Comment by SCdF
6 hours ago
> With AI, if well done, it would be useful nearly everywhere.
I fundamentally disagree with this.
I never, now or in the future, want to use AI to generate or alter communication or expression primarily between me and other humans.
I do not want emails or articles summarised, I do not emails or documents written for me, I do not want my photos altered yassified. Not now, not ever.
Keep in mine I said "if well done". That was not meant to imply that I think the current AI offerings are well done. I'd take "well done" to mean that it performs the tasks it is meant for as well as human assistants perform those tasks.
> I never, now or in the future, want to use AI to generate or alter communication or expression primarily between me and other humans. [...] I do not want emails or articles summarised, I do not emails or documents written for me, I do not want my photos altered yassified.
That's fine, but generally the tools involved in doing those things are designed to be general purpose.
A word processor isn't just going to be used by people writing personal things for example. It will also be used by people writing documentation and reports for work. Without AI it is common for those people to ask subordinates, if they are high enough in their organization to have them, to write sections of the report or to read source material and summarize it for them.
An AI tool, if good enough to do those tasks, would be useful to those users, and so it makes sense for such tools to be added by the word processor developer.
Again, I'm not saying that the AI tools currently being added to basically everything are good enough.
The point is that
(1) a large variety of tools and products have enough users that would find built-in AI useful (even if some users won't) that it makes a lot of sense for them to include those tools (when they become good enough), and
(2) AI may be unique compared to prior advances/fads in how wide a range of things this applies to and the speed it has reached a point that companies think it has become good enough (again, not saying they have made the right judgement about whether it is good enough).
How about machine translation and fixing grammar in languages you're not very familiar with? That's the only use of "AI" I've found so far. I'd rather read (and write) broken English in informal contexts like this forum, but there are enough more formal situations.
Remember, I am responding to this:
> With AI, if well done, it would be *useful nearly everywhere.*
I'm not saying it doesn't have uses.
Having said that, there are two things I never want AI to do: a) degrade or remove the need for me to express myself as a human being, b) do work I'd have to redo to prove it did it correctly.
On translation, sycophancy is a problem. I can't find it now, but there was an article I read about an LLM mistranslating papers to exclude data it thought the user wasn't interested in. So no, I wouldn't trust it for anything I cared about.
I do use AI: I'm literally reviewing some Claude generated code at the moment. But I can read that and know that it's done it right (or not, as the case often is). This is different from translation or summarisation, where I'd have to do the whole task again to prove correctness.
If you're not familiar, how could you possibly know if what you're conveying is accurate to your intention? And if you don't, why bother at all?