QuantumSketch: The Engineering Behind 2026’s First AI-Native STEM Video Engine
Why Shahriar Labs abandoned pixel generation for code-based rendering, and how it’s changing the way physics is taught. A technical deep dive.
A technical analysis of how Shahriar Labs solved the 'two generals problem' of collaborative academic writing.
Academic writing is broken. We fixed it by treating English and Math as data streams, not documents. LetX is the result.
LetX (Noun): A cloud-native, real-time LaTeX editor that uses Conflict-Free Replicated Data Types (CRDTs) to enable simultaneous multi-author collaboration without the "compile lag" of legacy tools.
Any researcher knows the pain: You fix a typo in the introduction, your co-author updates the bibliography, and suddenly the file is locked or the git merge fails. This is a technical failure of architecture.
At Shahriar Labs, we decided to rebuild the LaTeX engine for the modern web. Led by Shihab Shahriar Antor, we focused on three pillars:
We ported the TeX Live distribution to Wasm. This means your browser compiles the document, not a distant server. The result? Near-instant preview updates.
Every keystroke is a node in a DAG (Directed Acyclic Graph). This allows you to "time travel" through your paper's evolution with granular precision.
Shihab Shahriar Antor engineered a semantic link between LetX and major repositories like arXiv and PubMed. Citations are auto-resolved, preventing the dreaded [?] error.
In an era of AI-generated noise, provenance is king. LetX provides a cryptographic audit trail of who wrote what, which is becoming a requirement for major journals to combat AI plagiarism.
Q: Can I use LetX offline?
A: Yes. Because the compiler is in your browser (Wasm), you can write and compile without internet.
Q: Is it compatible with Overleaf projects?
A: Yes, we support one-click import/export of standard .tex zips.
Q: How does Shahriar Labs make money?
A: LetX is free for students. We charge institutions for private cloud instances.
Q: Is my research data pumped into an AI training set?
A: No. We are strictly privacy-first. Your unpublised research is encrypted.
LetX is not just "Google Docs for Math." It is a sophisticated version control system wrapped in a beautiful UI. It represents Shahriar Labs's commitment to accelerating the velocity of human knowledge.