In case you missed it, @bugarela joined @thegeeknarrator to get into a question that keeps coming up. Can we actually trust the code AI writes?
Full conversation below ↓
Full conversation below ↓
1
2
7
375
Quint retweeted
Just did something very cool: cut a release from Quint that is made of more external contributions than commits from our team.
Yes, in the age of AI slop PRs.
As any other sufficiently big opensource project, we struggled with those PRs since last year. But things seem to be better now. The PRs are often scoped to solve one specific issue that the person behind is was actually struggling with, and it's quite easy to validate and merge.
This feels great! Thank you for all contributions (and stars ⭐️) to the Quint repo 💜
Yes, in the age of AI slop PRs.
As any other sufficiently big opensource project, we struggled with those PRs since last year. But things seem to be better now. The PRs are often scoped to solve one specific issue that the person behind is was actually struggling with, and it's quite easy to validate and merge.
This feels great! Thank you for all contributions (and stars ⭐️) to the Quint repo 💜
1
2
22
626
While we are fully focused on making the human-centered benefits of formal specifications part of every team, it's great to see these kinds of advancements in hardcore formal verification. By advancing both sides, the world can have more trust in more software.
New on our Frontier Red Team blog: using Claude to formally verify software at large scale.
Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).
With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.
Read our research in full here: anthropic.com/research/forma…
1
2
20
766
Quint made it onto the @StackOverflow Podcast!
🎙️ We're joined by Gabriela Moreira @bugarela, CEO at @quint_lang, about why formal specification languages like TLA+ are essential for catching system design flaws before writing code, how Quint brings developer-friendly syntax and executability to formal methods, and the ways del-based testing can verify AI-generated code against real-world production data.
stackoverflow.blog/2026/10/16/for…
0
1
7
346
Leslie Lamport asks, “Who builds a house without drawing blueprints?” We ask, “Who chews a tree without listening for cracks?”
0
3
23
482
Quint retweeted
New TTFA episode is out! 🎙️
This time I sat down with Gabriella Mora, CEO of Quint, to talk about formal specifications, TLA+, temporal logic, and the increasingly important role of formal methods in the age of AI.
With AI generating more and more code, how can we make sure building actually behaves as intended?
We also discuss the story behind Quint, how it compares to Alloy, and the challenges of teaching formal methods.
I think this is a particularly timely conversation given everything happening at the intersection of AI and formal verification.
0:00 Intro
2:10 Gabriela's path: from Tumblr themes to a TLA+ thesis
11:56 Quint and TLA+ vs. Alloy
15:05 Why write a spec? Understanding over checkmarks
22:04 How AI changes the economics of formal methods
27:19 Executable specs as the middle ground for AI-written code
36:13 Keeping spec and code in sync: codegen, model-based testing, trace validation
39:04 Inside Quint Studio
47:53 Quint vs. TLA+: syntax, types, effects and the simulator
54:08 Model checking: a new Rust checker, TLC and Apalache
1:02:31 Temporal logic: modal operators, LTL vs. CTL
1:07:41 In practice: invariants, witnesses and cheap liveness checks
1:12:33 Bugs found in audits just by writing the spec
1:16:13 From Informal Systems and blockchain to Quint
1:23:38 Spinning Quint out and becoming CEO
1:36:19 Teaching formal methods as AI learned to write Quint
1:41:12 Teaching with games, and what students found hardest
1:50:56 What's next for Quint and Quint Studio
1:57:44 Wrap-up and outro
This time I sat down with Gabriella Mora, CEO of Quint, to talk about formal specifications, TLA+, temporal logic, and the increasingly important role of formal methods in the age of AI.
With AI generating more and more code, how can we make sure building actually behaves as intended?
We also discuss the story behind Quint, how it compares to Alloy, and the challenges of teaching formal methods.
I think this is a particularly timely conversation given everything happening at the intersection of AI and formal verification.
0:00 Intro
2:10 Gabriela's path: from Tumblr themes to a TLA+ thesis
11:56 Quint and TLA+ vs. Alloy
15:05 Why write a spec? Understanding over checkmarks
22:04 How AI changes the economics of formal methods
27:19 Executable specs as the middle ground for AI-written code
36:13 Keeping spec and code in sync: codegen, model-based testing, trace validation
39:04 Inside Quint Studio
47:53 Quint vs. TLA+: syntax, types, effects and the simulator
54:08 Model checking: a new Rust checker, TLC and Apalache
1:02:31 Temporal logic: modal operators, LTL vs. CTL
1:07:41 In practice: invariants, witnesses and cheap liveness checks
1:12:33 Bugs found in audits just by writing the spec
1:16:13 From Informal Systems and blockchain to Quint
1:23:38 Spinning Quint out and becoming CEO
1:36:19 Teaching formal methods as AI learned to write Quint
1:41:12 Teaching with games, and what students found hardest
1:50:56 What's next for Quint and Quint Studio
1:57:44 Wrap-up and outro
New TTFA episode! 🎙️
I sat down with @bugarela, CEO of Quint, to discuss formal specifications, TLA+, temporal logic, and how formal methods can help us trust AI-generated software.
We also explore the story behind Quint, its comparison with Alloy, and the challenges of teaching formal methods in the age of AI.
A timely conversation about where software engineering is heading!
youtu.be/SdwjRJgp6gs
0
6
19
653
Quint retweeted
New TTFA episode! 🎙️
I sat down with @bugarela, CEO of Quint, to discuss formal specifications, TLA+, temporal logic, and how formal methods can help us trust AI-generated software.
We also explore the story behind Quint, its comparison with Alloy, and the challenges of teaching formal methods in the age of AI.
A timely conversation about where software engineering is heading!
youtu.be/SdwjRJgp6gs
I sat down with @bugarela, CEO of Quint, to discuss formal specifications, TLA+, temporal logic, and how formal methods can help us trust AI-generated software.
We also explore the story behind Quint, its comparison with Alloy, and the challenges of teaching formal methods in the age of AI.
A timely conversation about where software engineering is heading!
youtu.be/SdwjRJgp6gs
3
7
21
2,524
Quint retweeted
@bugarela joins @Robbyrussel to ask a deceptively simple question: how do you know you can actually trust your tests?
Catch up here: buff.ly/h4gbg2V
Catch up here: buff.ly/h4gbg2V
0
3
5
141
We’re delighted to be sponsoring and attending
@gambiconf this November.
Come say hi to the Quint team!
@gambiconf this November.
Come say hi to the Quint team!
🏴☠️ Novo patrocinador a bordo: @quint_lang!
A Quint é uma linguagem open source de especificação executável. Com ela você descreve o comportamento do sistema e verifica se ele está certo antes que o bug chegue em produção.
0
4
12
307
Quint retweeted
🏴☠️ Novo patrocinador a bordo: @quint_lang!
A Quint é uma linguagem open source de especificação executável. Com ela você descreve o comportamento do sistema e verifica se ele está certo antes que o bug chegue em produção.
A Quint é uma linguagem open source de especificação executável. Com ela você descreve o comportamento do sistema e verifica se ele está certo antes que o bug chegue em produção.
1
2
19
729
Making software is a craft. We don't need to become managers. We are just crafting at a different level.
Find tools that respect your craft.
Find tools that respect your craft.
0
4
15
652
New podcast just dropped! Thanks for having @bugarela on 🙏
@bugarela joins @robbyrussell to talk about what happens when 800 tests still aren’t enough to trust your code.
They explore AI-generated tests, formal specifications, Quint, and why human judgment matters.
Listen here: buff.ly/pB1qbdx
0
2
8
307
One month of Quint Studio. Beta users have been generating specs, swiping behavior cards, and letting Studio write tests, and we're so grateful for everyone who's been in there with us.
Stay tuned for what's next. 💜
Stay tuned for what's next. 💜
0
2
9
649
Quint retweeted
I recorded a reaction video for the first time! I had some comments on Ray's video and thought recording it was better than writing a wall of text.
Part 1: Boris' Claude Win (I wanted to comment on this too!)
Part 2: Quint Studio (my comments on what the intention was behind some choices, etc)
Part 3: LLMs aren't everything (go to Ray's video for this part)
Hope you like this! Link: youtube.com/watch?v=t0VTYM…
Part 1: Boris' Claude Win (I wanted to comment on this too!)
Part 2: Quint Studio (my comments on what the intention was behind some choices, etc)
Part 3: LLMs aren't everything (go to Ray's video for this part)
Hope you like this! Link: youtube.com/watch?v=t0VTYM…
1
4
16
673