Here's a fun masterclass in #AI spin.
Can you read this Financial Times article and see *any* concrete use case, outside all the industry jargon interspersed with interview subjects saying that they're highly skeptical of this tech?
"Most software investors ... are betting the big winners will be the existing powers in the software industry—even if it is still unclear how or when the pay-off from #technology will come."
"Move Over Copilots: Meet the Next Generation..."
https://archive.ph/cp4Fx
I've written a few times now on the environmental and electrical grid cost of related data centres, including one piece pulling from The Guardian's analysis last week (linked below). Goldman Sachs had a research analysis back in June that also showed no likelihood of this AI hype bubble living up to its current $1 trillion investment valuation. So, definitely a bubble - but not one our climate crisis can afford to wait long on bursting.
https://www.theguardian.com/technology/2024/sep/15/data-center-gas-emissions-tech
@MLClark Cory Doctorow has written perceptively that while there are plenty of fun "low-stakes and fault tolerant" applications for AI, these aren't worth the high cost to run the models. AI's judgement can't be relied on for "high-stakes and fault-intolerant" applications without expensive and error-prone backup by expensive humans.
It's a classic bubble. I give it no more than another couple years.