http://simonwillison.net/2025-12-09T03:11:19+00:00Simon Willison2025-12-09T03:11:19+00:002025-12-09T03:11:19+00:00https://simonwillison.net/2025/Dec/9/formal-verification/#atom-tag

<p><strong><a href="https://martin.kleppmann.com/2025/12/08/ai-formal-verification.html">Prediction: AI will make formal verification go mainstream</a></strong></p> Martin Kleppmann makes the case for formal verification languages (things like <a href="https://dafny.org/">Dafny</a>, <a href="https://github.com/marcoeilers/nagini">Nagini</a>, and <a href="https://github.com/verus-lang/verus">Verus</a>) to finally start achieving more mainstream usage. Code generated by LLMs can benefit enormously from more robust verification, and LLMs themselves make these notoriously difficult systems easier to work with.</p> <p>The paper <a href="https://arxiv.org/abs/2503.14183">Can LLMs Enable Verification in Mainstream Programming?</a> by JetBrains Research in March 2025 found that Claude 3.5 Sonnet saw promising results for the three languages I listed above. <p><small></small>Via <a href="https://lobste.rs/s/zsgdbg/prediction_ai_will_make_formal">lobste.rs</a></small></p> <p>Tags: <a href="https://simonwillison.net/tags/predictions">predictions</a>, <a href="https://simonwillison.net/tags/programming-languages">programming-languages</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/ai-assisted-programming">ai-assisted-programming</a>, <a href="https://simonwillison.net/tags/martin-kleppmann">martin-kleppmann</a></p>

2024-04-27T19:31:08+00:002024-04-27T19:31:08+00:00https://simonwillison.net/2024/Apr/27/martin-kleppmann/#atom-tag

<blockquote cite="https://bsky.app/profile/martin.kleppmann.com/post/3kquvol6s5b2a"><p>I've worked out why I don't get much value out of LLMs. The hardest and most time-consuming parts of my job involve distinguishing between ideas that are correct, and ideas that are plausible-sounding but wrong. Current AI is great at the latter type of ideas, and I don't need more of those.</p></blockquote> <p class="cite">&mdash; <a href="https://bsky.app/profile/martin.kleppmann.com/post/3kquvol6s5b2a">Martin Kleppmann</a></p> <p>Tags: <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/martin-kleppmann">martin-kleppmann</a></p>

2020-09-28T21:03:57+00:002020-09-28T21:03:57+00:00https://simonwillison.net/2020/Sep/28/i-was-wrong-crdts-are-future/#atom-tag

<p><strong><a href="https://josephg.com/blog/crdts-are-the-future/">I was wrong. CRDTs are the future</a></strong></p> Joseph Gentle has been working on collaborative editors since being a developer on Google Wave back in 2010, later building ShareJS. He’s used Operational Transforms throughout, due to their performance and memory benefits over CRDTs (Conflict-free replicated data types)—but the latest work in that space from Martin Kleppmann and other researchers has seen him finally switch allegiance to these newer algorithms. As a long-time fan of collaborative editing (ever since the Hydra/SubEthaEdit days) I thoroughly enjoyed this as an update on how things have evolved over the past decade. <p><small></small>Via <a href="https://news.ycombinator.com/item?id=24617542">Hacker News</a></small></p> <p>Tags: <a href="https://simonwillison.net/tags/algorithms">algorithms</a>, <a href="https://simonwillison.net/tags/collaboration">collaboration</a>, <a href="https://simonwillison.net/tags/crdt">crdt</a>, <a href="https://simonwillison.net/tags/martin-kleppmann">martin-kleppmann</a></p>