His as-of-late frequent appearances on Bloomberg, CNBC, and Scott Galloway's podcast, as well as his collaborations with Ed Elson from Galloway's podcast must feel like a great vindication for him already. Quite impressive for a person who is self-taught in that area, just over the past two years.
edit: And, of course, the post is flagged. We must bury our heads in the sand again.
I don't get why there are so many people who need Zitron to be wrong. It reminds me of back in the days when people still liked Elon and every criticism of him or Tesla must be a short-seller spreading lies to move the price.
I think it's pretty easy to slot Ed's predictions into your possibility space even if you're still betting on a different outcome being more likley. Something else is going on.
But have you done the same for Zitron? Because the AI-Bros have consistently been more accurate and Zitron consistently pretty wrong.
Zitron proclaimed in (early) 2024 that LLMs were stuck because there was no more data to use, and synthetic data ruined models, so we were already at the brink of collapse.
AI is a tool, a useful tool, an innovative tool, but technology has introduced innovative tool after innovative tool for years. Apache, Twitter’s Bootstrap, Pull to Refresh, React, Kubernetes, all innovative, all delivered real business value. Nobody invested hundreds of billions of capital into infrastructure for React.
AI has endless use cases. Theres no evidence that these use cases deliver the value necessary to justify any capital expenditure. A technology can change the world and have no standalone business case.
Business is boring. Money in, money out. We are still waiting for Anthropic and OpenAI to show how LLMs have moved the needle on business.
Every single thing that an LLM can be used to deliver value for a business today can be achieved by a cheap to run open weight model that companies could host in house. Where does that leave the trillion+ worth of dollars being buried in the ground?
(also, of course, that there is a massive, insane misallocation of resources to the AI sector compared to the level of demand)
This part:
> I don’t think that’s Ed’s position. As I understand it, Ed’s position is that it has no real economic value.
Is not true. I can agree with one of his positions and disagree with the other, but it really is fair to say he repeatedly and confidently states that ai is useless.
then flywheel effect might kick off and impact carriers and construction companies building data centers and delivering things there.
which impacts investor trust and they might start pulling their money from AI companies, which in turn, they either need to increase prices or downsize their efforts
Long term, you're getting eaten by inflation, but when the power grid, and most cell sites are down, cash is still king.
The real lesson I've learned is that a small generator consumes about 10 gallons of gasoline per day. There's no way to store enough of it to last through a total collapse, even if you could keep it from being stolen.
It’s a long read with well-sourced numbers that make the point well.
But, OpenAI investors and management loosing money and power would be the one part of it I would enjoy.
> I think it’s very unlikely this happens, and if it does, it would be ruinous for everybody involved.
The full snippet is in TFA.
AI is being used to dehumanize people's jobs, provide bad customer service, a bunch of people are being fired for dubious benefit, people are being turned into accountability sinks (the AI is in charge of decisions, but the human takes the blame).
The actual numbers he presents are well sourced and, sadly, correct.
Even if AI is the future of computing, we will have a 9.x economic earthquake headed our way.
Read his articles carefully. Don't be distracted by the salt, the swearing is just there to help the numbers go down easier.
Work it out with a pencil. For those of us with the usual innumeracy about billions, trillions, etc, it can be helpful to draw diagrams on graph paper.
Also, @dang, unflag this please, it is well within the HN ambit.