Tag
A developer questions whether a high-RAM MacBook Pro for local AI models could match cloud frontier models like Claude for coding, considering long-term costs.
AI coding startup Cognition has acquired Poke, an AI assistant known for its friendly personality, in a deal valuing Poke in the low nine figures. The acquisition aims to bring Poke's interaction model to Cognition's coding assistant Devin, highlighting the growing value of AI personality.
OpenCode Superapp combines the power of Codex with locally hosted AI models and voice capabilities for coding.
pigo is a command-line AI Agent rewritten in Go, supporting headless script mode and interactive REPL, capable of reading/writing files, executing commands, searching code, and compatible with multiple large model protocols, accompanied by an e-book.
An Ars Technica article comparing two approaches to AI coding harnesses: the lean harness of Claude Code and the context-rich semantic retrieval approach of Augment Code, featuring an interview with Augment Code's VP of Engineering Vinay Perneti.
A personal review of using the Qwen 3.8 model for agentic coding, praising its ability to handle complex tasks like integrating llama.cpp with Godot while noting issues with looping.
Lightning Orchestrator is an agent skill that uses fast subagents (SWE-1.7 Lightning) for implementation, achieving ~5x faster execution while keeping a frontier model as planner and reviewer. It works with Devin and supports parallel execution for larger tasks.
Cline's agent harness is now available in JetBrains Air via ACP support, allowing users to power coding tasks directly in the app. A new Air release also adds Java language support and the ability to add various agents.
Type /graphify in your AI coding assistant to map your entire project into a knowledge graph for querying instead of grepping through files. The tool explains every edge in the graph, distinguishing source reads from resolved relationships.
Inkling is a large open-weights multimodal model (975B total, 41B active parameters) using a sparse MoE architecture, accepting text, image, and audio inputs and generating text outputs, intended for agentic systems, coding assistants, and chatbots.
Corbin Braun praises Fable 5's ability to autonomously implement complex plans and produce clean code, highlighting its practical utility as an AI coding agent.
Speculative cache warming pre-processes the system prompt and tools array while the user types their prompt, saving 10-20 seconds of wait time on local LLM inference. This feature is part of the open-source OpenFox harness for local AI, improving interactivity without breaking cache consistency.
A user seeks advice on choosing between a modded RTX 4090 48GB, dual AMD Radeon AI Pro R9700, or dual Intel Arc Pro B70 for running local coding LLMs, highlighting trade-offs in price, VRAM, software ecosystem, and inference speed.
A user compares the Codex and Opencode AI coding tools, questioning whether to switch given their local setup with Qwen 3.6 27b Q8.
A commentator notes that AI is in a messy middle phase where usage appears productive but costs are high, citing a Forbes article that AI costs more than the workers it replaced, with examples from Uber and Microsoft.
Laguna M.1, a 225B parameter model with 256k context for agentic coding, is now available for free in Cline.
A user reports that Fable 5's new classifiers misrouted 75% of a coding session to Opus, flagging routine coding as a cybersecurity risk and causing unexpectedly high costs.
Claude Fable 5 is now available in Cursor, topping the CursorBench leaderboard but with the highest cost per task.
A user shares positive impressions of the GLM-5.2 model, noting its impressive performance and cost-efficiency compared to Deepseek, and reflects on the educational value and limitations of AI agents for coding and learning.
Claude Sonnet 5 is now generally available and rolling out in GitHub Copilot. Early testing shows strong coding performance, especially on CLI tasks, with good prompt-cache utilization and competitive latency.