No judgment

This morning, for the first time, I used ChatGPT for my actual research. It was for a rapid text entry system that I’m developing.

I told it “list the 100 most common 3 letter words.” Worked like a charm. And the resulting list was indeed very useful.

A little later I told it “list the 240 most common 3 letter words.” This time it started out ok, but after 37 words it just started printing some random list of three letter words in alphabetical order.

This second interaction felt all too familiar, in a bad way. Like I was conversing with a student eager to give an answer — any answer — even if the student didn’t know the material.

As it happened, a few hours later I ran into Yann LeCun. We chatted a while and I told him of my experience this morning. He wasn’t at all surprised.

He pointed out that ChatGPT doesn’t really have any idea what you are asking it. So it actually doesn’t care whether its search of its vast database comes up with something incorrect.

Yet another reminder that ChatBots are not at all intelligent in the human sense. In particular, they have no judgment. Still, they are very useful tools if you accept them for what they are.

What didn’t happen

Today I had a near miss with disaster. Nothing bad happened, but something very bad almost happened.

And it occurred to me that our lives are continually sustained by a series of fortunate events, which conspire to keep us from disaster. We don’t notice them, because they are all about what didn’t happen.

Future blog

At some point in the future, I will just be able to write a few words here. My chatbot will expand those thoughts into a fully polished and well researched blog entry.

But by then people will no longer have the discipline to read through anything like a fully polished and well researched blog entry. So there will be software on the other side that will conveniently reduce that voluminous output down to a few words.

So I will only need to jot down a few words, and you will only need to read a few words. Everybody will be happy.

And nobody will be the wiser. In more ways than one.

Somebody saw this coming

Since we’ve been discussing ChatGPT, this is a good moment to pay tribute to Roald Dahl. You might know him as the author of Charlie and the Chocolate Factory, or James and the Giant Peach.

But he was much more than that. Dahl’s stories were often dark cautionary tales. They ran the gamut from children’s stories to other more challenging tales that were definitely meant only for grownups.

One story in particular, published in 1953 — nearly 70 years ago — was The Great Automatic Grammatizator. In this story Dahl rather precisely predicts chatbots such as ChatGPT.

But in this story he does more than that. He also predicts the potentially devastating effects such a technology might have on authorship in general. Maybe you should read it.

A bunch of bananas

Of course I tried ChatGPT. How could I resist? I learned that it is really good at some things, and terrible at others. It is wonderful at answering scientific questions, or giving code examples.

But my favorite thing about it is how you can ask it to describe any topic in the style of any author. I asked it to describe a banana in the styles of various famous people. See if you can guess the author of each description. They are roughly in order from least to most terrifying.


“The banana, yellow and sleek,
Lies coiled upon the plate.
Its smooth skin holds a secret,
A taste both sweet and great.

I peel back the layers,
And reveal the fruit within.
Its soft, creamy flesh,
A symphony of sin.

I take a bite and savor,
The taste that fills my mouth.
A hint of tropical sunshine,
A hint of the South.

The banana, strange and exotic,
A fruit both plain and grand.
It tempts and teases,
And holds me in its hand.”


“A banana is a product of capitalist society, created through the exploitation of the labor of the global south. Its bright yellow skin and sweet, tropical flavor are the result of the exploitation of Third World farmers and the plundering of the natural resources of the global south.

The banana is a commodity, produced for the benefit of the capitalist class and consumed by the proletariat. Its production and distribution are controlled by powerful multinational corporations that seek to maximize profit at the expense of the workers and the environment.

But the banana is more than just a commodity. It is a symbol of the global capitalist system and its inherent contradictions. Its production and distribution are linked to the exploitation of labor, the destruction of the environment, and the exploitation of the global south. The banana is a reminder of the need for revolutionary change and the overthrow of the capitalist system.”


“The banana sits on the counter, a bright yellow curve against the white porcelain. I pick it up, feeling the smooth skin give way under my fingers. It’s almost too perfect, too smooth and unblemished, like a plastic toy. I peel back the skin and reveal the pale, soft fruit inside. It’s almost obscene, the way it yields to my touch. I take a bite and feel the sweetness burst in my mouth, like a wound that won’t heal. It’s sickly, cloying, a reminder of all the things I can’t have. I toss it aside, the half-eaten fruit mocking me with its false promises of nourishment and sustenance. I turn away, my stomach churning with disgust and something like desire.”


“The banana is a curious and unsettling fruit, with its smooth, curved shape and bright yellow skin. It seems almost alien in nature, as if it were brought to our world from some far-off, eldritch realm.

As I peel back the layers, revealing the soft, white flesh beneath, I can’t shake the feeling that I am uncovering something ancient and forbidden. The sweet, tropical aroma fills my senses, almost hypnotizing me with its allure. But as I take a bite, I can’t help but feel that I am tasting something otherworldly, something that should not be known to mortal man.

The banana is a mystery, a portal to strange and ancient powers that threaten to unravel the very fabric of reality. Beware the banana, for it holds within it a terror that defies description.”


You wouldn’t believe it, folks, but I’ve got the best banana in all of the land. It’s huge, it’s beautiful, it’s yellow, it’s just incredible. I’m telling you, this is the best banana you’ve ever seen. It’s got the most amazing flesh, it’s so soft and creamy, it just melts in your mouth. And the smell, oh boy, let me tell you, it’s just unbelievable. It’s got this sweet and fragrant aroma that just fills the room.

But it’s not just about the looks and the taste, folks. This banana is a winner in every way. It’s so versatile, you can use it in so many different ways. Smoothies, cakes, curries, you name it. And it’s got the most incredible shape, it’s just like a curve of the female form. It’s just perfection.

So if you want the best banana out there, folks, you’ve got to come to me. I’ve got the best banana, and I’m not afraid to say it. Believe me, it’s the best. You won’t find a better banana anywhere. Trust me, folks.

A good day for birthdays

Today, December 16, is the birthday of Ludvig Von Beethoven, Jane Austen, Arthur C. Clarke and Philip K Dick.

Oh, and also Margaret Mead and Noël Coward. All in all, a good day for birthdays.

Wouldn’t it be wonderful if all of those people could be magically brought together into a room, simply to have a conversation with each other? I would love to be a fly on the wall for that!

Thinking about those people reminds me how important it is to remember the great minds of the past, and all that we have learned from them. As George Santayana said, “Those who cannot remember the past are condemned to repeat it.”

Today is his birthday too.

A gift

I just read a story that was written many decades ago. The author is long dead, as is the person who wrote the breathless introduction to the story in question.

As I was reading, it occurred to me that these people were part of a group, a cohort. They were writing with enthusiasm, and reading what the others were writing as well.

They likely threw parties, swapped stories, and gossiped behind each others’ backs. I am sure that friendships formed and fell away.

They were not thinking “Oh, one day we will all be dead and gone, our words a hollow echo of these lives once lived.” No, they were in the midst of living those lives, very much in the present.

I suspect that this is not what the author of the story would have liked me to be thinking about, but there it is. The ephemerality of the moment floats over every page.

And I am reminded how wonderful it is to be alive, right now. It is a gift not to be taken for granted

Future energy

I know that the fusion “ignition” achieved last week has no short term impact on energy infrastructure. Still, an important milestone was crossed.

Realistically, it will likely be several more decades before we can get to where fusion begins to live up to its potential as an energy source. But what will happen then?

In particular, if energy were to one day become vastly cheaper and vastly higher capacity, what would change? It seems to me that two obvious places are computation and transportation.

Currently about 30% of the cost of taking a long flight is the cost of fuel. But what if the cost of fuel went down by a factor of 100, or 1000?

In that case it might make sense to use far more energy for the same flight. And that might make it possible rethink the entire process. Some future variation on flight technology, designed around greater energy capacity, could end up being much cheaper.

And of course computation is bound mainly by the availability of electric power. Server farms are energy hogs, and the relationship of “energy in == computation out” is pretty well defined.

If it were possible to draw 100 or 1000 times more power, we could completely rethink the economics of computation. A.I. algorithms that currently take minutes or hours could be run at interactive rates — with every individual getting their own generous slice of the pie.

Imagine the possibilities.

Wednesday

I know it’s Tuesday, but today I would like to talk about Wednesday.

I am referring to the new Netflix series. I just saw the first episode, and I am astonished by how good it is.

We’re talking Andor level good here, and that is very high praise. I am still trying to analyze what makes this show so compellingly watchable, and I think I’m starting to figure it out.

Of course Jenna Ortega is brilliant in the eponymous role, but it’s more than that. I think it has a lot to do with the way the writers invite their viewers to be part of the meta-fiction at the heart of the Addams universe.

The appeal of Charles Addams’ cartoons was always that we, the audience, are in on the joke. The impossibly spooky Addams family manages to hold a perfect mirror up to the “normal” society whose norms it is defying.

Remember that Addams was doing this starting back in 1937. That’s 85 years ago.

And yet his fundamental point still resonates perfectly today: “Normal” society is incredibly weird. We just don’t care to admit it.

The creators of this new show have perfectly captured the humor and wisdom of that brilliant insight. Frankly, dear reader, I cannot stop watching.

Future creatives

Let’s skip forward to the near future where any painting you want to create or essay you want to write is manifested simply by describing it to an A.I. One problem with such a future is that an A.I. can only imitate styles that already exist.

And those styles do not come from computers. Computers, unlike us, are not sentient. They possess no actual judgment or opinions or creativity.

The wondrous A.I. that so many people are now playing with is an elaborate mirror. It is merely echoing back to us our own collective creative thoughts, which it obediently recombines in all sorts of ways.

As these sorts of A.I. become widely used, there will be less incentive for people to put the effort in to create new works that introduce original ideas. If we don’t want our culture to devolve into endless variations of the same unoriginal patterns, we will need to find a way for those ideas to make their way into the mix.

I wonder whether a new sort of specialty will emerge — creative people who continually insert new ideas into the ever more powerful A.I. recombination engine that everyone else will use.

It is not clear to me whether these future creatives will actually write entire novels or or paint pictures or pick out songs on their guitars. They may instead use some sort of computer interface that will immediately fill out and instantiate their creative ideas.

But they won’t be just using the A.I. engine the way most people will. They will have a more serious purpose — preventing future culture from devolving into a mass of endlessly repetitive and meaningless treacle.