AI runs on power grids that cannot meet its demands. “Useful” AI won’t get priority. Security and regime stability will. So even if AI lifts productivity & competence at the margin, a big share of the gains will get pulled into military, surveillance, and control—because those are the non-negotiables.
Very interesting analysis! The latest Chinese GLM-5 model is decentralized, uses about 100 small specialised AIs to produce a complete model, and has a much lower hallucination rate than the best American models. Trying to solve metaphysical anxieties of the American transhumanists with an excessively huge energy-hungry AGI model is not the best economical long term approach.
if learning is data in the west 'cluster' and inference is compute in the east 'hub', it's not clear how the map illustrates this as it shows the same number of clusters and hubs in both east and west. the bullet descriptions in the map repeat the same characteristics for both clusters and hubs using slightly different wording and what is the difference of terms cluster and hub anyway, and the title of the map is backwards - it has data in east instead of west and compute in west instead of east.
Decouple is one of very few publishers that analyzes problems and significance of state capacity, that must and will change as the chinese model and the rigors of multi polar relations are signaling every country to figure out how to increase its state capacity. scholarly domains and paradigms must and will shift to state capacity and other central challenges of transition to multi polar and automation. it is thorough going learning through practice and adjustment at every level and scope of hierarchy. Akamatsu, Lin YiFu, Carlos Lopes and similar political economists distinguish broad roles of capital and state capacity in the diffusion of demand and technique, the elevation of productivity through innovation in high education countries diffusing to others. State capacity equips capital with externalities and capital iteratively allocates to latent comparative advantage, advance is not possible without both functions. State capacity cannot determine comparative advantage and capital cannot determine infrastructure and commercial policy external to direct market exchange feedback. China has many advantages that helped it establish a better division of roles of state capacity and capital, this discovery prompts other countries few of which will have the combination of favorable circumstances to structure similarly, to engineer productivity ascent through public private roles. the advanced countries burden capital with expensive and frictive externalities due to lack of state capacity, a legacy of 1950s keynesian large state.
the professional state vs representative state is illustrated in this example of AI. cultivated experts and accumulated organization knowledge of agencies vs contesting ideas in electoral representative bodies. it seems likely two levels must distinguish everywhere, the higher strategic bodies that must operate hegemony, mediate the groups, define and agree on top requirements, and the lower agencies that professionally learn and operate their domain and crucially feed valid accessible information upward and report performance to and possible adjustment of the requirements from upper bodies.
the pertinent situation is that the western liberal formation exercised domination of political economy operation that prevented any change, absolute total blockage. Nativism is the only force so far that is dismantling the liberal complexes and reorganizing agency and corporate entities and their interaction. the advanced countries remain far far from the restructuring needed to become self guiding in economic and geopolitical advance. the power of nativism remains crucial to move forward and thoroughly remove liberal complexes. Nativism's ideology and source of power in grievance, and its new structures of corporate and political kingpin joint rewarded endeavor, arent the sharpest tools in the shed. But international competition will continue to pound hard toward pragmatic changes that work. that's where the analysis of Decouple and similar become decisive to our future.
We by nature and as reflected in our foundation documents don’t trust the government.
We never did we never will. Government and Legitimacy flows from restraint and being faithful to the Foundation.
(Note that Legitimacy isn’t the same as winning elections or power, democracy is eternally the buying of votes).
Also America is a Federation. Every American political arrangement from the Iroquois to the Internet is a Federation. This seems to be our nature. Certainly it keeps happening.
We don’t trust government and at present we don’t trust each other.
There won’t and can’t be any AI government regulation scheme.
AI runs on power grids that cannot meet its demands. “Useful” AI won’t get priority. Security and regime stability will. So even if AI lifts productivity & competence at the margin, a big share of the gains will get pulled into military, surveillance, and control—because those are the non-negotiables.
Very interesting analysis! The latest Chinese GLM-5 model is decentralized, uses about 100 small specialised AIs to produce a complete model, and has a much lower hallucination rate than the best American models. Trying to solve metaphysical anxieties of the American transhumanists with an excessively huge energy-hungry AGI model is not the best economical long term approach.
if learning is data in the west 'cluster' and inference is compute in the east 'hub', it's not clear how the map illustrates this as it shows the same number of clusters and hubs in both east and west. the bullet descriptions in the map repeat the same characteristics for both clusters and hubs using slightly different wording and what is the difference of terms cluster and hub anyway, and the title of the map is backwards - it has data in east instead of west and compute in west instead of east.
Decouple is one of very few publishers that analyzes problems and significance of state capacity, that must and will change as the chinese model and the rigors of multi polar relations are signaling every country to figure out how to increase its state capacity. scholarly domains and paradigms must and will shift to state capacity and other central challenges of transition to multi polar and automation. it is thorough going learning through practice and adjustment at every level and scope of hierarchy. Akamatsu, Lin YiFu, Carlos Lopes and similar political economists distinguish broad roles of capital and state capacity in the diffusion of demand and technique, the elevation of productivity through innovation in high education countries diffusing to others. State capacity equips capital with externalities and capital iteratively allocates to latent comparative advantage, advance is not possible without both functions. State capacity cannot determine comparative advantage and capital cannot determine infrastructure and commercial policy external to direct market exchange feedback. China has many advantages that helped it establish a better division of roles of state capacity and capital, this discovery prompts other countries few of which will have the combination of favorable circumstances to structure similarly, to engineer productivity ascent through public private roles. the advanced countries burden capital with expensive and frictive externalities due to lack of state capacity, a legacy of 1950s keynesian large state.
the professional state vs representative state is illustrated in this example of AI. cultivated experts and accumulated organization knowledge of agencies vs contesting ideas in electoral representative bodies. it seems likely two levels must distinguish everywhere, the higher strategic bodies that must operate hegemony, mediate the groups, define and agree on top requirements, and the lower agencies that professionally learn and operate their domain and crucially feed valid accessible information upward and report performance to and possible adjustment of the requirements from upper bodies.
the pertinent situation is that the western liberal formation exercised domination of political economy operation that prevented any change, absolute total blockage. Nativism is the only force so far that is dismantling the liberal complexes and reorganizing agency and corporate entities and their interaction. the advanced countries remain far far from the restructuring needed to become self guiding in economic and geopolitical advance. the power of nativism remains crucial to move forward and thoroughly remove liberal complexes. Nativism's ideology and source of power in grievance, and its new structures of corporate and political kingpin joint rewarded endeavor, arent the sharpest tools in the shed. But international competition will continue to pound hard toward pragmatic changes that work. that's where the analysis of Decouple and similar become decisive to our future.
I’m American.
We by nature and as reflected in our foundation documents don’t trust the government.
We never did we never will. Government and Legitimacy flows from restraint and being faithful to the Foundation.
(Note that Legitimacy isn’t the same as winning elections or power, democracy is eternally the buying of votes).
Also America is a Federation. Every American political arrangement from the Iroquois to the Internet is a Federation. This seems to be our nature. Certainly it keeps happening.
We don’t trust government and at present we don’t trust each other.
There won’t and can’t be any AI government regulation scheme.