I'm not a lawyer. I used to run an AI tech company until I sold it and early retired. I don't disagree with you that AI could be very deflationary. I've just been pointing out the other possible future scenario. Given that the discussion has veered back and forth across a very wide road without a lot of definition to what we are really discussing, a big part of the arguing here could be over people thinking of two different problems. Here is my framing of the discussion:
1. Time frame - In this whole discussion, I've been thinking about a time frame of 5-10 years and no longer, with a particular emphasis on potential inflation acceleration in the next 12-36 months.
2. Why this 5-10 year time frame - because we are running huge deficits currently that will have an impact on longer term rates for the next several years. The supply of Treasuries will be higher than historically, unless you go back to WW2, where we last ran deficits as high as a percentage of GDP. Coupled with that is a potentially increasing reluctance of other countries to buy as much Treasuries as historically. It will take 5-10 years of government action to get the debt as a percent of GDP under control. Out past 5-10 years, all bets are off because government actions, AI, and all sorts of other influences come into play.
3. What happens in this 5-10 year time frame - the government will pretty much have to let inflation run a little hot because they will need to get spending as a percentage of GDP down and WW2 is the perfect example of what the US did the last time. Now Trump and Elon are talking about massive spending cuts and I do believe they will do something meaningful, and that will help interest rates and slow down the economy, but the spending problem is so large letting inflation run a little hot must also be part of the tool kit, in concert with spending cuts and growth initiatives. Now letting inflation run hot means NOT hiking rates up as high, because you want to let it run a little hot. But the challenge is the risk I keep pointing to below, the 12-36 month inflation risk. By the way, running hot is letting inflation run at ~2.5-3.5%, not 5%, for example.
4. The 12-36 month risk - juggling spending cuts, the current momentum around wage increases, still relatively low unemployment, interest rates, etc. is a tough balancing act. We all learned in the 1970's and 2020's that if you cut rates too much too fast (or not raise them fast enough) that inflation can re-ignite, and then someone has to come in with draconian measures, ala 20% interest rates, to stop the inflation freight train. I put the odds on this scenario pretty low but not 0%. Conversely, cutting rates too slowly and spending too quickly can put the economy in the ditch, which creates a whole other set of problems that might or might not fuel inflation even more due to even higher deficits. The good news here is that employment is ALWAYS the last to go when you enter a recession, so the current relatively good unemployment numbers really mean nothing. We could be in a recession right now, for all we know, or just a soft landing growth scenario.
I do think you bring up a good point about AI. It won't suddenly produce a massive spike in unemployment. It will be a steady rise over the next 30 years in reducing resource requirements. So that definitely helps offset the potential inflation issue, and raises the odds of lower rates. I'm sure there are other factors that could slow inflation.
Again, I'll just re-iterate what I've now said at least five times. I'm not saying inflation is definitely going to take off again and rates will remain high. I'm saying that betting all your chips on a big rate drop in the near-term is a really bad bet, especially with an inverted yield curve still. In poker terms, your expected return doesn't justify an all-in bet. I personally own some 10 year bonds because I don't believe inflation will spike, but I only put a portion of fixed income into 10 year, and I'm still not much in longer term bonds. Most financial advisors today are also preaching caution on long term bonds.