Jevlish is a JavaScript/TypeScript frontend for the Jev judgment system that integrates semantic AI decisions into application logic. It provides functions like given and from to evaluate values and filter collections using plain-language questions alongside traditional code predicates, with results that include model confidence and decision evidence.
This paper introduces xeno-interpretability, a framework for studying internal representations in large language models that may lack human conceptual equivalents. The authors argue that LLM representational spaces exceed what can be expressed through human language, and propose methods to identify and characterize these model-native structures even when their semantic content cannot be fully translated to human terms. The work highlights implications for AI safety, as such representations could propagate unpredictably across interacting agents.
The article distinguishes between prompted intelligence (where AI assists human decision-making) and engineered intelligence (where AI judgment is built directly into software architecture). Smart software integrates intelligence as a core operational component rather than as a chatbot or external tool, requiring new software engineering practices that manage context, outcomes, uncertainty, authorization, and error handling within the system's control loop.
Geatsc is a TypeScript-to-C++ compiler built as a clean-room rewrite with strict design constraints: no boxing of dynamically-typed values, fail-closed guards, and source-neutral architecture. The compiler features layered components from frontend semantics through typed IR to C++ emission, with enforcement scripts that prevent architectural violations and comprehensive diagnostic reporting.
Semfont is a typography library that automatically highlights, colors, bolds, and italicizes text based on semantic analysis. It scores each word across four dimensions—valence, salience, surprise, and certainty—using dictionary lookups and contextual rules, then applies a second pass to handle negations and verb relationships across clause boundaries.
Researchers propose Cache-to-Cache (C2C), a new communication paradigm enabling direct semantic exchange between large language models through their KV-caches rather than text. The approach uses neural networks to project and fuse caches between models, achieving 3.1-5.4% accuracy gains over text-based communication and delivering 2.5x latency speedup.
Jev is a tool that visualizes concepts in 2D semantic space by positioning phrases based on their meaning relative to two user-selected axes, with scores from 0 (opposite) to 100 (close match).
Apache Ossie is an open-source specification standard for exchanging semantic metadata across analytics, AI, and BI platforms, using declarative YAML to define metrics, dimensions, and joins. It provides a vendor-neutral single source of truth to eliminate metric drift and ensure consistent data definitions across tools and organizations. The project was recently accepted into the Apache Incubator under its new name, having previously been known as Open Semantic Interchange.