Sander Spies retweeted
introducing TScert, a sound core for TypeScript.
- TypeScript functions accepted only with a proof
- a checker whose soundness is proved in Lean
- first target: its own front end, 628 of 684 functions verified
a research project, unfinished by design. read it, inspect it 👇
Sander Spies retweeted
O(x)Caml has been an incredibly good target for agentic code in my group, so I jotted a quick blog guide for those who might be persuaded to try out our favourite double-humped ungulate language with agents anil.recoil.org/notes/ocaml-…
Sander Spies retweeted
hey everyone, there's little time left to put in a talk or workshop proposal for FUN OCaml 2027 in Bangalore, India, January 25+26!
so please go ahead and put in some ideas!
Sander Spies retweeted
the FUN OCaml website is updated for 2027!
we're meeting in Bengaluru, India on January 25+26, 2027!
the CFP is open and you can find it linked from the website!
fun-ocaml.com/
Sander Spies retweeted
JUST IN: Man dressed as Darth Vader appears before San Diego City Council to defend its 550+ Flock cameras — declaring, “This is what the emperor needs.”
Sander Spies retweeted
I'll be back in 10 minutes’ is the biggest lie every developer has ever told. 💀
Sander Spies retweeted
Bring your whole team and dozens of AI coding agents into the same chat threads, then let @linzumi_ai keep the fleet coordinated and unblocked.
And for a limited time, try state of the art open-weights intelligence free: GLM 5.2 at high speed, via their @wafer_ai partnership. → linzumi.com
Sander Spies retweeted
I would give all the security clearances in the world to this guy.
The FBI urges everyone—especially those who hold, or have held, security clearances—to be cautious when approached by individuals online regarding career opportunities.
Learn more about how to protect yourself from foreign virtual targeting at: fbi.gov/news/stories/foreign….
Today marks a big milestone for Callstack! 🎉
We just launched Apex: our own language model fine-tuned for React Native development.
This means we are no longer only a consulting company helping teams ship cross-platform apps and adopt AI engineering practices. We can now train, host, and publish specialized models ourselves, on demand, at a much lower cost.
It unlocks so much potential: freedom from commercial AI labs, unlimited tokens for our teams (as long as infra holds), a base for building custom agents and swarms that talk to our API, you name it.
Excited for what's next! It is obviously just a beginning and the work is going to take months, but I love that kind of journey.
Will be fun!
Sander Spies retweeted
Dear frontend devs and UI designers. I bring you Liquid DOM, a complete and faithful implementation of Liquid Glass on the Web.
- Shape morphing
- All properties animatable
- Dynamic refraction and reflection
- Adaptive tint
- Adaptive specular highlight
- Dispersion
- Full html integration
- Super fast layout engine that works across Canvas and html
- Pointer event handling
- Framework and renderer-agnostic low level API
- High level React API
- Ootb @threejs and r3f integration
And lots more.
Read on for implementation details and demos.
Sander Spies retweeted
The Sam Altman and @miramurati texts from the day he got fired from @OpenAI in 2023 just became evidence in the @elonmusk v. @sama trial.
It felt like a meaningful moment in AI history, so I turned it into a musical.
The lyrics are the texts.
Sander Spies retweeted
A lot of AI safety stuff is overblown fearmongering for clout but this AI cheerfully going “Sure!” and popping a cap in this guy made me laugh out loud
Sander Spies retweeted
My friend Milla Jovovich and I spent months creating an AI memory system with Claude. It just posted a perfect score on the standard benchmark - beating every product in the space, free or paid.
It's called MemPalace, and it works nothing like anything else out there.
Instead of sending your data to a background agent in the cloud, it mines your conversations locally and organizes them into a palace - a structured architecture with wings, halls, and rooms that mirrors how human memory actually works.
Here is what that gets you:
→ Your AI knows who you are before you type a single word - family, projects, preferences, loaded in ~120 tokens
→ Palace architecture organizes memories by domain and type - not a flat list of facts, a navigable structure
→ Semantic search across months of conversations finds the answer in position 1 or 2
→ AAAK compression fits your entire life context into 120 tokens - 30x lossless compression any LLM reads natively
→ Contradiction detection catches wrong names, wrong pronouns, wrong ages before you ever see them
The benchmarks:
100% recall on LongMemEval — first perfect score ever recorded. 500/500 questions. Every question type at 100%.
92.9% on ConvoMem — more than 2x Mem0's score.
100% on LoCoMo — every multi-hop reasoning category, including temporal inference which stumps most systems.
No API key. No cloud. No subscription. One dependency. Runs on your machine. Your memories never leave.
MIT License. 100% Open Source.
github.com/milla-jovovich/me…
Readers added context they thought people might want to know
The claimed 100% LongMemEval score uses targeted fixes for the 3 failing questions and LLM reranking (held-out score: 98.4%). The 100% LoCoMo score uses top-k=50 exceeding session count with reranking (honest top-10 no rerank: 88.9%). github.com/milla-jovovich…





























