Full-text search across 66214 articles. Combine with topic and date filters; results sorted by relevance.
OpenAI published more than 700 math research papers generated by an unreleased model, with almost every result produced from a single prompt, according to the lab. The results cover previously unsolved problems, includin…
Meta's Muse Wonder Full scored higher than standard Wonder on three of six KingBench 4 coding builds — the emulator, keyboard configurator, and Pokémon-style game — but tied or scored lower on the other three, and landed…
A developer tested Bruno, Apidog, Insomnia, Elva, and Postman on the same API workflow — adding an optional tax_id field to a POST /invoices endpoint, sharing the updated request for review, and exposing the endpoint as …
A new benchmark called sudo L7, built by an unnamed team, measures staff-level software engineering judgment across 60 tasks authored by professional software engineers, most drawn from private production repositories at…
A developer released IronBee Gamer, an open-source tool that plays browser games by combining a 30ms local decision model, Laya, with an LLM trainer that writes each game's feature-extraction logic and action rules. The …
A developer workflow combining Vercel's v0.dev for AI-generated React frontends and Cursor IDE for AI-assisted backend implementation is proposed to speed up proof-of-concept development. The frontend-first approach uses…
Reflection introduced Beam, its first open-weight model, a sparse Mixture-of-Experts system with 501 billion total parameters and 23 billion active parameters built for coding, reasoning, and agentic workloads. Reflectio…
Artificial Analysis's Intelligence Index version 4.1.1 weights agentic workloads at 34% and general reasoning at 18%, a shift an analysis of 586 model evaluations argues distorts what the score measures. The analysis fou…
An agent harness is the runtime layer around a model — comprising a tool registry, model, context management, guardrails, an agent loop and a verify step — that makes an agent reliable, according to a technical explainer…
Anthropic's Claude Design, currently in beta, enables users to create on-brand slide decks by first using Claude Code to convert existing decks into a reusable design system, then generating editable slides in Google Sli…
A developer built a caching mechanism for Claude Code hooks to prevent performance degradation when checking token usage. The solution, cost_guard.sh, caches ccusage results for 60 seconds, throttling API calls to once p…
Spec-driven development spans three distinct levels, from spec-first to spec-as-source, according to Birgitta Böckeler of Thoughtworks, writing on martinfowler.com. The AI Unified Process, a methodology that keeps requir…
A developer's audit of migrating Claude Code skills to OpenCode reveals that skills often degrade silently, with harness-specific frontmatter fields such as allowed-tools and model pinning being dropped without warning. …
A developer's cost-monitoring script for Claude Code silently failed twice, once overcounting weekly spend by 4-6x and once going quiet due to set -e, leading to a redesign that prioritizes fail-open behavior so the dash…
Technical leaders should have the largest AI exhaust, according to an essay arguing that senior engineers must personally use coding agents to answer empirical questions about AI-assisted development. The piece, referenc…
A developer argues that AI will not replace software developers but will elevate them to a higher level of abstraction, shifting their role from writing code to defining intent, orchestrating systems, and owning outcomes…
Google Cloud's Database Migration Service now includes Gemini-powered AI code conversion to translate stored procedures, triggers, and functions from Oracle PL/SQL and SQL Server T-SQL into PostgreSQL PL/pgSQL, addressin…
A Substack essay by an author exploring formal verification argues that AI coding tools enable replacing elite programming competitions like the International Olympiad in Informatics (IOI) with contests that create usefu…
The traditional Plan-Build-Run model for IT organizations is being replaced by a four-phase Specify-Generate-Verify-Operate model due to the rise of AI agents and spec-driven development, according to an industry practit…
Prime Intellect's Prime Agent, a coding harness that replaces traditional tool menus with a live IPython kernel, can be run locally with DeepSeek V4 on NVIDIA's DGX Spark, and reportedly pushed Claude Opus to roughly 95.…