WTF Happened in 2025? (wtfhappened2025.com)

OpenAI's recent release of GDPVal reveals that frontier AI models are nearing 50% parity with human expert performance across various domains, including software development. The advancement is highlighted by the emergence of GPT 5.2 Pro, achieving an impressive 74.1% human performance benchmark. This trend reflects a significant shift in the capabilities of AI models, empowering developers like Pete Steinberger, who created OpenClaw using Codex, leading to surpassing established open-source projects. Additionally, Anthropic's Claude Code has rapidly grown, generating $2.5 billion in annual recurring revenue within its first year, with forecasts suggesting it may account for up to 50% of public commits on GitHub by year-end. On the technical front, Google DeepMind has made strides with its Gemini model, achieving gold certification at the International Math Olympiad through a purely token-based approach. This success indicates a new frontier in AI's ability to tackle complex mathematical tasks more efficiently than prior models. Furthermore, Cloudflare's rapid development of Next.js demonstrates a significant leap in engineering productivity, indicating that AI tools are becoming integral to accelerating software creation. An industry report also anticipates a pivotal change in the reliability of AI agents by 2026, alongside potential job displacement, highlighting both the advancements and challenges facing the AI/ML community.