NO/FOMO

Independent AI signal, once a day

The AI briefing worth opening.

ISSUE DATE2025-12-30DEFAULT EDITION
This issue
—
All time
—

Hacker News

5 stories
01

Show HN: Replacing my OS process scheduler with an LLM

This Hacker News 'Show HN' entry introduces an innovative project exploring the replacement of traditional operating system process schedulers with a Large Language Model (LLM). The 'brainkernel' initiative aims to leverage the advanced decision-making and pattern recognition capabilities of LLMs to manage system resources and process execution. This novel approach pushes the boundaries of LLM applications beyond conventional language tasks, proposing a shift towards AI-driven core system functionalities. The project represents an early-stage exploration into how sophisticated AI models could potentially optimize and intelligently control fundamental operating system operations, presenting a unique intersection of AI and systems engineering.

OS Process SchedulingLarge Language ModelOperating SystemsAI for SystemsResource ManagementArtificial IntelligenceLarge Language ModelAI Agent
02

Show HN: Brainrot Translator – Convert corporate speak to Gen Alpha and back

The "Brainrot Translator," recently showcased on Hacker News, is an innovative AI-powered web application developed to address the widening generational communication gap online. Built as an LLM-wrapper, its core function is the bidirectional translation of language between "Boomer" (representing standard or corporate English) and "Brainrot" (embodying Gen Alpha slang). This utility offers users a unique and often humorous way to decipher the evolving lexicon used by younger demographics, making it useful for both entertainment and practical understanding. Beyond linguistic translation, the application integrates an "Image-to-Rot" feature. This component employs computer vision capabilities to analyze user-uploaded images and subsequently generate descriptions of their content using the distinctive Gen Alpha slang. Positioned as a tool for fun, it also demonstrates the practical applicability of large language models and multimodal AI in interpreting complex cultural and linguistic nuances, ultimately fostering better cross-generational comprehension in the digital age.

Large Language ModelNatural Language ProcessingAI TranslationSlang TranslationComputer VisionLarge Language ModelNatural Language ProcessingComputer Vision
03

Non-Zero-Sum Games

The 'Non-Zero-Sum Games' initiative or platform appears to center on the fundamental concept of non-zero-sum interactions, a cornerstone of game theory. Unlike zero-sum scenarios where one party's gain precisely equals another's loss, non-zero-sum games allow for mutual benefit or mutual detriment among participants. This framework is crucial for understanding cooperative strategies, conflict resolution, and the dynamics of complex systems where participants can achieve win-win outcomes or face collective losses. While specific details of the platform's content are not provided in the prompt, the title suggests a focus on exploring these theoretical underpinnings and their practical applications. This could encompass areas such as multi-agent AI systems, economic modeling, social dynamics, or decentralized network architectures, where understanding shared incentives and interdependencies is paramount for designing robust and beneficial interactions. The resource likely aims to elucidate these concepts for a technical audience.

Game TheoryMulti-Agent SystemsCooperative AIStrategic InteractionsSystem DesignArtificial IntelligenceAI AgentMachine Learning
04

The 70% AI productivity myth: why most companies aren't seeing the gains

The widely circulated claim of 70% productivity gains from Artificial Intelligence within most companies is being challenged as a pervasive myth. This narrative suggests that despite the hype and significant investments in AI technologies, a majority of organizations are not realizing the anticipated substantial improvements in efficiency and output. The article likely delves into the underlying reasons for this discrepancy, which may include challenges in integrating AI solutions into existing workflows, a lack of strategic foresight in adoption, skill gaps among the workforce, and unrealistic expectations regarding the immediate impact of AI. It implies that successful AI implementation requires more than just deploying technology; it demands comprehensive organizational change, robust data infrastructure, and a clear understanding of specific business problems AI can solve. The piece aims to provide a more grounded perspective on AI's true impact on enterprise productivity, moving beyond aspirational figures to address the practical realities and hurdles companies face in leveraging AI effectively for tangible benefits.

AI ProductivityEnterprise AIAI AdoptionBusiness StrategyProductivity GainsArtificial IntelligenceOthersMachine Learning
05

Show HN: 22 GB of Hacker News in SQLite

A new project, 'Hacker News in SQLite', has been unveiled, offering a massive 22 GB dataset comprising historical content from the popular technology news aggregator, Hacker News. This significant compilation stores a vast archive of stories, comments, and user interactions within a single, portable SQLite database file. The initiative aims to provide researchers, data scientists, and developers with an easily accessible and queryable resource for in-depth analysis of trends, discussions, and community dynamics on the platform. By consolidating years of data into a structured SQLite format, the project dramatically simplifies the process of data extraction and analysis, bypassing the complexities of API limitations or web scraping. This ready-to-use dataset is expected to facilitate diverse analytical endeavors, from understanding the evolution of technology topics to studying user engagement patterns and AI/ML applications.

Hacker NewsSQLiteData ArchivingDatasetData AnalysisNatural Language ProcessingMachine LearningArtificial Intelligence

huggingface

6 stories
01

Coupling Experts and Routers in Mixture-of-Experts via an Auxiliary Loss

Mixture-of-Experts (MoE) models lack explicit constraints to ensure the router's decisions align well with the experts' capabilities, which ultimately limits model performance. To address this, we propose expert-router coupling (ERC) loss, a lightweight auxiliary loss that tightly couples the router's decisions with expert capabilities. Our approach treats each expert's router embedding as a proxy token for the tokens assigned to that expert, and feeds perturbed router embeddings through the experts to obtain internal activations. The ERC loss enforces two constraints on these activations: (1) Each expert must exhibit higher activation for its own proxy token than for the proxy tokens of any other expert. (2) Each proxy token must elicit stronger activation from its corresponding expert than from any other expert. These constraints jointly ensure that each router embedding faithfully represents its corresponding expert's capability, while each expert specializes in processing the tokens actually routed to it. The ERC loss is computationally efficient, operating only on n^2 activations, where n is the number of experts. This represents a fixed cost independent of batch size, unlike prior coupling methods that scale with the number of tokens (often millions per batch). Through pre-training MoE-LLMs ranging from 3B to 15B parameters and extensive analysis on trillions of tokens, we demonstrate the effectiveness of the ERC loss. Moreover, the ERC loss offers flexible control and quantitative tracking of expert specialization levels during training, providing valuable insights into MoEs.

Mixture-of-ExpertsAuxiliary LossExpert-Router CouplingLarge Language ModelsExpert SpecializationMachine LearningDeep LearningLarge Language Model
02

LiveTalk: Real-Time Multimodal Interactive Video Diffusion via Improved On-Policy Distillation

Real-time video generation via diffusion is essential for building general-purpose multimodal interactive AI systems. However, the simultaneous denoising of all video frames with bidirectional attention via an iterative process in diffusion models prevents real-time interaction. While existing distillation methods can make the model autoregressive and reduce sampling steps to mitigate this, they focus primarily on text-to-video generation, leaving the human-AI interaction unnatural and less efficient. This paper targets real-time interactive video diffusion conditioned on a multimodal context, including text, image, and audio, to bridge the gap. Given the observation that the leading on-policy distillation approach Self Forcing encounters challenges (visual artifacts like flickering, black frames, and quality degradation) with multimodal conditioning, we investigate an improved distillation recipe with emphasis on the quality of condition inputs as well as the initialization and schedule for the on-policy optimization. On benchmarks for multimodal-conditioned (audio, image, and text) avatar video generation including HDTF, AVSpeech, and CelebV-HQ, our distilled model matches the visual quality of the full-step, bidirectional baselines of similar or larger size with 20x less inference cost and latency. Further, we integrate our model with audio language models and long-form video inference technique Anchor-Heavy Identity Sinks to build LiveTalk, a real-time multimodal interactive avatar system. System-level evaluation on our curated multi-turn interaction benchmark shows LiveTalk outperforms state-of-the-art models (Sora2, Veo3) in multi-turn video coherence and content quality, while reducing response latency from 1 to 2 minutes to real-time generation, enabling seamless human-AI multimodal interaction.

Real-Time Video DiffusionMultimodal InteractionOn-Policy DistillationGenerative AIAvatar SystemGenerative AIMultimodalAI Agent
03

SmartSnap: Proactive Evidence Seeking for Self-Verifying Agents

Agentic reinforcement learning (RL) holds great promise for the development of autonomous agents under complex GUI tasks, but its scalability remains severely hampered by the verification of task completion. Existing task verification is treated as a passive, post-hoc process: a verifier (i.e., rule-based scoring script, reward or critic model, and LLM-as-a-Judge) analyzes the agent's entire interaction trajectory to determine if the agent succeeds. Such processing of verbose context that contains irrelevant, noisy history poses challenges to the verification protocols and therefore leads to prohibitive cost and low reliability. To overcome this bottleneck, we propose SmartSnap, a paradigm shift from this passive, post-hoc verification to proactive, in-situ self-verification by the agent itself. We introduce the Self-Verifying Agent, a new type of agent designed with dual missions: to not only complete a task but also to prove its accomplishment with curated snapshot evidences. Guided by our proposed 3C Principles (Completeness, Conciseness, and Creativity), the agent leverages its accessibility to the online environment to perform self-verification on a minimal, decisive set of snapshots. Such evidences are provided as the sole materials for a general LLM-as-a-Judge verifier to determine their validity and relevance. Experiments on mobile tasks across model families and scales demonstrate that our SmartSnap paradigm allows training LLM-driven agents in a scalable manner, bringing performance gains up to 26.08% and 16.66% respectively to 8B and 30B models. The synergizing between solution finding and evidence seeking facilitates the cultivation of efficient, self-verifying agents with competitive performance against DeepSeek V3.1 and Qwen3-235B-A22B.

Self-Verifying AgentsProactive Evidence SeekingTask VerificationLLM-driven AgentsReinforcement LearningAI AgentLarge Language ModelMachine Learning
04

Monadic Context Engineering

The proliferation of Large Language Models (LLMs) has catalyzed a shift towards autonomous agents capable of complex reasoning and tool use. However, current agent architectures are frequently constructed using imperative, ad hoc patterns. This results in brittle systems plagued by difficulties in state management, error handling, and concurrency. This paper introduces Monadic Context Engineering (MCE), a novel architectural paradigm leveraging the algebraic structures of Functors, Applicative Functors, and Monads to provide a formal foundation for agent design. MCE treats agent workflows as computational contexts where cross-cutting concerns, such as state propagation, short-circuiting error handling, and asynchronous execution, are managed intrinsically by the algebraic properties of the abstraction. We demonstrate how Monads enable robust sequential composition, how Applicatives provide a principled structure for parallel execution, and crucially, how Monad Transformers allow for the systematic composition of these capabilities. This layered approach enables developers to construct complex, resilient, and efficient AI agents from simple, independently verifiable components. We further extend this framework to describe Meta-Agents, which leverage MCE for generative orchestration, dynamically creating and managing sub-agent workflows through metaprogramming. Project Page: https://github.com/yifanzhang-pro/monadic-context-engineering.

Monadic Context EngineeringAI AgentsLarge Language ModelsMonadsMeta-AgentsAI AgentLarge Language ModelArtificial Intelligence
05

Robo-Dopamine: General Process Reward Modeling for High-Precision Robotic Manipulation

The primary obstacle for applying reinforcement learning (RL) to real-world robotics is the design of effective reward functions. While recently learning-based Process Reward Models (PRMs) are a promising direction, they are often hindered by two fundamental limitations: their reward models lack step-aware understanding and rely on single-view perception, leading to unreliable assessments of fine-grained manipulation progress; and their reward shaping procedures are theoretically unsound, often inducing a semantic trap that misguides policy optimization. To address these, we introduce Dopamine-Reward, a novel reward modeling method for learning a general-purpose, step-aware process reward model from multi-view inputs. At its core is our General Reward Model (GRM), trained on a vast 3,400+ hour dataset, which leverages Step-wise Reward Discretization for structural understanding and Multi-Perspective Reward Fusion to overcome perceptual limitations. Building upon Dopamine-Reward, we propose Dopamine-RL, a robust policy learning framework that employs a theoretically-sound Policy-Invariant Reward Shaping method, which enables the agent to leverage dense rewards for efficient self-improvement without altering the optimal policy, thereby fundamentally avoiding the semantic trap. Extensive experiments across diverse simulated and real-world tasks validate our approach. GRM achieves state-of-the-art accuracy in reward assessment, and Dopamine-RL built on GRM significantly improves policy learning efficiency. For instance, after GRM is adapted to a new task in a one-shot manner from a single expert trajectory, the resulting reward model enables Dopamine-RL to improve the policy from near-zero to 95% success with only 150 online rollouts (approximately 1 hour of real robot interaction), while retaining strong generalization across tasks. Project website: https://robo-dopamine.github.io

RoboticsReinforcement LearningReward ModelingRobotic ManipulationMulti-view PerceptionRoboticsMachine LearningDeep Learning
06

Self-Evaluation Unlocks Any-Step Text-to-Image Generation

We introduce the Self-Evaluating Model (Self-E), a novel, from-scratch training approach for text-to-image generation that supports any-step inference. Self-E learns from data similarly to a Flow Matching model, while simultaneously employing a novel self-evaluation mechanism: it evaluates its own generated samples using its current score estimates, effectively serving as a dynamic self-teacher. Unlike traditional diffusion or flow models, it does not rely solely on local supervision, which typically necessitates many inference steps. Unlike distillation-based approaches, it does not require a pretrained teacher. This combination of instantaneous local learning and self-driven global matching bridges the gap between the two paradigms, enabling the training of a high-quality text-to-image model from scratch that excels even at very low step counts. Extensive experiments on large-scale text-to-image benchmarks show that Self-E not only excels in few-step generation, but is also competitive with state-of-the-art Flow Matching models at 50 steps. We further find that its performance improves monotonically as inference steps increase, enabling both ultra-fast few-step generation and high-quality long-trajectory sampling within a single unified model. To our knowledge, Self-E is the first from-scratch, any-step text-to-image model, offering a unified framework for efficient and scalable generation.

Text-to-Image GenerationSelf-Evaluating ModelAny-Step InferenceFlow MatchingFrom-Scratch TrainingMultimodalGenerative AIDeep Learning