Tag
This article dissects the House of Apple 2 exploitation technique on modern glibc versions, providing an interactive GDB walkthrough and analyzing its viability with a sandbox environment.
Raymond Chen explains why the Windows API ReadDirectoryChangesW doesn't provide correlation for rename operations, due to an unwritten assumption that rename events are sequential, which may not hold with concurrent activities.
An analysis of the 2048 game played on a 3x3 board, exploring algorithms and strategies for the smaller grid.
The article examines the issues with MMLU benchmark scores, showing that different model builds can yield incomparable accuracies due to open evaluation variables, and proposes a content-addressed framework for better traceability.
Indicator Go is a Go library that provides 80+ technical analysis indicators, a backtesting framework, and integrates with AI tools via MCP for algorithmic trading.
The Ornith 1.5 35B A3B model's MTP tensors appear to be uninitialized, causing poor speculative decoding performance, and grafting the trained head from Qwen3.6-35B-A3B improves speed by 29%.
This article explains how reasoning traces work in AI models, discussing their implementation, concealment, and extraction techniques, with examples from models like GPT-OSS and DeepSeek.
This article provides a detailed introduction to Pi's coding agent's context compression mechanism, analyzing the reasons for context expansion, compression trigger conditions, and the cost of cache invalidation, aiming to optimize AI performance in long conversations.
The article explains how RuneScape in 2004 optimized data transfer to fit a multiplayer RPG on 56k dial-up modems, detailing network constraints, encryption, and client decompilation.
Su Jianlin reviews the K3 architecture, focusing on the combination of KDA + MLA + Stable LatentMoE + AttnRes. He explains the design trade-offs, MoE stability improvements, why MLA was kept, and the relationship between DSV4 and MLA.
A detailed analysis of the implementation of the physics system in Sonic games: it runs efficiently on a 7MHz 68000 CPU based on predefined collision tiles and simple sensor rules, rather than real-time physics calculations. It also introduces design details such as level collision data, sensor system, dual loop implementation, and the author's progress on an Amiga port.
Introduces the open-source Python tool InStock (myhhub/stock) for A-share quantitative analysis, supporting data scraping, technical indicators, stock selection, backtesting, automated trading, etc., and can be deployed via Docker.
This paper evaluates the ability of large language models to perform technical market analysis for trading, using a benchmark approach.
This article provides an in-depth analysis of WorkBuddy's six-layer technical architecture, pointing out that it is essentially an Agent operating system rather than a single model, and demonstrates real failures and fixes in Agent engineering through update logs.
Introduces patternfetch, an MCP server + REST API that provides a token-compact market-state brief instead of raw OHLCV data for trading agents, reducing token usage and hallucinations.
This article, based on the sharing of researcher Victoria Lin, systematically reviews the mainstream technical approaches of native multimodal large models (Chameleon, Transfusion, MOT) and their pros and cons. It points out that multimodal AI is still in the early exploration stage, with open problems such as gaps in scaling laws, inconsistency between image understanding and generation encoding, and connection with the physical world.
A blog post reveals that Slay the Spire 2's random number generators are correlated due to linearity in C#'s System.Random, allowing players to predict in-game outcomes. The post details how different RNGs initialized with seed+hash produce exploitable correlations, affecting various game events.
An in-depth analysis of ChatGPT Dreaming V3's memory architecture, explaining how it synthesizes a coherent memory state from raw sources and comparing it to other open-source memory frameworks like mem0, supermemory, and Letta.
An in-depth technical analysis of the FMV compression and playback techniques used in the Sega CD game Silpheed, including reverse engineering of its format.
This paper introduces FINESSE-Bench, a suite of eight specialized benchmarks with 3,993 questions for hierarchical evaluation of financial competencies in large language models, covering professional certification topics and applied trading tasks.