Was trying to update sqlite version for Turso and add more tests, and somehow ended up finding 3 optimizer bugs in SQLite. Wasn’t expecting that, but hey, I’ll take it.
Folks here might be interested in this , it was presented today morning in sigmod
"DBugScribe: Automatic Database Bug Reproduction from Community Reports"
https://t.co/He61lWcoaK
"There is no magical checkmark for software correctness."
At Bug Bash 2026, @bugarela made the case for chasing confidence instead, and showed why the AI era raises the stakes.
Here's a breakdown 🧵
This is an astoundingly rapid change in something so basic. In 15 years the proportion of people who say college is very important has decreased by more than half.
Few things i realised on day 2 of sigmod
- Quite a lot of ppl don't know turso.
- tech database twitter is smaller and a bubble
- most people are as nervous as u, all u need to do is say your name and convo happens naturally.
- need glasses lol cant see anything clearly
Outputs that sound right are not the same as outputs we can prove right.
The next phase of AI runs on verified systems: formal methods, theorem provers, runtime checks, provable agents.
It's time to verify AI !
Join us on the 10th of June at SF !
@PramaanaLabs@boldcapfund
SPS Fellowship deadline extended to Sunday, June 14 (AoE).
Remote, 3-4 months, project-based. Formal methods meets AI security, with Atlas Computing. 14 mentors, including Erik Meijer & Shriram Krishnamurthi.
Apply: https://t.co/ySwxkM8u84
One of the most valuable things about Quint is not just that it finds bugs, it exposes behaviors engineers often don’t think to test.
The Turso team used Quint to model SQLite’s C API contracts, automatically generate traces, and replay them against the real implementation.
The result: 10+ bugs uncovered in SQLite while hardening Turso.
What’s especially compelling is the kind of behavior Quint explored systematically:
→ unexpected API call sequences
→ subtle state transitions
→ edge-case interactions
→ assumptions hidden inside implementation details
Instead of relying only on humans to imagine “what could go wrong?”, Quint explored the state space automatically and generated concrete counterexamples when contracts were violated.
Huge kudos to @Pavan4820, @Glcst, and the @tursodatabase team for the fantastic work.
I chatted with @duswnchl, based in South Korea, who works at the open-source consultancy, and worker-owned cooperative, @igalia. We covered her experience over the last 15 years working on WebKit and Chromium.
There is no paywall; give it a read!
🚨 A huge moment for the open source community in India.
For years, Linus Torvalds and Dirk Hohndel have taken the keynote stage together at Open Source Summits around the globe - and next month this must-see conversation is coming to Open Source Summit India!
You can’t miss this. Join us 16-17 June in Mumbai at the Jio World Convention Centre.
Register today! https://t.co/esOg8WNA98
#OSSummit
@KSHartnett , former senior writer at @QuantaMagazine, has written the inside story of Lean. The Proof in the Code, out June 9 from @QuantaBks.
Grant Sanderson (@3blue1brown): "Kevin Hartnett perfectly captures the unlikely story of Lean's birth and development."
🔗More here: https://t.co/mpmR9KOE1D
Specification. Verification. Optimization. The manual for engineering with agents.
Breaking down systems and agents at @resonatehqio in my SD'26 talk.
Boston, Jul 27–28
Turso wants to match and surpass SQLite's reliability. When I say "surpass", usually ppl just look funny at me. But this is one such example: One of our OSS contributors have just found 10+ bugs in SQLite using validation he built for Turso. @pavan4820 used @quint_lang to build a formal model of the system and then executed its traces to find corner cases where SQLite deviated from the spec.
It is a great demonstration of how modern reliability tools, formal methods in particular, can lead to reliable systems and find *many* issues even on the most stable software on Earth.
Read more: https://t.co/Y6soBNs2PD