Goldman Sachs says AI's impact on the US economy was "basically zero" last year

[Endymio]

"Experts" publish a couple of such reports every week.
Usually one claims AI is a bubble that will go bust soon, and another claims AI will take some huge percentage of jobs within a year or two.

The AI 'bubble' will burst the same way the Internet dot com bubble did -- even at the very peak of the decline, Internet penetration and use, and spending on Internet-related goods and services continued to increase. AI isn't Dutch tulips: it's not going away simply because markets will eventually see a correction.

 

[emmzo]

You missed the point. Before the worker-displacing technology of the industrial revolution, the average person worked 12+ hours a day to be afford to afford no more than 1 or 2 sets of clothing, a diet that consisted of unflavored boiled grain, morning and evening, and a home that consisted of a one-room hand-made hut. Today, a middle-class family lives in most ways better than a lord of four centuries ago.

It's impossible to predict all the ways that AI will improve our standard of living. But some predictions are easy -- in 25 years, the average person will afford better healthcare than the billionaires of today have, with a team of medical specialists monitoring your vital signs and blood chemistry 24x7, able to instantly diagnose illnesses and conditions before you're aware of them yourself. A decade after that, young children will regard with disbelieving awe being taught that we once allowed people to drive their own vehicles, killing a million people each and every year, and maiming millions more.

Just two of the two hundred thousand ways AI will improve life, to the point that, one day a future "scoffer" can mock our current standard of living, just as you mock those 16th century peasants, happy that they don't have to choose between a pair of socks and going hungry.

Well, actually, before the Industrial revolution people worked less hours per year as from historical evidence, work was mostly seasonal. Also, they had a more diverse diet than just "boiled grain". I get you overstatements to prove your point, however things are more nuanced. Medieval life was not that bad as most people imagine, definitely not way worse than 19th century life. Sure, the average guy today lives better than a 16th century lord, it doesn't mean only technological evolution brought population welfare. A whole bunch of advancements in different other fields helped (sociology, economy, health and sanitation, etc.) and you must also remember evolution is not always a straight line. Change is opportunistic, branching, and full of contingencies, that's why I wouldn't bet on "sure" predictions for AI.

 

[Endymio]

Well, actually, before the Industrial revolution people worked less hours per year as from historical evidence, work was mostly seasonal.

This a long-since debunked claim:

"In 1986 economist Gregory Clark wrote a working paper that contained an estimate [of 150 work days/year in Medieval society]. It doesn't appear he published it, but it got cited. He actually did publish a new paper in 2018 raising that number up to an estimate of 250-300 days ... However, in the meantime a popular book was published by a sociologist that used that earlier lower number of 150. "The Overworked American: The Unexpected Decline of Leisure". Since then this number went wild on the Internet...."

I've seen similar reports since 1986: all are based on confusing a villein's labor obligation to his lord's desmense (effectively his rent cost) for the total amount of work a Medieval peasant needed to perform to survive.

Sure, the average guy today lives better than a 16th century lord, it doesn't mean only technological evolution brought population welfare. A whole bunch of advancements in different other fields helped (sociology, economy, health and sanitation, etc.)

All advances which were enabled by technology. When agricultural productivity is so low that 80%+ of the population needs to be employed as farmers, simply to keep everyone fed, it doesn't leave much labor left over to advance other fields.

 

[Theinsanegamer]

The AI 'bubble' will burst the same way the Internet dot com bubble did -- even at the very peak of the decline, Internet penetration and use, and spending on Internet-related goods and services continued to increase. AI isn't Dutch tulips: it's not going away simply because markets will eventually see a correction.

Funny how you leave out the collapse of big companies like worldcom in this rose tinted view of history. Tens of thousands lost their jobs and investors lost billions.

 

[BobDolesGhost]

And water deals. Many of the permits are being revoked before they actually come online. Things like Infrassound are a big concern for these things and they're running out of places to build them. So you are having big, noisy data centers right next to very wealthy neighborhoods and these wealthy individuals do not like the impact in their property value. They actually paid a lot for a nice home and don't want to live next to a noisy, industrial center

To space we go!! heh

 

[emmzo]

This a long-since debunked claim:

"In 1986 economist Gregory Clark wrote a working paper that contained an estimate [of 150 work days/year in Medieval society]. It doesn't appear he published it, but it got cited. He actually did publish a new paper in 2018 raising that number up to an estimate of 250-300 days ... However, in the meantime a popular book was published by a sociologist that used that earlier lower number of 150. "The Overworked American: The Unexpected Decline of Leisure". Since then this number went wild on the Internet...."

I've seen similar reports since 1986: all are based on confusing a villein's labor obligation to his lord's desmense (effectively his rent cost) for the total amount of work a Medieval peasant needed to perform to survive.


All advances which were enabled by technology. When agricultural productivity is so low that 80%+ of the population needs to be employed as farmers, simply to keep everyone fed, it doesn't leave much labor left over to advance other fields.

Well, debunked medieval studies or not, statements like they all worked 12 or more hours a day and only ate boiled grain are gross exaggerations. Also, this is not a debate about how technology made our lifes better, duh, but about how it's not a straight line with clear results. Ancient China, Greece, Rome achievements were partly forgotten and later rediscovered and quality of life varied. More to the point, AI could be a tool for both good and bad, it will surely change society, but it's not guaranteed higher quality of life regardless of politics, societal and economical changes.

 

[madboyv1]

To space we go!! heh

Space is for escaping the corruption of capitalism, not datacenters... XD

Lots of "free" energy to convert from solar. Whether you'd be able to reasonably attach enough panels to match the power requirements similar to a terrestrial datacenter would yet to be tested... However, for how much cooling datacenters need, the vacuum of space makes a very poor environment for traditional (read: cheaper, standard off the shelf) cooling solutions for that level of performance.

 

[Endymio]

Funny how you leave out the collapse of big companies like worldcom in this rose tinted view of history. Tens of thousands lost their jobs and investors lost billions.

No, you merely misunderstood my post. Yes, eventually some AI firms will go bankrupt, but even as that happens, the AI industry as a whole will continue to grow and thrive.

Oh, and Worldcom's collapse was due to accounting fraud, and only peripherally related to the dot com bubble. And despite the largest fraud in US history coupled with the dot com downturn, half of Worldcom survived and kept going as MCI, eventually to be acquired by Verizon.

 

[ZedRM]

This statement from Goldman-Sachs proves they are either complete out of touch with reality or they are flat-out fabricating information.

 

[Endymio]

We had this situation 150 years ago.

Back in the early days if the 1st Industrial revolution, they paid the people so poorly that none was able to buy the products who have been produced in these brand new factories.

Who do you believe were buying those products then? It certainly wasn't indigent peasant farmers keeping those factories in business, nor were the nobility filling their castles and manor homes with cheap gewgaws from these new assembly lines.

I don't know where disinformation like this arises, but during the Industrial Revolution, people flocked by the millions from rural farms to those urban factory jobs, specifically because the pay **was** better than what they could expect to receive anywhere else.

 

[GoldenGoat]

I think AI is the future, but it's not quite there yet. The amount they are spending is overboard for premature technology. But it is a race and they all want to win at any cost.