<
https://buttondown.com/maiht3k/archive/the-grimy-residue-of-the-ai-bubble/>
'Q2 earnings are in. According to Pitchbook data, venture capitalists put $27.1
billion into AI in the quarter—half of all VC investment for the time period.
But the past few weeks have shown more and more organizations doubting the
value of AI. The hype is starting to subside, it appears. The banking and
investment giant Goldman Sachs, who at one point was estimating that a quarter
of jobs would be exposed to automation, is now changing its tune and saying
that the technology is nowhere near when it needs to be to replace jobs at that
rate. In a new white paper, Daron Acemoglu, labor economist at MIT, has
estimated that productivity gains from AI will be less than 0.53% over the next
10 years.
With the Goldman Sachs report, the hype bubble is losing gas, and quickly. On
an investment call after its Q2 Earnings Report, Alphabet CEO Sundar Pichai was
grilled with questions about when its big bet on AI—to the tune of $12 billion
per quarter—would pay off. David Kahn at VC giant Sequoia Capital has written
that AI firms need to turn something like $600 billion in revenue for the AI
bet to pay off. It seems like both analysts and investors are starting to lose
patience in the hype cycle.
As a principle, I don't like to say "I told you so". Not only because the
victory lap doesn't do anyone justice, but more so because the destruction
caused in the wake of private equity and venture capital's salad days won't
recoup what's been wrought in the past several years of the AI hype cycle's
rein. In a longer piece from last year, Cory Doctorow asked "What kind of
bubble is AI?", comparing the technology to prior cycles of hype: dot-com and
blockchain, NFT and metaverse. He says:
Tech bubbles come in two varieties: The ones that leave something behind,
and the ones that leave nothing behind. Sometimes, it can be hard to guess
what kind of bubble you’re living through until it pops and you find out the
hard way.
Doctorow thinks that the residue of the bubble popping will be minimal—large
models will no longer be cost-effective to train, but small open-source models
will remain, adept for smaller, better scoped tasks. If that's all that the AI
bubble leaves behind, then we'd be in a better place for society and science.'
Via Wayne Radinsky.
Cheers,
*** Xanni ***
--
mailto:xanni@xanadu.net Andrew Pam
http://xanadu.com.au/ Chief Scientist, Xanadu
https://glasswings.com.au/ Partner, Glass Wings
https://sericyb.com.au/ Manager, Serious Cybernetics
