Goldman Sachs and Economists are Backtracking on Generative AI’s Value – Is BigTech’s investment in Generative AI warranted? Trillion dollar bets that look increasingly like gambling.
https://www.ai-supremacy.com/p/goldman-sachs-and-economists-are?
Goldman Sachs and Economists are Backtracking on Generative AI’s Value – Is BigTech’s investment in Generative AI warranted? Trillion dollar bets that look increasingly like gambling.
https://www.ai-supremacy.com/p/goldman-sachs-and-economists-are?
11 Comments
Submission Statement
The AI Investor hype bubble always seemed ultimately doomed. AI will be profoundly deflationary, and will likely lead us to end up dominated by a very different economic system than today, with a far smaller role for capitalism, stock markets, and investors.
This article is interesting as it neatly illustrates the schizophrenia at the heart of the AI investor worldview. On the one hand, it berates people who made claims that 300 million jobs would be automated – *because they’ve failed to live up to that “promise” to AI investors fast enough.*
What you never see is anyone joining the dots, and asking what sort of economic model society will evolve to when job automation is at that scale. (*Hint: It probably won’t have much room for high stock market or property prices, or prosperous investors).*
The main reason I didn’t invest in nvidia was because of the potential conflict between Taiwan and China, can talk yourself out of many things
All systems have inertia and the global economic system has a lot of it; the AI revolution was always going to hit the tar as it tries to roll out and find applications. Doesn’t mean it’s DOA. All booms will inevitably overshoot due to irrational exuberance before they correct, it always works this way.
I agree with the statement that the kinds of problems investors are looking to solve with AI and the truly transformative changes won’t happen until about 10 years or so if at all and that the productivity gains across the economy won’t be as dramatic as people might be expecting.
“Outside of ChatGPT and Character.AI, consumer Generative AI apps haven’t even been a big hit with, with many struggling to scale.”
This part made me laugh. Imagine saying, outside of ios and Android, mobile operating systems haven’t been a big hit, and then using that logic to software cell phones are a dud.
GPT4 was released in March 2023 and was the first actual LLM product. To expect anything spectacular a year later is ridiculous.
It doesn’t matter every company on earth has already gone all in on the AI marketing craze.
I had to do some market research and I looked into around 140 different e-commerce businesses last week.
Over 80% of them had the same boiler plate buzzwords regarding AI, innovation, automation, time savings..
It literally read like AI actually created it for them because it was literally the same shit across dozens and dozens of websites.
People are finally starting to realize that generative AI simply can’t continue to get “better” at its current pace unless models are made *dramatically* more efficient, and it’s not clear how possible that is. The power requirements are too high and there isn’t enough training data.
What you’ll see is marginal improvements in broad-application LLMs and slightly more (though still not dramatic) improvement in use case-specific LLMs. But there is still no clear path to anything resembling profitability for consumer AI, and even enterprise AI faces cost problems.
They couldn’t possibly be hedging the market right? Surely that isn’t something they’re capable of /s
When we finally remove the GPU requirement from modern neural networks, we will see the true utility of non-transformer based machine learning as a tool of humankind instead of a replacement. The current undue hype will define a generation.
AI is completely useless and people are finally realizing.
If the goal was to replace the workforce it might be a bit and a lot more money. But I don’t see how AI isn’t as big a disruption as the personal computer.