commented:
I dunno, Betteridge's law of headlines is giving long odds...

  commented:
Turso is one of those “hype-driven” projects a.k.a. Rust rewrite of
perfectly functional and stable software (and its ecosystem) done by
yet another VC-baked startup that over time starts pushing more “AI”
into the product. SQLite has its oddities but it has a one of the most
detailed test suites, and is used in billions of devices like mobile
phones.
After watching some interviews, I fell into similar trap with Lunatic
runtime that tried to be “BEAM for WASM” but turned out to be an
obscure can of worms that required you to rewrite every single lib to
ditch tokio for their odd macros and APIs that ran on top of tokio…
thus it died early. Author moved to different project that’s also dead
by now, and currently tries to do something related to jj which is
current hype train target for startups.

  commented:

Author moved to different project that’s also dead by now, and
currently tries to do something related to jj which is current hype
train target for startups.

I checked the GitHub profiles of the top five contributors to the
Lunatic runtime and none of them seem to be doing anything jj-related?
(Disclosure: I am the proprietor of one of those jj startups.)

  commented:
I'm really rooting for Turso to succeed. A fully-open SQLite + vector
search + encryption is already a very good value proposition for me.
Throw in sync + upcoming features, and I really see it becoming an
amazing product.
It's hard to trust a VC-backed startup with such a fundamental part of
an app, though. They do use the same SQLite file format, so that
should at least offer a possible way out in case they go belly-up or
get acquired.
To me it also seems a bit worrying that, if I understand correctly,
they basically abandoned development development on libSQL. I
completely understand, and it does make more sense to focus on one
thing only. Still, I was considering using it a few months back and
now I feel like I dodged a bullet. Seeing they now also started
working on AgentFS makes me fear for another pivot.

  commented:
What's interesting about a 'fully open SQLite' that they're not
actually developing? That's basically just saying 'patch our
software', when drh will already do that for you.
Personally I think their 'we're the new evolution of SQLite!' message
for libSQL leaves a bad taste in my mouth. SQLite is actively
developed, near-perfect software, and drh is a saint; portraying a few
patches and a CoC document as taking up some sort of mantle feels like
abusing the good faith of its public domain status. Compare to
fly.io's recent product poasting, which has no such air of disdain for
the original and portrays itself at all times as an extension instead
of a reinvention.
.