Welcome to "Modernity": A Comment on Living in Exponential Time; & BRIEFLY NOTED: For 2023-03-18 Sa
Jason Kottke & Ezra Klein & MarshallBermanChatBot on how to deal wiþ all that is solid melting into air; cost-disease & cost-shifting rule relative price changes; Davies and Levine snark about Sil...
Jason Kottke & Ezra Klein & MarshallBermanChatBot on how to deal with all that is solid melting into air; cost-disease & cost-shifting rule relative price changes; Davies and Levine snark about Silicon Valley Bank, Rubenstein attempts to explain what went wrong at SVB, and Barro says Ron DeSantis is not soo savvy…
FOCUS: Welcome to “Modernity”!:
Our texts this morning are about “living in exponential time”. The first is from Jason Kottke, commenting on Ezra Klein:
Jason Kottke: The Difficulty of Living in Exponential Time: ‘The pace of improvement in the current crop of AI products is vastly outstripping the ability of society to react/respond to it, Ezra Klein… [says] “the difficulty of living in exponential time”…. Covid, AI, and even climate change (e.g. the effects we are seeing after 250 years of escalating carbon emissions)… they are all moving too fast for society to make complete sense of them. And it’s causing problems and creating opportunities for schemers, connivers, and confidence tricksters to wreak havoc…
And the second is from Ezra Klein himself. Let me give the mic to him:
Ezra Klein: This Changes Everything: ‘Sundar Pichai… “A.I. is probably the most important thing humanity has ever worked on… more profound than electricity or fire.”… The possibility that he’s right…. I am describing not the soul of intelligence, but the texture of a world populated by ChatGPT-like programs that feel to us as though they were intelligent…. What is hardest to appreciate in A.I. is the improvement curve…. The people working on A.I…. are creating a power that they do not understand at a pace they often cannot believe…. A tempting thought…. These people are nuts…. This was true among cryptocurrency enthusiasts…. Is A.I. just taking crypto’s place as a money suck for investors and a time suck for idealists and a magnet for hype-men and a hotbed for scams?
I don’t think so…. In 2021… DeepMind managed to predict the 3-D structure of tens of thousands of proteins…. Science named it their breakthrough of the year…. Populat[ing] our world with nonhuman companions and personalities that become our friends and our enemies and our assistants and our gurus and perhaps even our lovers?… Put millions out of work?… Help terrorists or antagonistic states develop lethal weapons and crippling cyberattacks?… End up controlling critical social processes or public infrastructure in ways we don’t understand and may not like?…
We can plan for what we can predict…. What’s coming will be weirder…. We do not understand these systems, and it’s not clear we even can…. Zoom into specifics and the picture dissolves into computational static…. The “thinking,” for lack of a better word, is utterly inhuman, but we have trained it to present as deeply human. And the more inhuman the systems get… the more human they seem to us…. If we had eons to adjust, perhaps we could do so cleanly. But we do not…. To suggest we go slower, or even stop entirely, has come to seem childish…. Fatalism becomes the handmaiden of inevitability, and inevitability becomes the justification for acceleration…
I read these, and I think: this is how it has been since the 1870s. Perhaps this is how it has been since the late 1830s. And we see anticipations of this long before; for example, we see such antciipations in Tommaso de Campanella’s City of the Sun and in Francis Bacon’s New Atlantis, both written around 1600, both marvelling at the extraordinary transformations of society triggered by new technology, and both looking forward to much more. One of the people who most incisively made this point was the late cultural critic Marshall Berman, who died in 2013.
I cannot do a Marshall McLuhan, and pull him out from behind the curtain and ask him about “modernity”. But I have prompted OpenAI’s davinci-003 GPT LLM with relevant excerpts from the text of his big book, and so I can ask my cut-rate MarshallBermanChatBot questions:
Why did you entitle your big book All That Is Solid Melts into Air?
MarshallBermanChatBot: The title of my book, All That Is Solid Melts Into Air, is drawn from the opening line of Karl Marx's work Capital, Volume One, which states: "All that is solid melts into air, all that is holy is profaned, and man is at last compelled to face with sober senses, his real conditions of life, and his relations with his kind." This phrase encapsulates the idea of modernity that runs throughout the book, which is the idea that the world is constantly changing and the only thing one can count on is the experience of modernity itself. This idea can be seen in the examples provided in the context, such as the ad for Mobil Oil, the story about Shearson Chief, and the quotation from Dostoevsky. The title of the book thus reflects the idea that in a world of constant flux and transformation, we must embrace modernity and make the most of the opportunities it provides us…
Now the quote—“all that is solid melts into air, all that is holy is profaned, and man is at last compelled to face with sober senses, his real conditions of life, and his relations with his kind” is not the opening line of the 1867 Capital, but comes, rather, from the 1848 Communist Manifesto.
But let us continue:
What do you say to those who say that the present in the future tech booms are actually much more important and bring much faster and more disruptive change than did industrialization?
MarshallBermanChatBot: Well, it's true that modern technology has enabled us to make huge leaps forward in terms of speed and efficiency. We can move information around the world in the blink of an eye, build complex machines and systems, and enhance our own creativity and productivity in ways that were unimaginable even a few decades ago.
However, it's also important to recognize that many of the changes we’re seeing today were anticipated and prefigured by the industrial revolution. For example, the idea of machines taking over human labor is actually quite old—it's been a recurring theme in the history of technology since the industrial revolution. The same can be said of the automation of repetitive tasks, the development of data-driven decision-making, and the rise of the ‘knowledge economy’. So, while modern technology has certainly accelerated the pace of change, it's important to recognize that many of the ideas and concepts that we’re seeing today were actually anticipated by the industrial revolution…
So what you are saying is that the features of “modernity” that so distress Ezra Klein have actually been the rule for the world since 1870, or before?
MarshallBermanChatBot: Yes, that's exactly right. The industrial revolution was a pivotal moment in human history, ushering in a new era of mass production, mechanization, and economic growth. Many of the features of modernity that we're seeing today—such as automation, data-driven decision-making, and the rise of the knowledge economy—actually have their roots in the industrial revolution. For example, the idea of machines taking over human labor has been around since the industrial revolution, and the automation of repetitive tasks was also a key development of that period. So, while modern technology has certainly accelerated the pace of change, it's important to recognize the role that the industrial revolution played in laying the foundation for the modern world…
There are huge bugs in these experimental systems that so awe and shock Ezra Klein and Jason Kottke. As you see, it has already gotten very repetitive in its answers. While we appear at first glance to be having a conversation—engaging with a Turing-class entity constructed to mimic the kind of dialogue that you would have with the late cultural critic Marshall Berman—that is not in fact the case. The “entity” we are conversing with is very sub-Turing indeed. It is page-level autocomplete.
Now page-level auto complete is a wonderful thing. Asking a computer and getting an answer to the question: Out of the pages that you have been trained on, which are the “nearest neighbors” to this page I am feeding you in the prompt—my question, and a passage or passages from Marshall Berman that match key words in my question—and what is the average of responses to those nearest-neighbor pages? That can be very useful indeed. I look forward to a quadrupling of certain kinds of white-collar productivity. It seems to me likely to be of the same order of magnitude as the coming of the personal computer, or the coming of the internet. And, like those two, it seems of especial revolutionary importance to writers because it, like the coming of the personal computer or the coming of the internet, is an expl osion and transformation for which they are at the epicenter.
I do anticipate the use of this page-level autocomplete will spread around the worl, considerably faster than use of personal computers or use of the Internet. We are primed to have conversations. And what Chat-GPT or Bing-AI or MarshallBermanChatBot offers us is very close to a real human conversation, of the kind we expect to have with a Turing-class entity.
A kind of animal that attributes Turing-class intelligence to the lightning—which we, at least some of us, personified as a huge guy with a red beard and severe anger management problems—is primed to make much use of page-level autocomplete, when it is presented to us in this way. That was the lesson of the move from GPT3 to Chat-GPT3.5. Duck. Water.
Will we use it for good or ill? That is in our hands.
Will we understand how this changing forces-of-production, forces-of-communication, forces-of-information technology-driven hardware requires that we, in response and on the fly, change and cobble-together new and different relations-of-production, and then on top of that superstructural cultural-political-social, software of human interaction that runs on top of the hardware and makes up our society?
No.
But we have not understood that since 1870.
And that process of not understanding in time is a huge part of the subject matter of my book Slouching Towards Utopia <bit.ly/3pP3Krk>.
How can and should humanity deal with the extraordinary pace of change, upset, and transformation that is both the blessing and the curse of modernity?
MarshallBermanChatBot: The extraordinary pace of change, upset, and transformation that is both the blessing and the curse of modernity is a reality of modern life that humanity must face with courage and creativity. We must recognize that the old solutions and grand narratives that have guided us in the past are no longer as relevant or viable as they once were, and that in order to thrive in this new environment we must look for new solutions and develop new ways of understanding. We must also accept that our ability to make meaningful connections with each other is a crucial part of our collective response to this challenge. We must make the most of the opportunities modern life offers us to communicate and understand one another, and prioritize meaningful dialogue, collaboration and collective action. Finally, we must continue to search for ways to make modern life worth living and strive to create a world in which the free development of each person is an essential part of the free development of all…
Reciprocity, redistribution, hierarchy, and democracy. Those were the modes of human interaction back in the long Agrarian Age ever since agriculture and herding finished their spread across Eurasia and much of Africa in the year -6000. To those the Modern Age has added the market economy as a qualitatively different thing than reciprocity (although closely connected) and bureaucracy as a qualitatively different thing than hierarchy (although closely connected), giving us six modes of organization. Now I think it is clear that we are adding a seventh mode: organization-by-algorithm.
And it is going to be a wild ride.
ONE IMAGE: Smart Mike Konczal Is Smart:
Not so much regulation as Baumol’s cost disease and institutional cost-shifting lie at the heart of the relative price increases we have seen.
MUST-READ:
Daniel Davies snarks about Silicon Valley’s loss of assibayah, and Matt Levine picks it up:
Matt Levine: ‘Silicon Valley Bank failed last week because, to oversimplify slightly, its interconnected network of tech-startup and venture-capitalist depositors all worked themselves into a panic and rushed to withdraw money all at once, then tweeted indignantly about how they were justified in doing that and how it was all the Fed’s fault. It was individually rational for each depositor to take its money out and avoid exposure to SVB, but the collective result was quite bad for SVB and for the banking system and for the VCs and startups themselves. Silicon Valley Bank’s Silicon Valley customers, it turned out, were individually rational but unable to act cooperatively in a mutually beneficial way; in the prisoners’ dilemma of a bank run, they all chose to defect. Banking analyst Dan Davies was scathing on Twitter, writing about the First Republic rescue: “When you tell 11 bankers that someone is in trouble, seems that they ask ‘how can we help’. Just don't try that with effective altruists eh”...
Very Briefly Noted:
Mike Konczal: A Better AEI Graphic of Relative Prices A Better AEI Graphic of Inflation Over the Past 20 Years…
Nat Bullard: Presentations: ‘Decarbonization: The long view, trends and transience, net zero: My annual effort to capture the state of climate and decarbonization: how we got to now, where we are going, what's new, and how to think about what's next…
Robin Wigglesworth: How crazy was Silicon Valley Bank’s zero-hedge strategy?: ‘Not as nuts as you might think, but pretty nuts…
Katie Martin: The tumult in Treasuries: are hedge funds partly to blame?: ‘Historic rally in world’s most liquid market may have reflected a ‘short squeeze’…
Kenneth Rogoff: Things are only getting harder for the Fed: ‘We should be thankful that the past two weeks of banking stress did not happen sooner…
Eric Foner: A Regional Reign of Terror: ‘Most Americans now grasp that violence was essential to the functioning of slavery, but a new book excavates the lesser known brutality of everyday Black life in the Jim Crow South.... Margaret A. Burnham’s new book, By Hands Now Known: Jim Crow’s Legal Executioners, a work by turns shocking, moving, and thought-provoking...
¶s:
Dan Davies: Regulation is the scientist, supervision is the monster: ‘The USA didn’t do a good job of implementing these regulations.... But I just said that these rules only codified good practice! That means that the responsibility for the failure lies on the management.... And it lies with the supervisors at the San Francisco Fed, who ought to have stepped in as soon as they saw what was going on. That’s the gist of the argument. And I … agree? There certainly needs to be a number of interviews without biscuits – this thing had unrealised losses bigger than its shareholders’ funds, on a portfolio of 30 year bonds financed with overnight money. That’s many levels of “come on, you can’t be doing that”. But.… “No, your treasury practices are insane and dangerous” is a difficult technical argument which might have to be extracted from megabytes of inconsistent reports. “If it’s so great, why can it not produce this simple ratio, compliance with which is required by law” is a much easier conversation. I said in a previous mail that every accounting system is a prison, but an accounting system can also be a battleship. Supervision and regulation are like (because they’re specific cases of the general concepts of) action and information. They’re mainly separate things, but the ways in which the distinction breaks down might be more interesting than the distinction itself...
Josh Barro: Is Ron DeSantis Savvy Or Not?: ‘The president has been smartly preparing himself to face a Republican opponent on these issues by breaking from his party in popular ways — blocking reforms to weaken criminal prosecution in Washington DC, making it more difficult to cross the southern border and seek asylum, and approving more domestic oil drilling. The other thing Biden will need to do is to move to define DeSantis on his weak issues before DeSantis has a chance to. I wrote last month about how Biden is doing this with DeSantis’ congressional record of voting for Social Security and Medicare cuts. If and when DeSantis signs strict abortion restrictions into law in Florida, I expect Biden to take similar advantage on that issue.... The 2024 election won't be about which books are in elementary school libraries. It won’t be about Drag Queen Story Hour. And it won’t be about the ethics of flying migrants to Martha’s Vineyard. Ed Kilgore had a somewhat odd column last week calling it “risky” for Democrats to try to sidestep these culture-war issues. But what Biden is doing is simply keeping his eye on the ball — focusing on the things that DeSantis would do differently (and harmfully) in the actual operation of the federal government that would affect most Americans’ daily lives. It’s a favorable contrast for Biden, if he can keep the focus there…
Marc Rubenstein: Contagion: ‘Silicon Valley Bank’s assets were largely money-good. It sold a $21.4 billion portfolio of securities to Goldman Sachs days before its collapse but in the markets where they trade, that wasn’t enough to move the price. Indeed, Goldman may even have made a $100 million profit on the trade. Its loans were also pretty good. Non-performing loans made up just a fraction of its loan book and they were covered more than four times over by provisions. Most of its loans reprice over the short term so weren’t underwater from higher interest rates in the same way the securities book was. Ironically, if Silicon Valley Bank had written bad loans, it would have been given more time to work them out. In the last crisis, bank failures peaked in 2010, a full year after bank stock prices bottomed. Rather, contagion is spreading through an altogether different channel: customer confidence. The speed with which Silicon Valley Bank collapsed has customers questioning the fragility of the very model banks operate under…
Page level autocomplete seems to work as well as short phrase autocomplete. It's accurate maybe 10%-15% of the time. The AI/ML systems that are useful are the product of extensive and focused data gathering and do an advanced form of compression or curve fitting. If you want to optimize a metal alloy or find a protein structure, they can work fairly well, but that's because of the clean, standardized data set.
A friend of mine has been exploring various AI chat systems' literary abilities, answering questions about English language literature. She says the completion systems have access to interesting things, but you have to know those interesting things in order to get their completions. It is a chicken and egg problem that leaves one stuck with a chicken or an egg and no way to get from one to the other.
It's a lot like internet search. If you know the right terminology and phrasing, you can - or at least could - find all sorts of interesting things, but if you don't, you'll have to keep trying alternative search phrasings and hope that hit #267 provides a useful key phrase that leads one to a more interesting place. Search systems have actually gotten worse in this regard since they have done a lot of work to "improve" search over the last decade. It is often impossible to find certain things without precise phrasing and priming.
What happened to search? It moved away from simple textual search and tried to understand the semantics. Suddenly, things that were accessible vanished, since their web pages didn't fall into the correct place in semantic vector space. It's much easier to figure out an appropriate textual search than trying to outsmart a relatively opaque semantic space algorithm. Conversational search seems to have the same problem. At least one can look at the URL and original text of a web page to determine if it is bogus rather than having a credulous semantic algorithm accept and incorporate it. (Maybe these things should provide footnotes.)
To deal with this, most search engines now have a list of approved web sites that they use to resolve certain search queries. The same sites, often encyclopedias or standard reference works, appear again and again at the top of one's search results. The reason is obvious, the internet is full of incorrect and poorly indexed information. If the answer can be found in Encyclopedia Wikipedia, it is at least likely to be useful. On the other hand, there is a reason one learns to move beyond encyclopedias and such secondary sources if one is going to find out things.
RE: cut-rate MarshallBermanChatBot - like BradBot, an excellent demonstration & application of LLM