I have come to the conclusion that the most essential element of the Silicon Valley ideology is its collective faith in technological acceleration. More than the mix of libertarianism and tech determinism that Richard Barbrook and Andy Cameron described as the “Californian Ideology” in 1995, more than the breakdown of donations between Democratic and Republican candidates, the driving ideological conviction among the denizens of Silicon Valley is that the rate of technological innovation is accelerating.
Silicon Valley is the Church of Moore’s Law.
Tech acceleration is at the center of the entire Silicon Valley worldview. This dates back at least to 1993, when WIRED magazine declared in its opening pages that the digital revolution had put us on the cusp of “social changes so profound their only parallel is probably the discovery of fire.”
From this central assumption, it follows that:
- The entrepreneurs, engineers, and investors who travel in Silicon Valley circles (regardless of their actual zip code) are the driving force of history. The decisions that will reverberate through the decades are made on Sand Hill Road, not Capitol Hill or Brussels.
- Existing institutions – businesses, media, government, etc – will all soon be disrupted. The old rules on longer apply. New rules are being created.
- You judge the success of a business based on its growth, not its profit. The future is in the process of being created. The companies that monopolize an emerging sector are the companies that will define that future.
- And, finally, the future will be abundant. Everyone will reap the benefits brought about by Silicon Valley. So it is really not worth worrying about inequality, or about the distribution of those social benefits.
I have spent the past few years reading the entire back catalog of WIRED magazine, as part of a project tracing the history of the digital future. The surprising part is how little has changed in the intervening decades. Today’s tech barons speak and behave much like the libertarian tech-optimist ideologues of 30 years ago. And yet there is ample evidence that the pace of technological innovation has, if anything, slowed down in recent years.
Belief in the immutable nature of Moore’s Law is a matter of identity in Silicon Valley circles. But it is also a political project. Faith in Moore’s Law implies a distrust of government – government is too slow and cumbersome. Legislators and federal agencies cannot hope to keep pace with innovation, so we are better off demanding that they stay out of the way.
So I suppose it is unsurprising that, even as the pace of “internet time” seems to have stalled over the past decade or so, our tech titans and Silicon Valley evangelists have remained insistently committed to the idea of Moore’s Law. The digerati have been propping up the corpse of Moore’s Law, Weekend at Bernie’s-style, for years now.
Meanwhile, the only thing that has exponentially increased is wealth inequality. We now have centi-billionaires like Elon Musk, and a host of social problems that are the consequence of extreme wealth inequality. We ought to distrust the ideology that tells us this is nothing to worth worrying about.
I imagine most Crooked Timber readers are already familiar with Moore’s Law. But, as a refresher, let me explain where it originated.
Writing in 1965, Gordon Moore predicted that the number of transistors you could fit on a silicon chip would double roughly every two years, while the price of those chips would be cut in half at the same rate. This prediction turned out to be durable for decades to come. The tech booms of the past fifty years – from mainframe computing to personal computers to the Internet, World Wide Web, laptops, cell phones, smartphones, and tablets – all have Moore’s Law in common. Chips kept getting faster and cheaper, allowing for successive generations of digital technologies.
In The Long Arm of Moore’s Law, Cyrus C. Mody examines Moore’s Law as a social fact. Within the developing computer industry, Moore’s Law functioned as a coordination point and an industrial target (“More an engine than a camera”). Developing new computer chips is a capital-intensive process. For decades, Moore’s Law held true because a vast array of actors made it true. Various parts of the emerging industry (one that, for much of its history, could rely on a steady stream of public funding) worked to meet the 18-24 month development cycle that Gordon Moore had effectively set as a goal. The investment in new chip production facilities was worthwhile, because chip manufacturers had a clear sense of the timeline and price points that computer manufacturers were working towards.
Mody’s work focuses on the social fact of Moore’s Law within the computer industry, though. The ideological claim made by tech acceleration is that a similar exponential curve is underway at a civilizational level. Drawing from the work of futurists like Alvin Toffler and George Gilder, Silicon Valley in the 1990s developed a heady, brash certainty that a new day had dawned. Gilder, for instance, insisted that there would be parallels to Moore’s Law in telecommunications infrastructure. (He was wrong, but he was confident. And that confidence was well-rewarded).
In his 2005 book, What the Dormouse Said, John Markoff described Moore’s Law as “Silicon Valley’s defining principle. … [Moore’s Law] dictated that nothing stays the same for more than a moment; no technology is safe from its successor; costs fall and computing power increases not at a constant rate but exponentially: If you’re not running on what became known as “Internet time,” you’re falling behind.” It is, in other words, both a technical prediction about processor capacity and a deeply-held cultural belief about the pace of the digital revolution.
I have written about this elsewhere, but the pace of digital change was really quite noteworthy back then. The internet of 1993 bore little resemblance to the internet of 1998 or 2003. When technologists insisted that Congress and federal agencies could not regulate the emerging digital media environment that they did not understand it because it was evolving too fast, they were self-serving, but they also kind of had a point.
The social evidence of Moore’s Law was also much more visible as a mass cultural phenomenon back then. I recently spent some time combing through the first five years of product reviews in WIRED magazine. The magazine celebrated the latest in high-end consumer goods — $6,000 laptops, $8,000 flat screen televisions, $2,500 video cameras. If you had struck gold in the dotcom gold rush, you could afford those items today. If not, you could wait a couple years and buy a superior product for half the price. “Last year’s unattainable conspicuous consumption is the next year’s Christmas gift.”
This is just one of the ways that the mythology surrounding Moore’s Law has faded in recent years. The desktop computer I brought to campus in September 1997 could barely run Windows XP when it was released in 2001. By comparison, the iPhone 15 was just announced last week. It costs about the same as the iPhone 14. It replaces the iPhone 14’s lightning charging plug with a USB-C charging plug. If “Moore’s Law” was once synonymous with consumer technology getting dramatically more powerful and cheaper at a constant rate, then that version of Moore’s Law is now little more than a faded memory.
Faith in Moore’s Law as an inevitable social force has remained strong within Silicon Valley circles, though. The victors of that era – the young entrepreneurs who made their first millions in the first dotcom boom – became the investors and thought-leaders of the decades that followed. There is a self-justifying logic to today’s Silicon Valley ideology: the early entrepreneurs must have been right all along; Their insights being the essential ingredient of their much-deserved success. It’s survivor bias, plus the influence that comes from decades of unchecked wealth accumulation.
A decade ago, I was myself a believer in tech acceleration and Moore’s Law as a broader sociocultural phenomenon. I came to that perspective honestly: I was a graduate student in the ‘00s, and the pace of digital change was so fast that the research literature could barely keep up.
I became disabused of that notion in 2018, when I had my first opportunity to read the entire WIRED magazine back catalog. It became clear then that the Internet of 2018 was really quite similar to the Internet of 2013. Internet time had, effectively, slowed down.
The Internet of 2023 has many strong similarities to the Internet of 2013. It’s, for the most part, the same large companies, offering the same (enshittified) services. The intervening decade has featured plenty of grand pronouncements that a new era has just arrived (3D printing, robotics, self-driving cars, Google Glass, cryptocurrency, the Internet of Things, wearable electronics, Big Data, cryptocurrency again, the metaverse, etc etc etc). And while it is may still be too early to declare any one of these digital futures a permanent failure, we can confidently conclude at this point that the pace of their arrival is no longer speeding up as their promoters insist.
And yet 2023 has been something of a banner year for the acolytes of technological acceleration. Generative AI tools like Dall-E2 and ChatGPT seem solid in a way that “Web3” and the Metaverse never quite did. This time may be different. And even if ChatGPT starts to look like a cheap parlor trick after a few weeks of use, just think how much better the next version will be. Perhaps the past decade was just an interlude. Maybe the future is about to arrive.
Sam Altman, OpenAI’s President, wrote a sort of manifesto two years ago. It was titled “Moore’s Law for Everything.” It is quite a revealing document. If Silicon Valley is the church of Moore’s Law, then Altman, it seems, is a Cardinal.
Altman proclaims that the coming Artificial Intelligence revolution is “unstoppable,” part and parcel of the long arc of tech acceleration that we are living through. He writes:
“On a zoomed-out time scale, technological progress follows an exponential curve. Compare how the world looked 15 years ago (no smartphones, really), 150 years ago (no combustion engine, no home electricity), 1,500 years ago (no industrial machines), and 15,000 years ago (no agriculture).”
Fifteen years is an awkward anchor point. If he had instead chosen ten years, then he would have had little to point to. And it’s worth noting that, on a more reasonable timescale, the pace of technological progress doesn’t seem to be increasing at all. Back in 1997, a pair of futurists wrote a gaudy scenario for WIRED titled “The Long Boom.” They declared that 1980-2020 would feature a cluster of history-making scientific breakthroughs that would set us all on the path to global prosperity. (the authors have since taken stock of their predictions and decided “yep! Nailed it.)
1980-2020 encompass the rise of the personal computer, the World Wide Web, cell phones, and smartphones. Those are certainly nontrivial breakthroughs in the lives of mass consumers. But is that actually an increase in the rate of technological innovation? 1940-1980 was the television age and the nuclear bomb and The Pill. 1900-1940 was the automobile, the radio, the earliest commercial airlines, and the telephone. 1860-1900 included mass electrification and the end of slavery. 1820-1860 is basically the start of the industrial revolution. Moore’s Law only seems special because it’s the era we’ve been living through.
Altman insists that artificial general intelligence is inevitable. He also predicts that the first trillionaire will be someone involved in the AI revolution. And he speaks as though this will be a good thing. He urges his readers to preoccupy themselves not with how we regulate and shape his new industry, but how we prepare and remold society to adapt to the natural next-step in technological evolution.
Altman has also insisted that he welcomes regulation. OpenAI has invested heavily in educating Congress about the promise and perils of this new, world-changing, general purpose technology. But OpenAI opposes regulation by existing agencies. The company only supports regulation if it gets to select the regulator and influence the rules.
The most important difference between the 90s tech boom and today is the relative power of Silicon Valley. “Big Tech” in the 1990s simply wasn’t all that big yet. And the evidence of Moore’s Law – not as a prediction about the power and price of computer chips, but as a broader social phenomenon – seemed to be everywhere. The technologists were gaining influence, the technologies were moving to the center of daily life. It stood to reason that the future would be different and, hopefully, better.
We care today about the ideology of Silicon Valley both because of how digital technologies have been integrated into the social, economic, and political realms, and also because of the financial power of the tech titans themselves. Elon Musk is a centi-billionaire, and he is not the only one. In the months before FTX was revealed to be an elaborate ponzi scheme, Sam Bankman-Fried had enough wealth that he openly considered buying the island nation of Nauru. Those levels of wealth are deleterious to a functional democracy. Ignoring wealth inequality made the problem much worse.
We will need to address wealth inequality in the years to come. And we should do so knowing that Silicon Valley is ideologically unwilling to help.
{ 53 comments }
joeyjoejoe 11.02.23 at 3:16 pm
“We now have centi-billionaires like Elon Musk, and a host of social problems that are the consequence of extreme wealth inequality. ”
Elon Musk is worth $225 billion.
If we take $200 billion from him, and split it amongst 340 million americans, each citizen gets…$588.
What social problems are being caused by my lack of $588?
Joe
DCA 11.02.23 at 3:51 pm
There is a lot of merit in this argument, but I’ll point out that Moore’s Law isn’t
even a half of the internet revolution, only a third: there are two other technologies
that have had to be developed to make computers, and the internet, what they are. And they have different histories.
One is storage, as represented by magnetic disk drives, MDD’s, still the main storage medium, but gradually being replaced by electronic memory. Here it is the ability to magnetize very small regions–so, again the microworld, but iron and its relatives, not silicon. Moore’s law turned out to apply here also, though it might not have, and I believe some fundamental limits have already been reached.
The other, with the really different story, is moving bits around: here fiber optics has been key. This did revolutionize telecommunications, and that is the sector that developed it (Bell Labs, the UK Post Office) or offered possible profits to other sectors (Corning Glass). In the 1970’s nobody knew if fiber could be made to work (Bell was also pursuing high-capacity microwave waveguides)–it turned out that it could. Unlike silicon and magnetism, there isn’t any “just do this and the capacity will go up” trick, which is why you see Google laying undersea cables.
So Silicon Valley happens to have hit a triplet of revolutionary developments that really has no parallel anywhere else in the history of technology. They really have been born on third base.
Brett 11.02.23 at 4:19 pm
Saying that we’ve had exponential wealth inequality is wrong. Looking at FRED charts, the share of Total Net Worth held by the Top 1% of households in the US has gone from 28.3% in 1998 to 32% in 2022 – an increase, but an increase of 12-13% rather than something far more over the past 25 years. The visibility of a handful of tech super-billionaires tends to skew the perception of this – even most of those gains were in the past 2-3 years (between 1998 and 2020, that top share increased by only 6% on average).
Share of Total Assets held by the Top 1% tells a similar story – 25% to 29% from 1998-2022. Not nothing certainly, but not particularly impressive.
That does suggest we don’t live in an era of overall technological acceleration, but it also feels like an endorsement of Silicon Valley. In an era of much slower overall productivity growth and technological advancement, the areas they held direct sway over – telecommunications, software, computers – all had by far the most technological advancement and productivity growth.
Metalawyer 11.02.23 at 4:36 pm
This excellent piece stumbles on one point. It argues that ” Moore’s Law only seems special because it’s the era we’ve been living through.”
While I think that’s pretty debatable, it might be right. But there’s no way to tell, because Mr. Karp fails to give any working definition or benchmark to define “special”. And that’s kind of the axis of the argument, no? If we go with “rate of technological innovation”, I’d argue that the shift from the desktop computer to the laptop to the cell phone demonstrates significant technological innovation. I’d also argue that the Moore’s law revolution is leading to signficant advances in biotech, warfare, and social structures, mostly tangentially from the gross increases in inequality.
But it all depends on the definition. I’d like to agree, but there’s no basis for it.
And to be clear, I agree with nearly every other point he makes.
Chetan Murthy 11.02.23 at 7:25 pm
DCA: In “the biz” (I’ve been a professional computer programmer since 1986, PhD CS ’90), we think of Moore’s Law as not just about transistor density, but about the general exponential growth in capability along many dimensions. So your two (storage density, network bandwidth) both are included. More generally, there’s a famous quote from Jim Gray (father of transaction-processing) where he says (paraphrasing) “If you want to create a successful new business, find something that is costly today, but forecast to drop in price a lot (ideally, exponentially) and over use it in your business“. This is, in a nutshell, what the obsession with Moore’s Law is about: you find some way to use a currently-costly resource that is getting cheaper like crazy, such that as it gets cheaper, you turn a profit.
But I think there’s something missing from this focus on Moore’s Law: a lot of what happens in Sili Valley isn’t about Moore’s Law, except barely-tangentially. Consider Uber: its forecasted profitability wasn’t based on dramatically cheaper {computing,mobile comms}. You can say the same for food delivery. Instead, cheaper comms just made one small part of taxi-dispatching cheaper and more convenient. And so Uber used that as a crowbar to get into that business. But intrinsically, it wasn’t a unique crowbar, and really, the actual profitability story was all based on driving everybody else out of the market, and then jacking up rates. In short, it wasn’t about Moore’s Law, but about good ol’ monopoly. You can see the same in social media, and in ad networks: yes, initially the now-incumbents used new dramatically-cheapening technology to get their start. But pretty soon, they were using standard monopolistic practices to stay on top.
Heck, it’s a story as old as Microsoft and Windows, right? Or IBM and the Seven Dwarves.
Jake Gibson 11.02.23 at 8:02 pm
Is it just me or does it seem like Moore’s Law is an intellectual Ponzi scheme.
ancien regime 11.02.23 at 8:15 pm
Odd time to be deciding tech innovation has slowed down, being as we are in the early stages of what could be a transformative technology. Brought to us, by the way, by Moore’s Law, in the form of massive arrays of high performance GPUs.
Maybe AI ends up a bust, but it might be safer to wait a year or two before building on the premise that tech has become stagnant and Moore’s Law is dead.
By the way, Wired Magazine isn’t a particularly good source for understanding the tech industry, then or now. By the time it showed up the only underlying philosophy in tech was to make as much money as possible, preferably by creating a monopoly and/or selling shares to credulous investors. The techno-utopian stuff was primarily window-dressing. That was how the industry wanted to see itself and how it wanted others to see it, but the primary actual motivator was almost always direct self-interest.
That doesn’t mean the marks didn’t believe the story. Take a look at any Yahoo internet stock forum from the late 90s. (Or if they’re not around anymore you can substitute almost any crypto subreddit from 2020-2022.) Marks were supposed to believe the story.
But believing the Wired techno-utopian world view actually drove behavior in tech is a mistake. Money drove behavior. Still does. See, e.g., Sam Altman, who talks the utopian talk but whose actual behavior is invariably consistent with maximizing Sam Altman’s self-interest.
By the way, if you’ve read through Wired you must have found the Bill Joy 2000 piece on why technology was going to kill everyone. That actually was influential in tech. Joy was in the pantheon. Him saying this was all going to end badly was a major shock. And eerily timed, since that issue was on the stands when the internet bubble popped.
Albert 11.02.23 at 9:27 pm
Brett @3:
The 1% is a bad proxy for “extreme wealth” of the kind people have in mind thinking about Silicon Valley. It “only” takes $10-15 million of wealth to be in the top 1%.
How about looking at the relative growth of the top .01% ($227m) and .001% *(1b)?
Alex SL 11.02.23 at 10:11 pm
Yes to all of this. There are technologies that really do make a radical difference. In recent decades, the internet and the cell phone come to mind. But, and unfortunately I do not remember who the author of the piece was where I saw that pointed out first, if there is such a thing as the singularity, it was the broad deployment of washing machines, dish washers, vaccuum cleaners, and similar appliances. Nothing has revolutionised life as much as these, sharply demarcating the world before from the world after, in how they have freed humans (and, under patriarchy, women in particular) from suffocating drudgery. Any realistically achievable AI, as opposed to SciFi hallucinations, won’t get close to that, ever.
A key thread running through the post that I really want to stress is that in real life, progress is never exponential. It always runs into limits and diminishing returns. If there is an exponential curve, that simply means that we are looking at the first half of a logistic function, or s-shaped saturation curve. The reason is simple: the material world as it actually is outside of fantasy, AKA physics. There won’t be infinity processing speed, the question is simply when physical limits will be reached. There won’t be infinity people on this planet, the question is only if the saturation part of the curve is caused by family planning or by mass starvation. There won’t be infinite economic growth on this planet either, because even with a few percent growth per year, merely the waste heat from whatever those economic activities would be would boil off the oceans within few centuries from now.
The main underlying question is if the relevant tech bros believe in their exponential and accelerationalist nonsense or if they only claim to do so because it is a load-bearing part of an ideology that is politically useful to them. I am increasingly coming to the belief that most of them are true believers. That is partly because social media give us unprecedented insight into what some of them say in unguarded moments. What they post on Twitter etc. proves they really aren’t very smart or well-informed, and IMO that makes it more likely that they really believe what they say instead of cynically playing ten-dimensional chess.
joeyjoejoe,
The problem with billionaires is not primarily that they suck up money that other’s don’t have, although of course it remains morally relevant they only have that much money because did suck it up unfairly through, in the best case scenario, a combination of underpaying workers and market failure, mostly by establishing monopoly or monopsy power. (Worse case scenarios are stock price manipulation, defrauding retail investors, and similarly outright illegal activities.) The existence of a billionaire is by itself evidence of societal and market failure, because nobody can earn, in the sense of morally deserve as a reward for their contributions to society that much money. Nobody contributes to society 1000x as much as the average worker, and what they have supposedly contributed was >99.9% the expropriated contributions of their employees.
But again, a billionaire having 500 bucks that I don’t have isn’t the real problem. The real problem with billionaires is that no single human should have the degree of power and influence over others that this kind of wealth brings. It is undermining democratic governance and other people’s freedom.
John Q 11.02.23 at 10:22 pm
The crucial thing about technological progress in the era of Moore’s Law is that it is highly uneven. Once there is a digital technology to do something, such as videoconferencing, the cost drops rapidly to zero. Meanwhile, transport which used to be the defining form of technological progress (Steam Age, Jet Age, Space Age etc) has been virtually stagnant. Maybe self-driving cars will mark a new leap forward, but I’ve been burned making that prediction.
So, the issue now isn’t so much what is happening at the chip level. It’s whether a way can be found to use a digital technology to do work currently done as analog.
John Q 11.02.23 at 10:27 pm
Following up on this point as regards inequality. As with technological progress, the growth in inequality in the Reagan years was economy-wide. The fortunes being made in tech since the late 1990s are spectacular, but more confined to that sector, especially since the GFC put a crimp in the finance sector.
J-D 11.02.23 at 11:29 pm
There are many people who have problems which would be substantially lessened by having an extra $588, even if you don’t happen to be one of them, but at least as importantly you’re overlooking the ways that social problems can be caused not just by poor people’s poverty but also by rich people’s wealth. The way resources are distributed to the production of goods and the provision of services is heavily influenced by how much money people are willing to spend to purchase those goods and services; the greater the inequality of wealth, the more the operation of the economy is skewed towards providing what rich people want in preference to what poor people want. To take just one simple example, people starve partly because resources which could have been used to distribute food to the hungry are instead being used for other purposes preferred by rich people.
nastywoman 11.03.23 at 8:22 am
and about:
‘What social problems are being caused by my lack of $588’
you don’t have the $588 you wanted to give to the homeless person who made Elon Musk talk about ‘the Zombie Apocalypse’ in downtown San Fransisco – and as Elon Musk said that ‘the Zombie Apocalypse’ in San Francisco made him by Twitter –
and he –
until now lost over twenty Billion of his wealth by buying Twitter – he can’t give you $588
– which creates the ‘Social Problem’ that you can’t give it to the homeless person who made Elon Musk buy Twitter.
Right?
Right?!
Trader Joe 11.03.23 at 10:49 am
A couple of points:
First – you mention Musk’s wealth on several occasions but he is only peripherally a “tech” entrepreneur as its usually defined. He’s really more indicative of someone who harnessed silicon valley’s tech and applied it to more pedestrian pursuits – payments first and then later automobiles (and lately Twitter, but that’s more a value destruction). While his two main companies Tesla and PayPal invented many processes that are quite valuable, they invented minimal technology in the traditional R&D sense.
Second your comments have a heavy survivor bias to them. Show me a Google and I’ll show you a list of Netscapes, InfoSeeks and half a dozen others that were chewed up and spit out by the lack of progress you cite. I’d assert something canibalized all of the dead companies on the road to the internet superhighway and that something was decidedly technological advancement even if it may not have fully realized Moore’s law.
c1ue 11.03.23 at 11:48 am
An excellent writeup of the disparity between Silicon Valley hype and actual performance.
I would note that this is largely because of fundamental misunderstanding of what really underpins Moore’s law: that information has zero mass.
Zero mass means all you need to transmit, process or store information is a bit, and so sizing scale down works. In reality, it was never quite that clean but the problem wasn’t in the information part, it was in the definitions of the transistor state changes that the information was based on.
And the problem with this fundamental lack of understanding of information vs pretty much anything else is the notion that anything else can scale like information.
Energy, food, health care, jobs, whatever – these are all categories involving mass.
They will never scale like a zero mass category.
Furthermore – once we get past the truly quantum leaps in productivity from compute scaling (compute = information processing) of the 1980s to 2000s, what has Silicon Valley wrought since then?
Cured cancer? No
Society-wide peace, progress or wealth? No
What we got is very close to Thiel’s comment about flying cars promised but 140 characters delivered: Twitter, Ebay, Facebook. Or put another way: the dystopian cyberpunk worlds of megacorporations and various flavors of crooks/enforcers only it turns out this is a description of the online world.
DavidtheK 11.03.23 at 1:05 pm
Lots of interesting things to think about here. Paul Campos at LGM has suggested 70 year chunks might give you a better sense of how technological changes and scientific ideas reflect in the way a median person lives. So 1950 compared to today, 1880 to 1950, 1810 to 1880, a few hops and you are back to where changes need millenia.
Bryan Steele 11.03.23 at 2:39 pm
The question has been raised “What social problems are being caused by my lack of $588?”
The conflict between centralized wealth and democracy derives from the undue influence on the political system by those who control disproportionately oversized wealth. This is why Thomas Jefferson favored the estate tax, because he feared concentrated wealth passed from one generation to the next would be used to disproportionately influence the political process.
Ebenezer Scrooge 11.03.23 at 3:43 pm
I don’t think that technology has much to do with Silicon Valley, and hasn’t had much to do with it for a decade or so. “Technology” is a magic totem that legitimates what is really going on, much like the Bible legitimates hard-right Christianism. Our Silicon Valley overlords may think that they have a faith in “technology.” But their true faith is in a certain form of disruptive capitalism: mostly a quest for the monopoly power inherent in unregulated informational intermediation. Not the same thing.
I do believe that there is such a thing as technology: mastery of our physical, biological, and informational worlds. “Artificial intelligence” (whatever that may be) is a technology. So is a stone axe. Both can be used to improve lives, or hurt them. Both are closely entwined with their cultural milieus, but are conceptually separable from them. A stone axe is a stone axe, whether today or in the Paleolithic era.
Technologies are not in themselves business models, although they may enable them. Silicon Valley is about business models and the financing thereof, not their underlying technologies.
Seekonk 11.03.23 at 7:03 pm
“Silicon Valley is ideologically unwilling to help”
They’re also unworthy of their riches.
A “new” tech idea is almost always a relatively tiny accretion which embeds the cumulative scientific inquiry of the ages: Archimedes, Pythagoras, Descartes, Galileo, Newton, Curie, Faraday, etc., plus countless, nameless others. Fairness doesn’t require that Musk, Gates, Jobs, etc., be granted untold wealth for their “discoveries”; and if their idea is beneficial for general application, it’s anti-social to give them a monopoly on its use.
Cheez Whiz 11.03.23 at 8:15 pm
I wonder if anyone has addressed what the effect of constantly accelerating change might have on human society. Change is often confusing, and can easily move to threatening, even if just subliminally. Now add an ever increasing rate of change. What would a society where some substantial percentage felt subliminally threatened look like, and how might it behave?
both sides do it 11.03.23 at 8:30 pm
Couple-three things to keep in mind:
1) Productivity overall, YoY growth, is essentially linear since 1950. Moore’s Law (as discussed by DCA above) is about material constraints in computing, but how that translates into our economy or society is governed by entirely different dynamics.
Which leads to . . .
2) The pattern of Silicon Valley’s “growth” has been finding chokepoints in the economy and getting partial or total control over them. Facebook, Amazon, etc. There’s a similar dynamic to financialization of an economy here: bankers can invent esoteric financial vehicles and Line Go Up, but no value/growth in the underlying process of the productive economy is actually being created. Similarly, SV innovates technological / political / social tools (including code but also venture capital, cultural changes in the workplace, etc.) that are efficient at finding at niches in the economy to establish chokepoint holds over, and centralizing those profits to relatively few people with holds on the SV power structure.
Which leads to . . .
3) It’s not so much the total amount of money being centralized in a few hands, contra the inequality discussion above. It’s what that centralization does to the social and political fabric. Critical nodes in communication infrastructure are bought for chump change and become explicit tools advancing individual interests (Twitter, but also the Washington Post and, an earlier example, Fox News). Money from relatively few sources overwhelms other competing interests in the political system (the couple-three people bankrolling Clarence Thomas and relieving Brett Kavanaugh’s 6-figure debt which brought us women’s healthcare armageddon in a few short years, but also Sam Bankman-Fried shaping the Democratic Party’s data and consulting infrastructure for years on end). Etc., etc., etc. That is . . . bad. I’d pay more than $600 to not have it happen.
John Q 11.03.23 at 11:03 pm
Cheez Whiz @20 “I wonder if anyone has addressed what the effect of constantly accelerating change might have on human society”
This was the central theme of the 1973 hit book Future Shock, by Alvin Toffler, mentioned in the OP. I don’t think the thesis has stood up very well. At least in part that’s because of the fact that stupendous progress in ICT has been accompanied by near-stagnation elsewhere. My kitchen is pretty much the same as it was in 1973 (except for the microwave oven I bought a few years later). I have a 35-year old car which I can’t be bothered upgrading. Air travel is slower and less comfortable (though quite a bit cheaper) than it was then. OTOH, when I’m looking at the computer, I’m in an entirely new world.
c1ue 11.03.23 at 11:11 pm
@joeyjoejoe
Your analysis is highly flawed because it isn’t just the existing value of Musk’s (or any other billionaire’s) pile of money that is at issue – it is the future earnings and ongoing payments said pile of money generates.
$225 billion invested into say, 30 year Treasury notes would generate $11 billion a year, each year, for the next 30 years.
Both sides’ spending for the 2020 Presidential election was $1.85 billion, and all campaign spending was $14.4 billion. So the 100% safe interest earnings from Musk’s money would be only slightly less than every single campaign finance dollar spent in the 2020 elections.
The impact of SBF’s spending of $70 million in 2020 made him the 2nd largest single donor to the Democrat party, which puts the $11 billion into even greater perspective.
I believe this is much more indicative of the impact of large sums of money.
Alex SL 11.04.23 at 12:27 am
Further to my comment above that said, what they [tech bros] post on Twitter etc. proves they really aren’t very smart or well-informed, Eliezer Yudkowsky happened to tweet the following yesterday:
Among the dangers of AI is that LLMs dual-trained on code and biology could enable computer viruses to jump to DNA substrate. Imagine getting a cold that compromises your immune system and makes it start mining Bitcoin
It is impossible to fully convey how stupid this tweet is without resorting to very colourful language that I hesitate to post here. It betrays fractal ignorance across everything from computing to biology. Nobody who writes something like this should be taken seriously on anything, ever. If tomorrow this person claims the sky is blue, you’d better step outside to check, because they have proven to be so wrong that the null hypothesis should from now on always be that the opposite of what they say is true. If have dinner with them, you’d better manage health and safety risks by only giving them a plastic spoon, because that tweet demonstrated that they can never be trusted with something as complicated as a fork.
To be fair, he now claims it was a joke, but that would be more convincing if it that claim wasn’t the sixth tweet in the thread after his tweets 2-5 defended the first.
Now, admittedly, Yudkowsky himself isn’t strictly speaking a tech bro; the entire problem with him is that he knows nothing and understands nothing, and as far as I understand from reporting about him, he doesn’t have formal expertise in anything except writing fan fiction and collecting donations. But he is part of that culture. Many Silicon Valley tech bros venerate him as something between guru and cult leader and take him very seriously, and (this is the crucial point here) that is very revealing about them. What is more, although the words may be slightly rearranged from case to case, the entire discourse about AI risks being on the level of global heating are variations on the theme of Yudkowsky’s statement – whether offered by a Google manager or by a UK prime minister, none of them make any more sense than that. They all, in the end, boil down to, I don’t have to provide evidence or make sense, because the AI will be so smart it can do anything, including things we know are demonstrably impossible to do. It will just be that smart, believe me.
These people know how to order staff around to do things for them, are wealthy, and have maybe read a few Iain Banks novels or perhaps only absorbed that ideology by osmosis. And from that they conclude that they know everything anybody needs to know and are smarter than everybody else. Every time they say something, experts in the fields they opine on fall off their chairs laughing, yet they still get taken seriously.
Plarry 11.04.23 at 6:24 am
Trader Joe @ 14 hits the nail on the head. Where this piece misses is the failure to recognize that it is not Moore’s law that has fertilized Silicon Valley, it the fact that it is built and grows on a graveyard of startups and failures. What makes Silicon Valley successful and unique is that innovation is funded and tolerated, with the knowledge that most of the investment will fail. For every Facebook/Meta or Google, there are hundreds or thousands of companies that few ever heard of, there are companies that lived short and beautiful lives then died (Silicon Graphics, Sun Microsystems), and so on.
Peter T 11.04.23 at 9:42 am
On change, what counts is the social impact. The period roughly 1820-1880 was massively disruptive of social patterns that were literally millennia-old: speed and accessibility of transport (from weeks and months to hours and days), associated constraints on the size of cities and the ability to muster and sustain numbers – whether of workers or soldiers – in one place, on the centrality of ownership of land to political and social power. It’s no accident that the Communist Manifesto came out in 1848.
We are still working through the flow-ones in areas like gender and workplace relations and it’s these fuelling a lot of the current angst. In the longer view, SV is small beer.
fish bicycle 11.04.23 at 12:54 pm
the constant redefinition (towards meaninglessness) of what constitutes “innovation” has been key to sustaining this church.
there’s some level of self-awareness here too, in that “disruption is actually cool and good” is another rhetorical device employed by these particular thieves and mountebanks. this emerged once there were enough people starting to wonder about maybe sorta thinking about employing the precautionary principle for some of the value-free garbage coming out of silly valley. (maybe there is a timeline of the emergence of these twin rhetorical devices mappable to the WIRED magazine history?)
so, creation of better things (for who?) is good in and of itself, and so is the roiling of existing networks and mechanisms that actual human beings rely on for their day to day lives.
thats a mighty big space you’ve carved out there for financial speculation.
bekabot 11.04.23 at 2:08 pm
“What social problems are being caused by my lack of $588?”
You need dental work and can’t pay for it. In order to get your teeth fixed, you switch from regular Uncle-Sam Medicare, a government program, to Medicare Advantage, a business scheme. Medicare Advantage programs feature dental services and regular old Medicare doesn’t. So you get the problem dealt with, the sun shines brightly, and everything is fine. Then you begin to age and start and breaking down. You experience some kind of bodily malfunction, and in order to recover from it fully you need therapy. Your doctor recommends you to a therapy center and the center is willing to take you. Then you check with your insurance guys and find out they won’t come up with the money to pay the bills. (You, who originally were short of the few hundred bucks you needed to fix your teeth, of course can’t pay them on your own.) You and your doctor and the people at the center and your kids and your spouse and your friends and relations all remonstrate with your insurance firm, but remonstrations are in vain. They’re not paying, they’re not obligated, and that’s the end of it. The outcome is that you either go without therapy or get a few sessions, which are just enough to give you the energy to be really resentful. At home, you’re the focus of everyone’s attention, and not in a good way. Your spouse, like you, is getting older and requires a certain amount of physical aid, but can’t get it because you’re sicker than they are and your needs come first. Eventually they move in with a younger sibling who can take care of them, though the sibling has a spouse and kids and responsibilities of their own and it’s an awkward fit. One of your daughters quits her job in order to look after you, and the loss of income, and the drain on her concentration, has repercussions on her family. At length you die, maybe three or four (or five or six) years after your original breakdown, but ten or twelve years earlier than you would have died if you’d been allowed to regain your health. Your daughter is now not only minus a parent but minus a job. Her family is strained. Your spouse’s sibling’s family is strained. Nobody can quite afford to get their teeth cleaned. Rinse and repeat. In the meantime, politicians on Both Sides make inspiring speeches about the resilience of the American people.
All for lack of a nail, my friend.
Cheez Whiz 11.04.23 at 2:14 pm
John Q@20: Haven’t read Future Shock, but the summary on Wiki sounds remarkably shallow, even by 70’s standards. Just because a hack stumbled across a catchy phrase and beat it into submission doesn’t mean the idea is invalid. I think my point is that “entirely new world” seems to be leaking into our old one.
Greg Daneke 11.04.23 at 6:13 pm
Brilliant stuff, cheers. Your analysis past issues of Wired is very insightful. However, the more disturbing aspects of the Silicon Valley Fever goes well beyond product “enshittification”, it is the enshrinement of algorithmic vassalage and techno-feudalism. (Note: https://medium.com/@daneke/andreesen-and-the-techno-feudalist-apologia-dbe18ae28847).
banned commenter 11.04.23 at 10:03 pm
Musk doesn’t have billions of dollars in treasures and bank accounts. He owns businesses. He may very well be up the wazoo in debt, trying to maintain and expand his businesses. He has what Taleb calls “skin in the game”. He can’t just give 200 billion to pay for everyone’s teeth cleaning; that’s nonsense.
bad Jim 11.05.23 at 6:23 am
The power of Moore’s law is everywhere, but it’s easy to overlook. Phones, for example. I can remember my father struggling with the idea that the handset was not tethered to the base unit. Now it’s the rare household that even has a landline.
People still use phones for conversation, but especially in my beach town they’re as often employed as cameras, their gigabytes of cheap memory filled with selfies. I’m guessing that internet traffic runs mostly to streaming video.
It’s nice that washing machines are now controlled by chips instead of intricately stacked motor-driven switches, but too many things now have to go through lengthy start-up sequences instead of being instantly on. The text editor on my Mac is more sluggish than the one I had on my Osborne.
There’s sort of a Parkinson’s Law working in opposition to Moore’s Law, that increases in hardware performance will not linearly translate into software performance, the losses being in no small part due to what might be termed beautification.
Alan Peakall 11.05.23 at 10:10 am
Alex @ 24, Thanks for sharing that gem. I clearly recall that at the time of the Morris Worm one of those tasked with the post mortem grudgingly conceded that even the worst of the contemporary print media had not published anything quite as witless as the idea cross substrate infection.
engels 11.05.23 at 1:33 pm
What social problems are being caused by my lack of $588?
Elon Musk
bekabot 11.05.23 at 2:26 pm
“He can’t just give 200 billion to pay for everyone’s teeth cleaning”
I never suggested the Elon Musk could pay to have everyone’s teeth cleaned nor that he should. I was asked what a person could do with $588, and I replied.
Phil H 11.05.23 at 6:25 pm
I don’t think this piece makes a great argument. It suggests that Moore’s law is an ideology, and then argues against that ideology by pointing to some facts about how current technological progress is not as exciting as some commentators make it out to be.
This seems wrong to me, because Karpf is not following up on his original observation, that Moore’s law went from being a historical factoid to an ideology. Forget whether it works as a factoid. Karpf should follow up on the more interesting question: How does Moore’s law work as a belief system? I think he wants to suggest that it doesn’t work well, because Silicon Valley types seem very comfortable with extreme inequality.
But I want to suggest that compared to all the other ideologies out there, Moore’s law is fantastic. It suggests that the future will not be very much like the past, so all decisions we make today are contingent. That is, Moore’s law contradicts any totalising ideology, like a religion, and says, no, there are no eternal truths (possessed by your high adept bishops, or ours). Moore’s law undoes the primary pathology of ideologies, which is to insist that we are right, and so right that it’s worth restricting your freedom and killing you to enforce our rightness. The Moore’s law ideology doesn’t allow for that, only for buying you out, or for outcompeting you on speed.
Moore’s law as an ideology also tends to be inclusive and international, because those companies desperately want Asian, and ultimately African markets.
And it’s somewhat nonconformist, though it’s building its own conformities as it grows.
So if Karpf actually believes his premise, that Moore’s law has become an ideology, then he owes it to us to examine what that means. Vague complaints about inequality aren’t the meat of this argument.
Curtis Adams 11.05.23 at 10:26 pm
Moore’s law has been dying or dead for personal electronics for years. A Sony Vaio had a clock speed of 2.8 Ghz in 2004: in 2020 an HP Pavilion can do – 2.0 Ghz. The HP has multiple processors, to be fair; but nothing like the 2000 that would be required for Moore’s law to apply to to that many 18-month doubling periods. On-board memory kept on an exponential trajectory until the mid-teens, but has now greatly slowed; storage is falling off its exponential trajectory now. A 2023 computer isn’t even remotely close to 4 times as good as a 2020 computer in any aspect. Moore’s law has run its course and is over.
What’s made smartphones become so much more impressive is not the moderate hardware improvements but the still-exponential growth in wireless transmission speeds. That now is hitting a different limit; there really isn’t much use for anything faster than streaming a movie, and we’re there now. So even if it could continue it won’t make much difference.
Alex SL 11.05.23 at 11:05 pm
Phil H,
A full evaluation of the tech bro ideology of exponential progress bleeding into singularitarianism would likely require a lengthy book, a PhD thesis or suchlike. As somebody who doesn’t study ideologies for a living but merely watches from the side-lines, I would, however, say that the two beliefs outlined below and their logically following conclusions are at the core of the ideology, in as short a comment as I can make it. (Sorry.)
Belief the first, and the core of the original post: Progress is accelerating exponentially, and if we just let it accelerate (as opposed to constrain it with regulations), it will soon get to the point where humanity sees unprecedented prosperity and is granted technological solutions for every social and environmental problem there is. There will be immortality, interstellar space flight and colonisation first of Mars, then of planets orbiting distant stars, mind uploading, unlimited energy, unlimited mineral resources, and a unicorn for everybody, because resource constraints and the laws of physics just slink away if a smart enough mind turns its attention to them and isn’t hampered by annoying government busybodies.
Two conclusions follow from this belief. A, there should be no regulations to protect people from exploitation, fraud, misuse of their data, theft of their intellectual property, etc., because those just get in the way of progress and the required risk-taking. The consequences of this conclusion are obvious: vulnerable/poor people will suffer exploitation, fraud, misuse of their data, theft of their intellectual property, etc. B, because there will be a technological utopia soon that solves all our problems, we don’t actually need to do anything as tedious as reduce carbon emissions, protect biodiversity, keep our water supply clean and plentiful, preserve arable soils from erosion, or pay taxes to maintain infrastructure or public services. Whatever problem arises, there will be a tech fix delivered by Genius Entrepreneurs and/or the Super-AGI they have created, and if all else fails, we just move to Mars. The consequence is that if this belief wins out, our current global civilisation will go the way of the ancient Indus Valley Civilisation, the Minoans, or the Sassanian Empire, because it fails to act on threats to its very existence, with billions of deaths during the collapse. (And Mars happens to be much, much less habitable than Antarctica was during the last ice age, quite apart from the logistical issues of getting there.)
An important accessory argument to this first belief is that every counter-argument based on physical reality can be rejected by claiming that the interlocutor is just not imaginative enough and can’t predict what will yet be possible. In other words, “five hundred years ago you would have said that airplanes are impossible, so who are you to say now that immortality and mind uploading are impossible?” With that logic, any potential future tech achievement can be envisioned, no matter how disconnected from reality.
Belief the second: Value is created not by people working hard to create it, through incremental social processes such as academic research or iterative improvements in engineering, or supported by public investment, but entirely through an Entrepreneurial Genius having a Great Idea all by themselves and Taking Risks to see it through. The hundreds or thousands of people who annotated the training data for the AI, who coded the software and keep it running, who did the detailed design plans and painstaking prototype testing, who build the devices in a factory, who actually had the idea that the investor claims as his own after he bought them out, etc, are nothing but a weird background fog that can be ignored. No, the only person who ever creates value is the Entrepreneurial Genius who stands on the stage and unveils the new product to adoring tech journalists. The consequence of this belief is a general hostility to, again, regulations that keep consumers and the broader public safe, but also specifically to organised labour and labour protections, resulting in increased precarity and exploitation.
Plarry 11.05.23 at 11:32 pm
Curtis Adams @ 37: Moore’s Law says that the density of transistors in chips will double every 18-24 months (working out to about an order of magnitude per decade). It makes no claims about CPU clock speeds. We could debate about whether Moore’s law still holds, but in 2003, the AMD Athlon 64 was introduced with a transistor size of 130 nm, and in 2023 the Apple M3 chip was built using a 3nm technology. So not quite two orders of magnitude/20 years but not too bad.
The reason that processor speeds are more or less the same has to do with diminishing returns of higher clock speeds caused by bottlenecks in the CPU architecture, heat dissipation, and the power consumption caused by higher clock speeds. Architecturally, it has made much more sense for the industry to “go wide”, i.e., increase parallelism, rather than try to boost speed. Digging into it you will find that GPUs, which have massive parallelism and are responsible for much of the AI computation going on today, actually have slower clock speeds than the CPUs.
clew 11.06.23 at 2:24 am
compared to all the other ideologies out there, Moore’s law is fantastic. It suggests that the future will not be very much like the past, so all decisions we make today are contingent
Changing the sign on the discount rate seems like it should be interesting. In practice “your regulations now are bad because they $STEP2 torture billions of future people” comes out on the same side as all the corrupt bishops of the past, so it doesn’t seem to make any difference.
clew 11.06.23 at 2:27 am
there really isn’t much use for anything faster than streaming a movie, and we’re there now. So even if it could continue it won’t make much difference
Oh, maybe I can think of one — public key signing of realtime video transmissions by several people, so the difficulty of faking recordings doesn’t drop to zero? Totally paperback mirrorglasses handwaving here, though.
both sides do it 11.06.23 at 6:20 am
“there really isn’t much use for anything faster than streaming a movie”
rich complex AR, 3D projections in space, etc; that hardware’s not there yet tho, incentive might not exist to get there
roger gathmann 11.07.23 at 9:38 am
Joe. What social problem would be caused by taking all of Elon Musk’s money save a million dollars? I don’t see the problem. Bright side: more people would have more cultural, political and social power, once the billionaire class is appropriately shrunk. Down side: Elon Musk’s feelies would be hurt.
joejoe 11.07.23 at 2:24 pm
“What social problem would be caused by taking all of Elon Musk’s money save a million dollars?”
The establishment of a government that takes stuff from its citizens because other citizens are jealous.
Joe
Kenny Easwaran 11.08.23 at 11:10 pm
It’s true that the pace of change of the web, and of the hardware capacities of consumer electronics, has been much slower in the past 10 years than in the period from 1980-2010 or so. But it’s also true that the amount of influence these things have on the life of ordinary people has changed much more drastically in these past 10 years than in that previous period. Back in 2010 we were still waiting on the dream of access to the music catalogs of all the major record companies, and you had to limit your local travel in case you accidentally stayed out past the last bus or train. But Spotify and Lyft really have, for better or worse, significantly changed how many ordinary people structure their days. All of the many benefits and harms of social media also date to this post-2010 period – whereas in 2010, the average person wasn’t really that much more impacted by web media than they were in 1980, before the web even existed.
Tm 11.09.23 at 5:29 pm
Paul Krugman used to point out that indoor plumbing and electrification had more impact on daily life of most people than the computer or the internet.
Otoh, many people nowadays spend most of their waking time in and around the internet. That for sure is a huge impact on daily life, just mot necessarily an improvement.
engels 11.09.23 at 6:20 pm
Back in 2010 we were still waiting on the dream of access to the music catalogs of all the major record companies
Unless we had already got everything we could possibly want to listen to (and hundreds of GBs we couldn’t) from Bit Torrent etc years ago (not that I would condone such behaviour).
Tm 11.10.23 at 4:15 pm
BTW:
“On a zoomed-out time scale, technological progress follows an exponential curve. Compare how the world looked 15 years ago (no smartphones, really), 150 years ago (no combustion engine, no home electricity), 1,500 years ago (no industrial machines), and 15,000 years ago (no agriculture).”
This time series is not exponential but hyperbolic. If it were real, the next breakthrough would occur in 1.5 years, the next in .15 years and so on. There‘d have to be an infinite number of further breakthroughs around the year 2025. Maybe that is what they call the singularity. A hyperbolic series ends in a vertical asymptote.
engels 11.12.23 at 1:18 am
Compare how the world looked 15 years ago (no smartphones, really)
The horror! The horror!
roger gathmann 11.12.23 at 10:56 am
Ah, intangible negative effeccts! Why it is almost as if you don’t believe your quantitative point! Which makes one ask – why pose it?
As for the government “taking” Elon’s gains, which came from him “taking” Tesla’s collective net profits for a decade and puttting them in his pockers – well, too bad. If the state wants to take my billion and pare me back to millionaire status, the qualitative benefits far outweigh the loss to a Randian philosophy that thinks taxation is theft, but profittaking is justice.
I want my $588 dollars!
J-D 11.12.23 at 10:20 pm
But that’s not a problem!
(I don’t think it’s ever going to happen, but if it did, it would not be a bad thing. If you think it would be a bad thing, I suggest you haven’t thought about it enough. If you think I haven’t thought about it enough, tell me what I’ve missed.)
Andrew 11.14.23 at 12:12 am
As someone who lives/works in Silicon Valley, in semiconductors, I would clarify the above as applying more to venture capitalists and those who are in software startups? People in semiconductors and other high tech hardware (datacenters, basebands, etc.) are dealing with limitations of laws of physics, and so the rhetoric above doesn’t come as easily to us bc we are bashing our heads (i.e. spending tons of time/$$$ to overcome) against the physical limits of transistor size reduction and the difficulties of getting to 6G. People who work at AWS or adjacent to it, for example, aren’t going to sound anything like this – even though AWS has done so much to change our lived realities with the expansion of the cloud. But AWS has to sit and wait for building permits to come through so a datacenter can get built. It’s very telling that Horowitz and his friends aren’t involved in speculating on the technological frontier much at all (look at a16z investments – mostly software services and NFT-stuff).
A lot of ex-Wall Street types migrated to tech/Silicon Valley following the 2008 crash and became “business development” professionals (M&A for corporates) or product managers for software/internet services. The fact that the move into actual operations was in product management in software/internet is very telling, as there’s a delusion that in the world of software/internet, there are no limits on what can be done? A few lines of code and voila! I think that this naive view of the power of software (even though, of course, software well done can be very powerful indeed) mirrors a view of the economy as frictionless/seamless due to the artificial environment of trading through Bloomberg terminal (remember the story of oil traders who found out one day that the oil futures they had been trading one day meant that they were obligated to physically take receipt of barrels/tanks of oil, and were scrambling to find storage or to offload the futures contracts which had led to them having said obligation)
So that makes potentially TWO different (though overlapping) Silicon Vallies (hardware vs software, mature firm vs venture-backed, or aspiring-to-be-VC-backed startups). Then, in the East Bay Area there are also large biotech and pharmaceutical campuses which house massive R&D teams (as well as the IP lawyers who enforce royalty/trademark/IP rights). There, three Silicon Valleys.
All this being said, I’m a longtime CT reader and in a past life studied political economy under Fred Block and then had a grad advisor who himself studied under Michael Mann, so I’m not the typical tech bro…but I think my point about there being “many different Silicon Valleys” still stands.
David in Tokyo 11.14.23 at 6:47 am
Plarry wrote:
“We could debate about whether Moore’s law still holds, but in 2003, the AMD Athlon 64 was introduced with a transistor size of 130 nm, and in 2023 the Apple M3 chip was built using a 3nm technology. So not quite two orders of magnitude/20 years but not too bad.”
Careful there. The definition of “nm” has changed. I forget the exact details, but the “improvements” since the 24 or so nm level of things have been not in the actual size of things, but in the fine details of the fabrication technology, and the “nm” ratings are now just names of the generation of the technology not the actual feature size. At the start of this changeover, the game was that the “nm name” reflected the move to devices having a bit of a 3D structure on the chip, so it was an “equivalent nm” thing where you actually did get proportionally more devices per sq. mm, but after that it was just a generational thing. The term “nm” isn’t “officially” used, and hasn’t been for a while.
Not that things aren’t getting better, but the improvements are way slower than the used to be in the period 1985 to 2005 or so, over which microprocessors when from running some number of millions of instruction per second to some number of billions of instruction per second. Three f-ing orders of magnitude in 20 years. We get more processors and more memory than we did in 2005, but the best individual cores are “merely” a tad over an order of magnitude faster than they were almost 20 years ago.
On the other hand, the improvements have been way more interesting. The games that Intel and Apple play to get more instructions executed per clock cycle are seriously amazing. But that has introduced a problem that the more games of that sort you play, the more power these things draw. So you get another zillion gates, the IPC goes up 30% or so every 3 years, but the power per instruction executed goes through the roof.
So now both Intel and Apple give you both “efficiency cores” and “performance cores” because making things faster uses too much power, and you can’t actually use those performance cores except for very brief periods if you are a notebook.
Comments on this entry are closed.