[messe]

> Suspicions aroused, I clicked on the “Document History” button in the top right and saw a clean history of empty document – and then wham – fully-formed plan, as if it had just spilled out of someone’s brain, straight onto the screen, ready to share.

This isn't always a great indicator.

I can't stand Google Docs as an interface to write with, so use VIM and the copy/paste the completed document into it.

[el_benhameen]

Yep. I do this because I explicitly do not want a third party to see my thought process. If I wanted the reader to see my edits and second thoughts, I would have included them in the final document.

[like_any_other]

> my thought process

Don't forget about typing patterns, that could be used to deanonymize you across different platforms (anywhere that you type into a webpage that runs javascript):

https://www.bleepingcomputer.com/forums/t/759050/improve-ink...

[kashyapc]

Yeah, I often write locally in Emacs Org-mode, generate the HTML from it, and then copy/paste that into GDocs.

[twothamendment]

Another copy/paste reason - I can't count the number of times I've written up something for work on my own google account by mistake, then paste it into a new doc on the work account so I can share it.

[QuercusMax]

You really should use separate browser profiles...

[yjftsjthsd-h]

Or separate machines. It's not impossible to maintain sufficient separation in software, but it's a lot easier to skip the whole mess.

[tlavoie]

Even still, I have to make sure my KVM switch is connected to the correct laptop. (Work on one side, personal on the other, and shared keyboard / trackball)

[zephen]

> This isn't always a great indicator.

Right. Certainly not dispositive.

> use VIM and the copy/paste the completed document into it.

But he did mention tables. You'd think if they weren't just ASCII art, there'd be _some_ google docs history about fixing them up.

[jadedtuna]

Fairly certain e.g. copy-pasting from Obsidian will copy over the tables as well.

[fragmede]

every other time, paste with styling is the devil, but ever so occasionally, I do actually want that

[Izkata]

Different-sized headers too.

[GaryBluto]

It's bizarre to me that this didn't occur even slightly to the post author.

[NitpickLawyer]

As with many other things (em dashes, emojis, bullet lists, it's-not-x-it's-y constructs, triple adjectives, etc) seeing any one of them isn't a tell. Seeing all of them, or many of them in a single piece of content, is probably the tell.

When you use these tools you get a knack for what they do in "vanilla" situations. If you're doing a quick prompt, no guidance, no context and no specifics, you'll get a type of answer that checks many of the "smells" above. Getting the same over and over again gets you to a point where you can "spot" this pretty effectively.

[pessimizer]

The author did not do this. The author thought it was wonderful, read the entire thing, then on a lark (they "twigged" it) checked out the edit history. They took the lack of it as instant confirmation ("So it’s definitely AI.")

The rest of the blog is just random subjective morality wank with implications of larger implications, constructed by borrowing the central points of a series of popular articles in their entirety and adding recently popular clichés ("why should I bother reading it if you couldn't bother to write it?")

No other explanations about why this was a bad document, or this particular event at all, but lots of self-debate about how we should detect, deal with, and feel about bad documents. All documents written by LLM are assumed to be bad, and no discussion is attempted about degrees of LLM assistance.

If I used AI to write some long detailed plan, I'd end up going back and forth with it and having it remove, rewrite, rethink, and refactor multiple times. It would have an edit history, because I'd have to hold on to old drafts in case my suggested improvements turned out not to be improvements.

The weirdest thing about the article is that it's about the burden of "verification," but it thinks that what people should be verifying is that LLMs had no part in what they've received. The discussion I've had about "verification" when it comes to LLMs is the verification that the content is not buggy garbage filled with inhuman mistakes. I don't care if it's LLM-created or assisted, other than a lot of people aren't reading and debugging their LLM code, and LLMs are dumb. I'm not hunting for em-dashes.

-----

edit: my 2¢; if you use LLMs to write something, you basically found it. If you send it to me, I want to read your review of it i.e. where you think it might have problems and why you think it would help me. I also want to hear about your process for determining those things.

People are confusing problems with low-effort contributors with problems with LLMs. The problem with low-effort contributors is that what they did with the LLM was low-effort and isn't saving you any work. You can also spend 5 minutes with the LLM. If you get some good LLM output that you think is worth showing to me, and you think it would take significant effort for me to get it myself, give me the prompts. That's the work you did, and there's nothing wrong with being proud of it.

[satisfice]

You may be missing the point. The author’s feeling about the plan he was sent were predicated on an assumption that he thought was safe— that his co-worker had written the document that he claimed to have “put together.”

If you order a meal at a restaurant and later discover that the chicken you ate was recycled from another diner’s table (waste not want not!) you would likely be outraged. It doesn’t matter if it tasted good.

As soon as you tell me you used AI to produce something, you force me review it carefully, unless your reputation for excellent review of your own is well established. Which it probably isn’t— because you are the kind of guy who uses AI to do his work.

[jandrese]

Or the tell that the guy who usually writes fairly succinctly suddenly dumps five thousand words with all of the details that most people wouldn't bother to write down.

It would be interesting to see the history where the whole document is dumped in the file at once, but then edits and corrections are applied top to bottom to that document. Using AI isn't so much the problem as trusting it blindly.

[plorkyeran]

Dumping the entire file into google docs and then editing and corrections applied top to bottom is exactly my normal workflow. I do my writing in vim, paste it into google docs, and then do a final editing pass while fixing the formatting.

[like_any_other]

> the whole document is dumped in the file at once, but then edits and corrections are applied top to bottom to that document

This also happens if one first writes in an editor without spellchecking, then pastes into the Google Doc (or HN text box) that does have spellchecking.

[Lerc]

I have seen a number of write ups where I think the only logical explanation is that they are not conveying what literally happened but spinning narrative to express their point.

There was an article the other day where the writer said something along the lines of it suddenly occurred to them that others might read content they had access to. They described thenselves as a security researcher. I couldn't imagine a security researcher having that occur to them, I would think that it is a concept continually present in their concept of what data is. I am not a security researcher and it certainly something I'm fairly constently aware of.

Similarly I'm not convinced the "shouldn't this plan be better" question is in good faith either. Perhaps it just betrays a fundamental misunderstanding of the operation being performed by a model, but my intuition is that they never expected it to be very good and are feigning surprise that it is not.

[SkyeCA]

A world full of AI generated content and a world where we trust what we see seems to be mutually exclusive. I expect the default for a number of people going forward (myself included) will be extreme suspicion when presented with new images/videos/documents.

[pgwhalen]

It probably did, but they didn't feel the need to fully explain why they were confident it was AI generated, since that's not the point of the article.

[clickety_clack]

I also interact with Google Docs as little as possible. I draft in Notes or Obsidian and copy the text in. I just hate the platform.

[el_benhameen]

Oh, another fun one: I once got an offer letter via Docs. The edit history included the original paste from another candidate’s offer letter, including their name and salary. Useful for benchmarking!

[superultra]

I don’t need everyone seeing the dirty laundry of my first draft and edits. I too work in a working doc and then when completed I drop it at once into the final google doc.

[jchw]

On another similar but different note, I don't think I've ever uploaded any code written by LLMs to GitHub, but I do sometimes upload fully complete projects under my "initial commit". Some people may legitimately just hide the edit history on purpose just because they don't want to "show their work". It's not really a particularly good habit, but I think a lot of us can relate.

[LanceH]

A legit reason to hide your edit history is you might not remember what was in there. Say you have a moment of frustration and type out "this is an absolute garbage assignment by a braindead professor". Or you jot a quick note from the doctor because it happens to be open.

The simple fact is that the reader has no business reading the edit history, and the ability to make this happen should probably be far more prominent in document applications like Word or Google Docs.

[Veen]

Yes, I write everything in Obsidian and use "Paste from Markdown" in Google Docs. It's a habit I picked up years ago when Docs was much less reliable and lost work.

Plus, I want to deliver the completed document, not my edit history. Even on the occasions that I have written directly in Google Docs, I've copied the doc to obliterate the version history.

[mystifyingpoi]

Same here. Confluence web editor has a thousand options but no option to comfortably edit text. I always write the entire document in Neovim and then format it later (or never, in case of yet another "please explain this thing only you know but we will ignore this page and call you anyway when it breaks").

[exe34]

yep, emacs, version control that doesn't suck, all my notes in one place. I'll copy and paste what I need to share into whatever hellscape you want to live in, but my copy will remain safe.

[nereye]

Also, in some countries (e.g. Germany) applications explicitly do not track that information (such as how long a documented was edited) for legal reasons related to privacy laws.

[elgertam]

I do something similar. I write markdown, then render it and copy-paste that in

[necubi]

You can now paste markdown directly into Google docs (Edit -> Paste From Markdown)

(I have the same workflow, via Obsidian)

[kianN]

I do the exact same thing and this was my first thought. To be fair, I would probably not be able to format tables in a single cope/paste

[guerrilla]

Me too.