technocracy

arXiv:2603.08755v1 Announce Type: new Abstract: We present textbf{Turn}, a compiled, actor-based programming language — statically typed for schema inference, dynamically typed at the value level — for agentic software: programs that reason and act autonomously by delegating inference to large language models (LLMs). Existing approaches augment general-purpose languages with frameworks, encoding critical invariants (bounded context, typed inference output, credential isolation, durable state) as application-level conventions rather than language guarantees. Turn introduces five language-level constructs that address this gap. […]

Ver mais

Introduction If you follow the message streaming space at all, you know it is dominated by a few big names. Apache Kafka has been the 800-pound gorilla since LinkedIn open-sourced it back in 2011. I’ve covered streaming-adjacent projects a few times in this series, like WarpStream, Apache Paimon, and Proton. So when I started seeing chatter about a message streaming platform written from scratch in Rust that was processing millions of messages per second, I had to take […]

Ver mais

arXiv:2603.23571v1 Announce Type: new Abstract: Effective navigation intelligence relies on long-term memory to support both immediate generalization and sustained adaptation. However, existing approaches face a dilemma: modular systems rely on explicit mapping but lack flexibility, while Transformer-based end-to-end models are constrained by fixed context windows, limiting persistent memory across extended interactions. We introduce StateLinFormer, a linear-attention navigation model trained with a stateful memory mechanism that preserves recurrent memory states across consecutive training segments instead of reinitializing them at […]

Ver mais

arXiv:2604.13108v1 Announce Type: new Abstract: AI coding agents spend a substantial fraction of their tool calls on undirected codebase exploration. We investigate whether providing agents with formal architecture descriptors can reduce this navigational overhead. We present three complementary studies. First, a controlled experiment (24 code localization tasks x 4 conditions, Claude Sonnet 4.6, temperature=0) demonstrates that architecture context reduces navigation steps by 33-44% (Wilcoxon p=0.009, Cohen’s d=0.92), with no significant format difference detected across S-expression, JSON, YAML, and […]

Ver mais

Large language models (LLMs) with reasoning abilities have demonstrated growing promise for tackling complex scientific problems. Yet such tasks are inherently domain-specific, unbounded and open-ended, demanding exploration across vast and flexible solution spaces. Existing approaches, whether purely learning-based or reliant on carefully designed workflows, often suffer from limited exploration efficiency and poor generalization. To overcome these challenges, we present HELIX — a Hierarchical Evolutionary reinforcement Learning framework with In-context eXperiences. HELIX introduces two key novelties: (i) a diverse […]

Ver mais

Two years back, shipping a SaaS product meant throwing a bunch of cash at devs, running out of money, and then desperately crossing your fingers that your first version wouldn’t fall over on launch day. Fast forward to now, and a solo founder with a subscription to Claude and a three-day weekend can actually get something really close to the same result. That’s genuinely great for builders, but defensibility is a different story. When everyone has access to […]

Ver mais

Centralized training is the standard paradigm in deep learning, enabling models to learn from a unified dataset in a single location. In such setup, isotropic feature distributions naturally arise as a mean to support well-structured and generalizable representations. In contrast, continual learning operates on streaming and non-stationary data, and trains models incrementally, inherently facing the well-known plasticity-stability dilemma. In such settings, learning dynamics tends to yield increasingly anisotropic feature space. This arises a fundamental question: should isotropy be […]

Ver mais

arXiv:2602.02543v1 Announce Type: new Abstract: Model editing has emerged as a practical approach for mitigating factual errors and outdated knowledge in large language models (LLMs). Among existing methods, the Locate-and-Edit (L&E) paradigm is the dominant framework: it locates MLP parameters implicated in expressing a target fact, and then performs a localized update to rewrite that fact. However, long sequences of edits often trigger abrupt model collapse in L&E beyond a critical point. We empirically identify a strong correlation […]

Ver mais

arXiv:2604.18697v1 Announce Type: new Abstract: Indistinguishability properties such as differential privacy bounds or low empirically measured membership inference are widely treated as proxies to show a model is sufficiently protected against broader memorization risks. However, we show that indistinguishability properties are neither sufficient nor necessary for preventing data extraction in LLM APIs. We formalize a privacy-game separation between extraction and indistinguishability-based privacy, showing that indistinguishability and inextractability are incomparable: upper-bounding distinguishability does not upper-bound extractability. To address this […]

Ver mais

arXiv:2602.06142v1 Announce Type: new Abstract: The phase ordering problem has been a long-standing challenge since the late 1970s, yet it remains an open problem due to having a vast optimization space and an unbounded nature, making it an open-ended problem without a finite solution, one can limit the scope by reducing the number and the length of optimizations. Traditionally, such locally optimized decisions are made by hand-coded algorithms tuned for a small number of benchmarks, often requiring significant […]

Ver mais