Japan ramps up tech ambitions with ¥10 tril for AI, chips

This is a sure bet, of course in the sense of being lost as the only possible outcome. Relying on a further AI development or breakthrough in practice is a misleading path. That is not going to happen, not with more money, energy and resources and not by useless scaling up of processors or more data centers. In the best case, it is a little bit useful as an untrusted tool, not more not less. Efficiency gains will of course all eaten up by the checking and validating of every single AI output, as it intrinsically cannot work properly. And therefore the necessary massive checking and validating of all AI results isn't even anymore possible under a society in shrinking and aging trend.

Taiwan is home to some of the richest and most diverse forests in the world over 4,000 years old. Illegal tree poaching has taken a toll on the precious forests, and Taiwan still suffers from disproportionate forest loss. The yellow cypress, or Hinoki, was the first type of tree poached while Japan occupied Taiwan between 1912 and 1945. 

They used to loot art and take gold, timber and ivory from African colonies; now you just lift semiconductor chip factories and move it right to Arizona in your own country. And then you just tell them China will invade you, so hand it over here now.

U.S. President Joe Biden's administration is also pouring money into building chip factories on American soil, including $6.1 billion to Micron and $6.6 billion for TSMC.

We cry and cry over stated funded industries. How can we compete against the US government pouring money into the chip companies. It's unfair.

The supply chain disruptions caused by the COVID-19 pandemic, the war in Ukraine, and over-reliance on Taiwan are key driving forces behind this move. It’s not just Japan; almost every country has been affected. The U.S. CHIPS Act offers billions in subsidies and tax incentives to promote domestic chip production. Meanwhile, the European Chips Act aims to double the EU's global market share in semiconductors to 20% by 2030.

The AI bubble is plateauing. The US tech sector is sacking people left, right and centre. Government crackdowns are curtailing innovation, governments are demanding 'sovereign' data centres and may wipe out most of the major web 2.0 services over the next few years. AI data centres may not be particularly profitable as AI often produces unreliable data at great expense, and it may never become reliable enough. The cloud is losing its shine. Large data holdings are now seen increasingly as a risky privacy liability rather than an asset to pimp for profit.

The power issue will limit data centres. Any government that rations power to people just so bit barns can have it, will be out of office very rapidly. It will be tough enough to go green with the power we need now, never mind any extra for AI.

AI may not deliver savings. Few technologies actually do. The things people spend money on just changes. But the amount they have to spend rarely decreases.

Self-driving cars may never happen. Autonomy may simply have stalled the push for EVs and driven up prices.

Really advanced chips have limited uses. We have not needed more powerful chips in systems for decades, except for games and AI. And most industrial chips are not particularly high end. Clustering simpler chips may be a better bet looking ahead.

The future may be in the use of lower power, simpler systems. In some cases, decomputerising procedures will save cash, energy and increase resilience. If something works, why spend a tonne of cash to computerise it, if you are just increasing risk and cost, and reducing resilience.

Obviously chip companies will push for governments to give public subsidies to chip companies. But politicians should remember that hype is designed to sell stuff to punters and open the wallets of unsophisticated investors. As those spending money on the metaverse and 3D glasses discovered.

Other technologies may be more important in the future.