One of the most interesting things about LLMs as a technology is that you don’t have to use it to be powerfully deranged by it. You can get snowcrashed by using it. You can get snowcrashed by hating it.

i don’t think ai fully understands the lengths i’ll go to in order to make sure my writing sounds different than all the other stuff out there.

Even as L.L.M.s get better at producing fluid and plausibly human text, these persistent stylistic tics remain interestingly abrasive—in a single short answer, presented to you in a vacuum, A.I. text is as smooth as can be, but when you’re confronted with an overwhelming amount of it, the strangeness that’s been fine-tuned out really begins to re-assert itself.

I also had a conversation with a writer friend Anna Gát, and she was saying that she feels like she writes in a more punk way because she wants to prove that her work is human. And what that means is she’s looser with the rules of grammar. She is more experimental with her prose.

Here’s the thing, we can’t distinguish effort. How do I know you put a little bit of effort in?

Let’s draw the line: docs generated entirely by AI should never ship to users. You can use AI judiciously to audit your docs for gaps, to spot inconsistencies, to generate first drafts that you’ll edit later.

But the moment you let any LLM have the final word on what users will read, you’ve abandoned the basic contract of docs: that someone who understands the system has taken responsibility for explaining it truthfully.

AI has given people “absolute impunity to ignore reality […] AI is a direct attack on the way we verify information: AI both creates fake sources and obscures its actual sources.”

The best code is no code, programming still sucks and always will, and yet, I find myself still searching for the claw, the mark of mastery. Because that mark comes from people who want to reach other people directly. I want to see the claw, because if there is a claw, it means there is a living, breathing lion on the other side of the screen building the software that elevates us and binds us together as a community of software engineers.

The only thing I wish is that they had a default tag or notice that was always produced—“Summarized by AutoPR” or just “#AutoPR” would be enough. People should know when their content is made by bots.

I’m going to take this assertion one step further and make a stronger claim: in fact, the only thing computers can write is poetry. I think this is true for all methods of algorithmic writing, but it’s especially true of text that is generated with language models.