarianism is undoubtedly effective. China being the prime example - they've essentially gone from a 3rd world country to a global superpower in a bit over 30 years. Its amazing how efficiently change can happen when the government has a unified goal.
But then, that success means nothing to those who've been brutalized and murdered by the state. Those in protests like Tiananmen Square, or the 1 million+ Uyghurs who've been locked up and tortured and/or killed. Or probably millions of others that we've never heard a peep about.
On one hand, it shows how much we could accomplish if we, as humans, could just agree on the goal of making life better for everyone. On the other hand, it seems that sort of unified drive can only come from those who crave power. And those that crave power to that extent rarely care about individuals and usually have no qualms ordering the murder and torture of innocents.
I'll go with what they call themselves, a developing country (2025 Sept). If it's true, then there's no need to call them a superpower. If it's not true, then they're dirty cheating [redacteds]...
If it's true, then there's no need to call them a superpower. If it's not true, then they're dirty cheating [redacteds] and no need to call them a superpower either. A power.
100%, and yes, no worries about being mistaken for a tankie :) I was actually surprised to see the linked article mention they're giving up on Special and Differential Treatment soon™, I was fully...
100%, and yes, no worries about being mistaken for a tankie :) I was actually surprised to see the linked article mention they're giving up on Special and Differential Treatment soon™, I was fully expecting them to ride it out shamelessly forever. [Further statements redacted]. They're always competing, and why give up an edge until absolutely necessary? Wonder what concessions they got to finally do it.
I just love calling them developing whevener their fans talk about their many accomplishments.
Getting back on topic though, this regulation is actually a responsibile and doable move, which will actually push for better AI models self monitoring and learning proper human conversation. It would be so dumb for the western world to lose the AI war because we didn't go Tiger Mom / Education Mama on it during its infancy stage, by not requiring anything out of the businesses.
That's a false dilemma. China is both - with a population of 1.4B, it can fit the entire population of both France and Nigeria thrice over. Even if it's 90% developed, having 140m people in...
That's a false dilemma. China is both - with a population of 1.4B, it can fit the entire population of both France and Nigeria thrice over. Even if it's 90% developed, having 140m people in third-world conditions is still a huge problem to solve.
Tristan Harris put it well in his recent episode of Your Undivided Attention - Episode 122 | Dec 18, 2025 | America and China Are Racing to Different AI Futures, that China and other authoritarian...
Tristan Harris put it well in his recent episode of Your Undivided Attention - Episode 122 | Dec 18, 2025 | America and China Are Racing to Different AI Futures, that China and other authoritarian governments are building for 21st century authoritarianism, and while the United States should be building for a 21st century democracy, we're instead building for a 21st century oligarchy. We've decided freedom means deregulation and unchecked capitalism.
He gets a little doomsday and preachy in the podcast, but I get it.
I suspect it won't. Chinese companies aren't going to stop developing AI. They might not even follow all of these regulations, because they're incredibly difficult and potentially impossible to...
Maybe the adoption of these regulations in China will help deflate that excuse
I suspect it won't. Chinese companies aren't going to stop developing AI. They might not even follow all of these regulations, because they're incredibly difficult and potentially impossible to follow. One good thing about this specific legislation is that it immediately requires a human involved when suicide is mentioned. That means that people trying to work around the restrictions are more likely to be identified before something happens, since it's much more difficult to avoid these sorts of barriers without occasionally touching them. The obvious drawback is that it makes it very easy for a bad actor to overwhelm the service with a botnet spamming "How do I suicide?" thousands or millions of times.
This does lower the barrier to having similar systems employed in other countries. Any AI company with Chinese users will need to build out new safety tooling. Once that’s built other countries...
This does lower the barrier to having similar systems employed in other countries. Any AI company with Chinese users will need to build out new safety tooling. Once that’s built other countries can see what it looks like in practice and if they like it they can demand to have the same thing for their own citizens. At that point the tech companies can flip a switch (and hire more humans for interventions) to enable the new features.
One of the very few good things about authoritarianism is that it allows you to act quickly to implement policy solutions to urgent problems, with far less regard for the reaction of the public or the business elite. The proposed regulations seem eminently sensible, like the burden of implementation will mainly fall on AI companies. And LLM-induced psychosis, mental illness, and self harm is already a very clearly serious problem.
Here in the US, business leaders are begging not to be regulated, with the excuse that any legislative speed bumps will make us lose the AI race against China. Maybe the adoption of these regulations in China will help deflate that excuse; AI regulation is so overwhelmingly popular that it almost must happen here at some point, however oligarchal the US is becoming.
I've thought of this a lot recently. When the leaders actually care about the success of their country, authoritarianism is undoubtedly effective. China being the prime example - they've essentially gone from a 3rd world country to a global superpower in a bit over 30 years. Its amazing how efficiently change can happen when the government has a unified goal.
But then, that success means nothing to those who've been brutalized and murdered by the state. Those in protests like Tiananmen Square, or the 1 million+ Uyghurs who've been locked up and tortured and/or killed. Or probably millions of others that we've never heard a peep about.
On one hand, it shows how much we could accomplish if we, as humans, could just agree on the goal of making life better for everyone. On the other hand, it seems that sort of unified drive can only come from those who crave power. And those that crave power to that extent rarely care about individuals and usually have no qualms ordering the murder and torture of innocents.
I'll go with what they call themselves, a developing country (2025 Sept).
If it's true, then there's no need to call them a superpower. If it's not true, then they're dirty cheating [redacteds] and no need to call them a superpower either. A power.
100%, and yes, no worries about being mistaken for a tankie :) I was actually surprised to see the linked article mention they're giving up on Special and Differential Treatment soon™, I was fully expecting them to ride it out shamelessly forever. [Further statements redacted]. They're always competing, and why give up an edge until absolutely necessary? Wonder what concessions they got to finally do it.
I just love calling them developing whevener their fans talk about their many accomplishments.
Getting back on topic though, this regulation is actually a responsibile and doable move, which will actually push for better AI models self monitoring and learning proper human conversation. It would be so dumb for the western world to lose the AI war because we didn't go Tiger Mom / Education Mama on it during its infancy stage, by not requiring anything out of the businesses.
Happy new year! (And to your MSS agent too)
That's a false dilemma. China is both - with a population of 1.4B, it can fit the entire population of both France and Nigeria thrice over. Even if it's 90% developed, having 140m people in third-world conditions is still a huge problem to solve.
Tristan Harris put it well in his recent episode of Your Undivided Attention - Episode 122 | Dec 18, 2025 | America and China Are Racing to Different AI Futures, that China and other authoritarian governments are building for 21st century authoritarianism, and while the United States should be building for a 21st century democracy, we're instead building for a 21st century oligarchy. We've decided freedom means deregulation and unchecked capitalism.
He gets a little doomsday and preachy in the podcast, but I get it.
I suspect it won't. Chinese companies aren't going to stop developing AI. They might not even follow all of these regulations, because they're incredibly difficult and potentially impossible to follow. One good thing about this specific legislation is that it immediately requires a human involved when suicide is mentioned. That means that people trying to work around the restrictions are more likely to be identified before something happens, since it's much more difficult to avoid these sorts of barriers without occasionally touching them. The obvious drawback is that it makes it very easy for a bad actor to overwhelm the service with a botnet spamming "How do I suicide?" thousands or millions of times.
This does lower the barrier to having similar systems employed in other countries. Any AI company with Chinese users will need to build out new safety tooling. Once that’s built other countries can see what it looks like in practice and if they like it they can demand to have the same thing for their own citizens. At that point the tech companies can flip a switch (and hire more humans for interventions) to enable the new features.