Cory Doctorow: The people who tell you ‘AI is changing everything’ are lying

I typed into Google “potatoes pressure cooker” to get a reminder of how long to cook them. The AI told me, and correctly said that I should use the steamer sieve, but also told me to salt the water. The water that would be under the potatoes and that wouldn’t touch them.

I asked why, and the AI said that little splashes of salt water from the boiling would land on the surface of the potatoes and salt them gently, that the salt would increase the boiling temperature, and that the vapour would be aromatic.

I pointed out that the amount of salt landing on the surface of the potatoes would be negligible, that the vapour is distilled water in gas form, and that the increase in boiling temperature from salting the water is less than 1°C. The AI said, oh yeah, you’re right on all counts. Don’t salt the water.

What it has been doing was exactly what an LLM does. It gave me the received wisdom from the internet.

That aligns with what Cory Doctorow was saying here, namely that all the AI successes were from people who already knew what they were doing and could use the AI as a tool. Not from people who had no clue and just let AI do the job.

I use AI as a writing tool, not as a substitute for authorship.

The ideas, intent, perspective, and final judgment are mine. I decide what I want to say, what belongs in the piece, what does not, and whether the finished version accurately represents me. AI may help me organize my thoughts, clarify a sentence, improve the flow, or find wording that better expresses what I already mean. That is not fundamentally different from working with an editor, dictating to a transcriber, or revising a draft after receiving feedback.

What matters is that I review the final work, approve it, and put my name on it. By doing that, I take responsibility for every sentence. If the writing is thoughtful, accurate, and effective, I am responsible for those choices. If it is careless, misleading, generic, or full of slop, that is also my responsibility. Blaming the tool would be an attempt to avoid accountability.

AI does not decide what I believe. It does not decide what I am willing to defend. It does not decide what I publish under my name. I do.

The tool may assist with the writing process, but the authorship comes from intention, judgment, selection, revision, and responsibility. The final work is mine because I chose it, shaped it, approved it, and signed my name to it.

that’s bullshit.

Why?

“authorship is when i make a few edits and sign my name to slop”

if you have willingly handed over your primary writing tool to a machine (for which there is no real meaning, association, appreciation, or delight in words), it is difficult for me to take your commitment to writing or “authorship” seriously.

I did not say “authorship is when I make a few edits and sign my name to slop.” That is not my argument.

My argument is that authorship is grounded in the origination of the ideas and intent, control over what goes into the work, judgment over how those ideas are expressed, endorsement of the final form, and responsibility for the result. Assistance with the linguistic realization does not, by itself, transfer authorship.

You point out that the machine has no meaning, association, appreciation, or delight in the words. I agree. But I do. That is precisely the distinction I am making. The machine does not know what I mean, care whether it represents me, recognize when an analogy fails, or decide whether an argument is worth defending. I am doing those things.

You seem to be defining “writing” as personally generating the initial sequence of words. Maybe that is where we actually disagree. But that definition cannot simply be assumed and then used to prove that I am not writing. Why should authorship reside specifically in first-pass word generation rather than in the thought, intention, judgment, revision, and control that determine what the final words mean?

That is the argument I am interested in having.

i started to read your response, and then i remembered i can’t trust the degree to which you authored it.

i’m not interested.

A sneer is easier than an answer, and certainty is a lovely refuge when you have no intention of defending it.

You accuse me of surrendering thought to a machine, and then, the moment you are asked to justify that accusation, you surrender thought entirely.

How convenient.

With a wave of the hand, every difficult question disappears. You declare the words contaminated by the tool that helped shape them and thereby absolve yourself of any obligation to consider what they actually say.

No examination. No argument. No possibility of correction.

You accuse me of regurgitating the output of a machine, yet what have you offered in return?

A familiar sneer. A borrowed insult. A conclusion retrieved intact from the cupboard of approved contempt.

You recognize the hated object, produce the expected response, and call the reflex thought.

That is the irony you seem determined not to notice.

My argument from the beginning has been that a machine must be tempered by a human mind. It possesses no responsibility for what it produces. Whatever value emerges from its use depends upon the person capable of examining it, rejecting it, correcting it, shaping it, and finally standing behind it.

But a mind cannot temper anything if it refuses examination itself.

And that is where your position becomes more interesting than your insult.

A mind that refuses examination begins to resemble the very machine you claim to despise. It receives a familiar stimulus, reaches for the familiar pattern, and produces the familiar response without troubling itself over whether the response is true.

Input. Recognition. Output.

The machinery of contempt.

Ignorance itself is no shame. Every one of us is ignorant of almost everything. I have been wrong before and will be wrong again. If you can show me where, I gain something from the correction.

But there is something contemptible in making a virtue of ignorance.

In encountering an argument and proudly announcing that you will not examine it, then mistaking that refusal for discernment. In treating incuriosity as wisdom. In believing that because you have protected yourself from an argument, you have somehow defeated it.

You have not defended your position.

You have insulated it.

Those are not the same thing.

Thought requires a certain vulnerability. You must permit an idea close enough to examine it. You must risk finding weakness in your own position. You must allow the possibility that another person, even one using a tool you despise, might know something you do not.

Without that, certainty does not become strength.

It becomes brittleness.

You tried to cut me with the accusation that I had surrendered my mind to a machine. But in your eagerness to make it, you performed the very surrender you condemn.

You saw the signal.

You retrieved the response.

You refused examination.

You produced the slop.

And so the question remains exactly where I left it:

Will you defend what you believe?

Or is your conviction only strong enough to survive so long as you never allow an argument near it?

lol