I think a reasonable summary of the study referenced is that: "AI creates the perception of productivity enhancements far beyond the reality."
Even within the study, there were some participants who saw mild improvements to productivity, but most had a significant drop in productivity. This thread is now full of people telling their story about huge productivity gains they made with AI, but none of the comments contend with the central insight of this study: that these productivity gains are illusions. AI is a product designed to make you value the product.
In matters of personal value, perception is reality, no question. Anyone relying heavily on AI should really be worried that it is mostly a tool for warping their self-perception, one that creates dependency and a false sense of accomplishment. After all, it speaks a highly optimized stream of tokens at you, and you really have to wonder what the optimization goal was.
There were vehicle sourcing credits in the Inflation Reduction Act, which are granted per kWh of battery with sufficient domestic mineral content. Was $4,000/vehicle if I recall.
It’s a bad thing if it increases China’s confidence that it can win. Wars are often launched as soon as one side thinks it has a winning strategy that might slip away over time.
In most modern cars, there is a pressure booster in the power brake system, enabling braking force that can overcome the engine’s horsepower and quickly stop the car. But they are designed to continually maintain the braking force for a limited amount of time—30 seconds or so—after which this boosting ability is depleted. Once that happens, braking must be fully supplied by muscle power via the mechanical backup.
This is challenging if the engine is stuck in a wide open throttle (WOT) state, because the driver must overcome the cars weight in addition to the engine.
For a small car like a Toyota Corolla, this requires a few hundred pounds of downforce on the pedal. For a large 300hp SUV, this could require a thousand pounds of downforce.
As you said, the brakes can bring the car to a stop, but the car will start reaccelerating if the engine isn’t shut off.
The treasury would auction off securities—probably short term bills—in excess of the debt ceiling. I am confident that this auction would succeed, in that dealers would be happy to bid, possibly at high yield. The settlement process following this auction has the Treasury and the dealers both sending out instructions on FedWire to transfer the bill to the dealer and the cash to the Treasury account.
The Fed’s conundrum is: do they possibly break the law, or do they certainly destroy the economy? I am supremely confident they will not destroy the economy. They’ve done gray-area actions to prevent economic catastrophe before. So the settlement will take place, Treasury will have cash, and everyone’s happy.
Then what? A lawsuit I presume, but who would sue the Treasury, and under legal theory? Not a lawyer, but I think you need to show injury to have standing to bring a case. But a case of this magnitude would be decided on political calculations as much as legal principles, and I cannot imagine the Supreme Court wants to be the party clearly responsible for world economic chaos.
And what kind of order could the Supreme Court issue? It would be weeks after the fact—maybe months. Some sort of unwind of the treasury auction, where the dealers put the bills back. (The bills might be expired though.) That would not solve a damn thing, because of the debt ceiling: there will be no money, and no way to raise any, to return to the dealers. I cannot overstate how damaging any such attempt to take back debt would be. This is similar to the collapse of a big bank or exchange, but orders of magnitude more so.
So since there really is not an effective legal remedy available to the court, why would they issue such an impossibility? I think the court will find a way to make the administration look bad, sure, but why would any court force a constitutional crisis? They are in a sense the weakest branch, especially in an emergency situation.
The standard works on the other side as well. A Toyota with a NACS plug can charge on the supercharger network, but also other networks that are also adopting the NACS standard. A non-Tesla car that charges on a non-Tesla network doesn’t generate any obvious revenue for Tesla, unless there are licensing fees. We have not seen the agreements, but I have to assume Tesla gets something.
And Tesla’s ability to do enormous markups only works if they have monopoly pricing power. Perhaps the superchargers, by virtue of being first, will sit at the prime locations, giving Tesla some serious pricing advantages. Or maybe Tesla chargers will be better maintained and overall easier to use, so Tesla may have additional pricing advantages.
But that would fall short of full monopoly-level pricing power, and we see that people have the ability to respond to differentials in fuel prices by driving to cheaper stations. This will be easier in cars that integrate charging prices into navigation.
One thing I would like to know is: will Tesla allow their cars to charge on non-Tesla NACS chargers?
The single most important assumption in this paper is that energy consumption will increase by 2% per year. This kind of exponential growth leads to outlandish estimates for the amount of tidal energy that society will demand.
Energy consumption has decoupled from population growth rates and economic growth.
How much energy will we consume in 1,000 years? Most projections of the population have it stabilizing at around 15 billion. But continuing at its current growth rate (an optimistic assumption I think), gets us to about 150 trillion humans in 1,000 years.
And at 2% growth rate, each of those humans will consume 20,000 times more energy than a circa 2023 human.
Now state of the art technology wastes about 80% of the energy consumed, so this is equivalent to 100,000 times more useful energy consumed per human.
So the physics in this page is a good examination of the surprisingly large compounding effects of unchecked exponential growth.
The US installed 11.8GW solar last year. That is 453MW/every 2 weeks. :-)
Yes, that’s nameplate capacity, so you can’t count on solar to delivery the power to get us over the line by itself. But we are nearly on track to the net zero goal when you add in …
Was recently in a fairly new construction house in a very cold climate state that did not have a full basement below the frost line. There was a partial 8’x8’ corner that sunk down to where the water line came in, but the rest of the footprint was a crawl space. Losing the basement does affect the style of building that is economical to build on top though, mainly due to building codes.
Minivan seems like a great form factor for an EV. Unfortunately, the small number of companies that still make minivans are also the companies that are slowest to adopt EVs.
This is not the way I would do stock valuation. PE is the ratio of future earnings will all its growth to current earnings, discounted. In the future, some of that growth will be in the past, and the PE will be expected to come down. This being Tesla, you could imagine some new fancy product, and not value it strictly as a car company.
Edit: I see that you actually addressed the PE roll down. Missed that on the first read.
Exactly. As an example, a typical 30-year fixed will be halfway paid off after 15 years, and so is about half as sensitive to interest rates as a 30-year interest-only.
This should not be a surprise to the management of First Republic.
It was mainly just the mortgages. The quirk is that they marketed to extremely high-net worth homeowners—I think this is the penthouse condo crowd. In order to attract them, they offered 3% interest-only loans. Without principal pay down, these mortgages are more sensitive to interest rate rises, leaving the bank with roughly $35bn in losses as interest rates have been rising. They have been insolvent for several months, but a typical bank didn’t have losses on their mortgages which were nearly so steep.
As compared to SVB, this is the same basic situation: interest rate losses led to insolvency, which could be temporarily ignored because they were “small” banks. However, once deposits started fleeing, the losses could not be ignored when they needed to sell the impaired assets for actual cash.
They have been in limbo for a few weeks thanks only to the injection of $30bn from other banks.
Another similarity in the two banks’ situations is that the same catalyst of rising interest rates cause asset losses and drive deposit flight.
It’s a weird quirk of accounting that they are allowed to ignore these losses for the life of the assets. But the other extreme is weird too, because sometimes the market value of assets can undergo a “V” shaped dip before recovering, and it would be bad to make a bank insolvent because some flash crash. The accounting rules try to split the difference by letting the bank partition its assets into buckets that take losses immediately, or at the end of life of the asset. This is an easily abused system.
It seems, hopefully, that the three failures were exceptionally badly run banks, and that this doesn’t indicate a wider wave of bank failures. Not yet.
Primary energy infrastructure planning takes place at a 20-30 year timeframe. GDP for 25 years is nearly $600 trillion USD, so the Master Plan is approximately 1-2% of available resources.
Even within the study, there were some participants who saw mild improvements to productivity, but most had a significant drop in productivity. This thread is now full of people telling their story about huge productivity gains they made with AI, but none of the comments contend with the central insight of this study: that these productivity gains are illusions. AI is a product designed to make you value the product.
In matters of personal value, perception is reality, no question. Anyone relying heavily on AI should really be worried that it is mostly a tool for warping their self-perception, one that creates dependency and a false sense of accomplishment. After all, it speaks a highly optimized stream of tokens at you, and you really have to wonder what the optimization goal was.