On Wed, May 20, 2026 at 4:31 PM Francisco Vila <[email protected]>
wrote:

> Of course. First, forgive my impulsiveness but that really struck a chord
> with me.
>
Well played, sir :-)
To the rest, it's very evident you're very invested in your work here,
which I find inspiring, thanks for sharing that.

> So yes, I’m just as proud of my work, even though these days an AI can do
> it better and faster than I can
>
Faster: yes, very likely
Better: no, I don't think so. Not today at least, in my experience. I see
human-language content that "locally" (a couple phrases? a half page?) is
quite good, but the "overall vision" is still very very drifty with these
things. It completely need human guidance.

If I may offer a different angle on the situation:
 - AI will provide a translation today that is not as good as what you
would provide. However, it's possible that at times when you are busy with
other work, this translation, though less good, might provide a level of
value to your users in a stop-gap kind of form, where they are able to
"limp along" instead of being completely blocked
 - if you elevate this approach to apply to the whole system, this puts you
in better control of where to spend your time: now that everything has a
bit of a "wash" of a basic translation, you can bring it to the level you
want it to be in a manner where you're selective and deliberate about where
your work goes, instead of having to chase a large number of rabbits all
running in different directions. I do realize this is more of an issue at
the start of a translation project where nothing is available, rather than
the ongoing incremental work that I would guess you currently face "daily",
all the same, it seems to me you being in control of your time is good for
you, no?

Personally I would expect human-curated translations will always be better
(or at least for a long time).
If the workflow is one where you're guiding an assistant, possibly you gain
"leverage" and one hour of your time at the machine, with luck, might
deliver N times the translated work (when compared with the previous
workflow), where N is ideally larger than 1 :-)

And just so we're clear: I strongly agree with where I see your heart
sitting,
the whole essence of this project and its use is to make things beautiful,
as good as they can possibly be,
so I simpatize with you quite strongly :-)

Thanks again for taking the time to write this all down,
I believe this is valuable for this conversation

L

-- 
Luca Fascione

Reply via email to