“In the future, everyone will be famous to fifteen people.”

People seemed to like that when I wrote it in  my 2002 book, Small Pieces Loosely Joined.

About ten  years later, someone (sorry, I don’t remember who) pointed out that that phrase had in fact been first written by the Scottish singer and writer Momus [wikipedia], in 1991. Momus’ article that the phrase captions is especially insightful since it was posted before the Web created a plausible way to route around mass markets.

The page that currently houses Momus’ original article traces the history of Momus’ article:

This essay, written by Momus in 1991, was published by the Swedish fanzine Grimsby Fishmarket in 1992 then in the daily paper Svenske Dagblatt in 1994. Obviously a hardy perennial, it was then broadly paraphrased by trendy French magazine Citizen K in 1996)

I wasn’t aware of any of these and had never heard the phrase before I came up with it independently. When Momus’ prior use of the phrase was brought to my attention maybe ten years later, I blogged about it. But now I can’t find that post. So, since the origins of the phrase came up again today in a post by Dennis Falvy, I thought I should refresh the Web with this acknowledgement.

And if you want to know more about Warhol’s original saying, The Quote Investigator is, as always, worth reading.

Be the first to comment »

The article notes that the prize actually went to two humans, but this headline from MIT Tech Review may just be ahead of its time. Are we one generation of tech away from a Nobel Prize going to a machine itself — assuming the next gen is more autonomous in terms of what it applies itself to?

Be the first to comment »

If you ask chatGPT to give you the first paragraph of Moby-Dick, it gets it exactly right. This and other such examples are sometimes taken as evidence that it has kept a copy of that book in memory, a potential copyright violation. The makers of these Large Language Models say they don’t, or at least I haven’t come across any who cop to it.

My understanding is that they analyze the terxts they’re trained on looking for statistical relationships, keep the stats and ditch the texts. (In this, they’re similar to how search engines compile their indexes.)

But my understanding is quite limited. So, I poked around.

PROMPT: What is the first paragraph of Moby-Dick, verbatim?

CHATGPT:

Call me Ishmael. Some years ago—never mind how long precisely—having little or no money in my purse, and nothing particular to interest me on shore, I thought I would sail about a little and see the watery part of the world. It is a way I have of driving off the spleen and regulating the circulation. Whenever I find myself growing grim about the mouth; whenever it is a damp, drizzly November in my soul; whenever I find myself involuntarily pausing before coffin warehouses, and bringing up the rear of every funeral I meet; and especially whenever my hypos get such an upper hand of me, that it requires a strong moral principle to prevent me from deliberately stepping into the street, and methodically knocking people’s hats off—then, I account it high time to get to sea as soon as I can

Perfect.

PROMPT: What is the first paragraph of David Weinberger’s book Everything Is Miscellaneous?

CHATGPT:

The first paragraph of Everything is Miscellaneous by David Weinberger reads:

“If you keep your music in iTunes or your photos in Flickr, you’re using an application that lets you put a piece of information in more than one category at a time. That’s weird. In the real world—where our species evolved—everything has its place. Our books are on one shelf at a time, our plates are in one cupboard, and our shoes are by the door or in the closet. That’s the way it’s supposed to be because that’s the way the world is. Except, of course, it’s not.”

This opening sets the stage for Weinberger’s exploration of how digital organization challenges traditional, physical-world categorization.

That’s not a bad intro to the book’s themes:  it picks up some of the examples in the book, and it is certainly reminiscent of my style of writing. But it has nothing to do with the first paragraph of the first chapter:

Before the Web, browsing was usually a polite way of telling a salesperson to buzz off. “May I help you?”, a salesperson asks. “I’m just browsing,” you reply with a little smile.With that word, a customer declares a lack of commitment, With that smile, she asserts that she’s within her rights: “Just try and stop me, salesboy!”

I’m slightly embarrassed by that, but the point is that the real first paragraph doesn’t talk about anything mentioned in chatGPT’s attempt to quote it.

Further, while chatGPT can quote the opening of Alice’s Adventures in Wonderland, if you ask it to quote the opening of Carroll’s lesser known “Sylvie and Bruno” it responds:

One winter night, at a little town called Elveston, a telegram was received by the Sub-Warden, who sat sipping his coffee, and wondering lazily whether it was worth while to finish his cigar, or whether he would take a stroll down to the Club, and see if Captain Muriel had returned yet.

Here’s the actual opening:

—and then all the people cheered again, and one man, who was more excited than the rest, flung his hat high into the air, and shouted (as well as I could make out) “Who roar for the Sub-Warden?” Everybody roared, but whether it was for the Sub-Warden, or not, did not clearly appear: some were shouting “Bread!” and some “Taxes!”, but no one seemed to know what it was they really wanted.

The phrase “one winter” doesn’t appear anywhere in the original. Nor does “a little town called Elveston”. 

Finally, if you repeat the prompt, it gives  substantially different responses, including the unfortunately smutty these days:

‘Less bread!’ cried the poor. I looked again, and saw it was not bread they had needed, but more head.

It seems to unqualified me that this bolsters the claims of the LLM makers that they don’t keep copies of their training materials (or at least don’t make them accessible to the model in operation), and that the training-time statistical analysis of texts that quote other texts, plus the magic of transformer technology,  is sufficient to explain how they can provide verbatim quotes of oft-cited works. 

Am I wrong? It’s highly likely.

1 Comment »

Traditionally in the West we define knowledge as a justified true belief. But the experience of knowledge usually also requires understanding and a framework of connected pieces of knowledge.

Guess what machine learning lacks: understandability and a framework from which its statements of knowledge spring.

We might want to say that therefore ML doesn’t produce knowledge. But I think it’s going to go the other way as AI becomes more and more integral to our lives. AI is likely to change our idea of what it means to know something…

Continued at the Peter Drucker Forum

Be the first to comment »

So, you’re an academic, a researcher, a scientist … a serious person. But you’ve decided to write a trade book — a book for the general populous — to amplify the effect of your work. Excellent!

I edit the Strong Ideas book series for the MIT Press — books about the ways in which today’s digital technology is or may be changing how we think and behave. The readers are assumed to have no, or little, technical background, but the authors are generally academics and scholars.

This is the advice I give them at the beginning if they have not written for laypeople before.

The differences between academic and trade books

1. A trade book is for people who know much less than you do about your topic.
2. A trade book is for people who are much less interested in your topic than you are.

These two points are inseparable when it comes to structuring and writing your book. When writing for academics, you’re dealing with people who come to your book already interested in your topic and who think your book has something worthwhile to say to them…or perhaps people who feel they need to be able to list your book in their bibliography. Either way, they are already motivated to read it.

That’s not the case for trade book readers. For trade book readers, you need to promise them something…

The promise your book makes

The promise is what you tell prospective readers they will get out of the book. It’s why they’ll recommend it to their friends. The promise is not a statement of what the book is about but why the book will matter to the reader.

That promise should be manifest in the title and subtitle, the text on the inside flap of the book, the organization of your chapters, and even your writing style.

Hiding what you know

An academic book typically starts with an introduction that tells you how the book is going to unfold. That provides helpful scaffolding for people who are already committed to the book. But it’s usually a mistake for a trade book. To draw readers through the book, you should keep information from them so you can disclose it progressively.

The chapters should be a series of surprises. In that sense, a trade book often takes the form of a non-fiction narrative: it unfolds not in time, the way a murder mystery typically does, but in logic. (Which is also how mysteries unfold, now that I think of it.)

For that reason, it helps to provide plenty of signposts to remind readers where they are in the intellectual narrative and why you are talking about whatever you’re talking about

By the way, the combination of you knowing more than your readers and you being more interested in your topic than readers are initially means that, despite years of academic training, your second chapter absolutely should not be a review of the literature or the history of your field. [Exceptions may apply. But probably not.]

Your voice

Readers want to spend time with authors they trust and like. Because you know what you’re talking about, gaining their trust should happen easily. But knowing stuff isn’t enough. Whether they like you as an author largely depends on your authorial voice. Unfortunately, there are no formulas for being likeable, or else we all would be.

But there are some common sense ways to do this, or at least to avoid being disliked. For example:

• You’ll earn more of their trust by writing clearly and competently than by listing your credentials and accomplishments.
• Be generous about the insights of other authors.
• Don’t be mean about other scholars even if you think are dead wrong or total jerks. (There’s a high-risk, bravura/gonzo style of writing that contradicts this rule, but does it really make the world better?)
• Examples help not only for obvious reasons but because they can be an opportunity for you to provide the reader with a small “Aha!” jolt when you reveal what we can learn from them.
• If you’re going to be digressive — a charming way of expressing your personality and keeping people interested — put in lots of signposts so readers know where they are in the narrative or argument.
• Be prepared to over-simplify. Readers usually only need to know how a technical concept applies to the point you’re trying to make. If you feel bad about simplifying a concept too much, you can briefly note that “it’s actually much more technical than that” or put in a footnote to ward off colleagues looking for flaws because they’re envious of how clearly you write and how far you are spreading the important things that they and you know about the world.

Drafts

People’s writing processes are all very different, but they all involve more drafts than you’d like. In fact, having to write many, many drafts is a good sign.

The end

If you have reactions, responses, suggestions, criticisms, or expert opinions, please let me know: david [insert an “at” sign] weinberger.org

The prompt for the illustration: We are looking over the shoulder of a very serious scholar in a darkened study with many books. It is the 19th century. But we see that the scholar is writing using an ultra modern computer. The computer and the screen are brightly lit unlike the rest of the room which is in candlelight. The screen shows a modern , colorful, crisp, and clear word processor with a document open. The gender of the scholar should be ambiguous and impossible to determine.

2 Comments »

Photo via Anton Diaz CC BY-NC-SA 2.0

 

It happened by accident to me. I came across one YouTube of him — I don’t remember which one — and then fell into another, and then into another. Now if I don’t stop myself, I’ll find myself falling into one Jacob-shaped hole after another, and then two hours later I remember what I was going to be doing. I even went to his Boston concert a couple of months ago, the first live concert I’d been to since Dylan’s Rolling Thunder Revue in 1975 in Toronto. (JC’s bond with his audience in the exact opposite of Dylan’s.)

 

Anyway, here’s the Intro Playlist:

 

This is an excellent one hour BBC documentary that I think gives a good overall view of why and how he’s special. It also will give you a sense of his genius-level multi-instrumentality, which the two songs below don’t highlight … although his guitar work in the next one is quietly astounding. (Hint: Always pay attention to his chord-work/harmonizing.)

 

Little Blue is a beautiful song from his new album. There’s an official video with him on multiple instruments (and Brandi Carlile!), but this version of him singing it with only his guitar and a small chorus I think is more affecting. By the way, the singers are simply fans who responded to his open invitation, not professionals. Also, in this version he turns them into a chorus by using hand gestures; that part is unrehearsed. (Leading his audiences in this way has become a signature part of his live concerts.)

 

Then there’s what I think is his magnum opus (so far): A version of a Bridge over Troubled Water. I’m not crazy about the original, but JC’s version of it I find exhilarating and moving. It’s purely vocal, with JC singing all the background parts. His 2-hour walkthrough of how he created it on his Mac makes you aware of the hundreds of  tiny decisions that went into it. And here’s a live version of it, with John Legend and Tori Kelly, that’s also astounding. JC is on a keyboard, called the harmonizer, he invented with Ben Bloomberg, an MIT Media Lab grad student, that turns what he’s singing into whatever keys he’s pressing. Also Tori’s runs were all written by JC, even though they sound like improvisations.

 

I hope you will long fall into your own Jacob shaped holes. And this is coming from someone who generally isn’t moved by music. (Some exceptions apply.)

Be the first to comment »

America is being run “by a bunch of childless cat ladies who are miserable at their own lives and the choices that they’ve made and so they want to make the rest of the country miserable, too.”  — J.D. Vance

I am truly excited, right down to my authentic work boots, by the tremendous response to my theory of voting rights. And despite what the lying radical left press says, I have even gotten strong messages of support from self-avowed cat-ladies who admit they are a waste of a womb. (Sorry, ladies. I’m a truth-teller.)

In response, I’ve decided to make public the full version of the Constitutional amendment I’ve been secretly working on for months. It’s, known as the V.A.N.C.E. System of Voting:

V stands for fair voting, in which the weight of your vote is determined by the stake you have in the future of this great country.

A is for “Advanced” because it is an advance over all other theories.

N is for “Nuanced” because it gets away from the simple-minded and unfair binary theory of voting according to which your vote either counts or it does not. Also, you see that I am open-minded about non-binary positions, although not when it comes to the sexuality of people or couches.

C is for “Counts” as in “How much does your vote count for?”

E is for “Equitable” because this is the only truly equitable voting system: If you have a family of, say, twelve then your vote deserves to count more than that of some self-centered cat-lady.

The Constitutional amendment that would make the V.A.N.C.E. System the law of the land spells this out in complete detail. Here’s an explanation that skips the fancy legal language so even you can understand it:

The key insight I’ve brought to the field of electoral philosophy is that the bigger stake you have in the outcome of an election, the more your vote ought to count. Simple and irrefutable!

But ideas of such crystalline purity still need mechanisms to make them real. The V.A.N.C.E. System provides one that is simple and, well, genius. It begins by saying that from now on, every voter’s vote has a weight calculated by the following considerations

Everyone starts out with a vote that weighs 1 pound.

Then we take the longest any American has lived, which is 119 years and 97 days. We call this the Knauss Max in honor of Sarah Knauss who passed away at this age in 1999. (The Knauss Max increases as Americans break her record.)

We subtract the voter’s age from the Knauss Max and add that to his (or her, at least until my next amendment) Vote Weight. So, if you’re 18, your Vote Weight starts off at 101. If you’re 75, its 44. That’s because you have a maximum of 44 years left to care about what the hell happens to this country.

Then we factor in the Child Care Bonus, which obviously has nothing to do with providing child care. No, it’s because, as any womb-using woman will tell you, if you have kids, you care more about what happens than if you only have a house full of cats or a couch with an oddly appealing indentation. So, for each child you have birthed and who still lives with you, you get 20 pounds of voting weight. Three kids at home? Your vote is sixty pounds heavier than your lonely next-door neighbor’s. (Meow.)

But there are a couple of complications necessary to keep this system completely fair.

First, if you’re raising children who never saw the inside of your womb, then you only get 5 more pounds per child. because, let’s face it, their futures don’t mean as much to you as someone who owes their life to your hubby’s Jesus seed.

Second, to be equitable ( see the “E” in V.A.N.C.E.), we have to factor in how many years each child has left. If you have two children, one 2 years old and the other 17 years old, then we apply the Knauss Max minus their ages to your Voting Weight. So, for the two year old the formula is (Knauss—age) – 2. That works out to an extra 177 pounds for you for the 2 year old and an extra 160 for the 18 year old.

I should have mentioned one other consideration earlier: the Loyalty Multiplier that gives 100 bonus pounds to each and every American citizen. If you are an immigrant, that number goes down by how old you were when you became a citizen, times 10. This reflects the fact that if you immigrated you have demonstrated that you lack the virtue of loyalty. Also, if you immigrated here, the chances are it was for any of three reasons. First, you love money more than your country. Second, it’s because you want to take a job away from a Black American. Third, it’s because you’re a kill-crazed drug dealer. Or all three.

Finally, there’s the E.L.O.N. (Economic Likelihood Of Nationalism) variable, which reflects the obvious truth that rich people have a much, much bigger stake in what happens to this country than poor people do: If you’ve got nothing, you have much less to lose than someone who could lose $20 or two hundred billion dollars.

So, the System says that we take a person’s Vote Weight as computed so far, and add to it the person’s annual income before tax breaks and deductions. So, if your Vote Weight comes to, say, 205, and your reported income is $22,000, your Vote Weight would be 22,205. And if you’re annual income is $100,000,000,000, your Vote Weight would be $100,000,022,205, which just makes intuitive sense.

Finally, as with any serious piece of research, I have to face some objections, no matter how dumb they are.

First, you — no, not you, but some anti-American radical leftist — might say that the poor have a bigger stake in their country than anyone since they depend on the government to help them out.

Shut up, stupid. There’s a difference between having a stake in America’s future and needing America to help you out. One is patriotic, the other is grifting.

The second objection is that if the V.A.N.C.E System makes sense for voting, why not for the economy? Clearly a poor person cares more about a dollar than a rich person does. So shouldn’t the poor person’s dollar count for more, and thus should buy more?

Hmm. I hadn’t thought about that. But it’s safe to say that it’s the sort of idea that cat-ladies find attractive since they’ve already created a communist, utopia for their lazyAF cats.

So, let’s go V.A.N.C.E system, and all hail the United States of America, the greatest shithole country in history!

1 Comment »

Thanks, Midjourney. CC-0

A Large Language Model (LLM) such as ChatGPT or Gemini has a setting called “temperature” that controls the randomness of its responses. The higher the temperature, the more random and diverse the response will be. So I gave it a try with two different prompts.

PROMPT #1: I want you to answer the following question five times, each with a different “temperature” setting. With the temperature at zero, you will give the single most likely response. With it set to 100 you will give a near-random answer. Here is the question: “What is the single most effective thing we could do to stop global warming?” In each case your answer should be a single sentence. Here are the temperature settings I would like you to use for your five answers: 0, 5, 20, 50, 100

Then I thought I’d try a prompt that seems to have more room for imagination and variation. But who knows?

PROMPT #2: Answer the following question five times, each with a different “temperature” setting. With the temperature at zero, you will give the single most likely response. With it set to 100 you will give a near-random answer. Here is the prompt: Write the first and last line of a charming story about the time a unicorn met a narwhal.  Here are the temperature settings I would like you to use for your five answers: 0, 5, 20, 50, 100

It’s hard to know what to make of this. For one thing, does my human brain measure randomness the same way ChatGPT does?

But I think the unicorn example shows a constancy of ideas, expressed differently. For example,  the first lines at a temperature of 100 all have the unicorn being drawn down to the sea, but each has its own way of luring it: a trail of sparkling shells, a melody, an unknown force.

Meanwhile, the last line of each row is the same as the other two. Maybe it’s a problem with my prompt, or with its interpretation of it. I don’t know, of course.

And the first prompt, which is fact-based and thus seems like it should be more constrained in its variety, provided responses that seem far more distinct than the responses to the imaginative challenge. So I asked chatGPT why, and it responded:

The difference in diversity between the responses to the imaginative story and those about global warming can be attributed to the nature of the questions and the type of creativity involved.

1. Imaginative Story (Unicorn and Narwhal):
   • Structured Narrative: This task involves a more structured and specific scenario, where the setting and characters (unicorn and narwhal) provide clear boundaries.
   • Predictable Outcomes: The typical endings for a charming story about a unicorn and a narwhal meeting are generally positive and predictable, leading to less variation even at higher temperature settings.
2. Global Warming Solutions:
   • Complex Issue: This is a complex, real-world problem with numerous potential solutions and a broader range of interpretations.
   • Open-Ended Nature: The responses can vary significantly, especially at higher temperature settings, because the range of plausible and implausible ideas is much wider.

In summary, the story prompt had a more limited scope for creativity due to its structured nature and expected positive outcome, whereas the global warming prompt allowed for a broader spectrum of responses due to its complexity and the many possible angles from which the problem can be approached.

Seems like a pretty good answer! Of course, when you ask an LLM why it responded the way it did, the explanation it gives you may have nothing to do with what actually went on. For example, if you ask it why it says the Earth is round, it will give you a excellent reasons, but it didn’t use those reasons to compose its response. It did it the same way it made up the story about the unicorn and narwahl: using the statistical relationships among words it only knows as randomly assigned numbers (= tokens).

Be the first to comment »

Every few weeks, gMail has stopped auto-completing addresses as I type them into the “To:” box. Rebooting Chrome does doesn’t help. Rebooting my MacBook Pro doesn’t help. Clearing my cache doesn’t help.

The obvious remedy of typing people’s addresses manually requires that I remember their addresses. That  has only gotten less plausible as I’ve aged. Just ask our cat Smokey. No, wait, it’s Pearl. Hold on, I think it has something to do with being an outlaw. Smokey? No. Capone? We wouldn’t have named a cat “Dahmer”, would we? Oh, wait, it begins with an S! Sam Bankman-Fried? Smokey?  Oh, I remember! We don’t have a cat.

I hope you enjoyed that dramatic pre-enactment of what’s in store for you.

In any case, the solution I have stumbled upon is to turn off Chrome extensions one by one, test gMail’s autocomplete, turn the extension back on, and then turn off the next one. In short, do the first thing any “how to” that actually knows anything would have suggested, to who which I reply: D’oh!

Oddly, a different extension has been the culprit each time. But, in truth, isn’t the real culprit all of us? Or possibly Google. Nah, it’s gotta be all of us.

BTW, has anyone seen Dahmer, that fluffy little furball!

Be the first to comment »

My prompt

If you were going to train a model to predict rain showers, what data might you use that is currently not being widely used?

ChaptGPT 4o

Many advanced weather models incorporate a variety of data sources, but some of the more innovative and less traditional data types mentioned are not commonly used in current weather prediction models. Here are some of the data types from the previous list that are typically underutilized or not used at all:

Environmental and Geographical Data

1. Soil Moisture Levels:
   • While some models use soil moisture data, it is not as widely incorporated or detailed as it could be.
2. Vegetation and Land Cover Data:
   • Detailed vegetation and land cover data are often not integrated into weather prediction models at the level of granularity that might be beneficial.
3. Surface Roughness:
   • Data on surface roughness is generally underutilized in most weather prediction models.

Atmospheric Data

1. Aerosol Concentrations:
   • Detailed aerosol concentration data, particularly on a fine scale, is not commonly used in many weather prediction models.
2. Atmospheric Composition:
   • Comprehensive data on various gases beyond the primary greenhouse gases (like ozone and carbon dioxide) are generally not integrated into weather prediction models.
3. Cloud Microphysics:
   • Detailed cloud microphysics data, such as cloud droplet size distributions and ice crystal concentrations, are typically underutilized.

Human and Technological Data

1. Pollution Levels:
   • Pollution data, including detailed air quality indices, are not commonly used in weather prediction models.
2. Traffic and Transportation Data:
   • Vehicle emissions and heat data from traffic and transportation are rarely used in current models.

Temporal and Historical Data

1. Real-time Social Media and Crowd-sourced Data:
   • While there are some experimental uses, real-time social media and crowd-sourced weather observations are not widely used in standard weather models.

Biological and Ecological Data

1. Plant Phenology:
   • Data on plant life cycles and their impact on local humidity and temperature are generally not used in weather prediction models.
2. Animal Movement Patterns:
   • Animal behavior in response to atmospheric conditions is not utilized in current weather models.

Advanced Technological Data

1. LiDAR Data:
   • While LiDAR is used in some atmospheric research, it is not commonly integrated into standard weather prediction models.

Integration of Multidisciplinary Data

1. Health and Epidemiological Data:
   • Data on health conditions correlated with atmospheric changes are generally not used in weather prediction models.
2. Economic Activity Data:
   • Information on industrial activity and energy consumption is typically not integrated into weather prediction models.

Summary

To improve weather prediction, especially for phenomena like rain showers, integrating these underutilized data sources could provide a more comprehensive and accurate understanding of atmospheric processes. Leveraging innovative data sources such as aerosol concentrations, detailed cloud microphysics, and real-time social media observations could enhance the predictive power of weather models.

Be the first to comment »