ChatGPT: Everything you need to know about the AI-powered chatbot (techcrunch.com)

ChatGPT has evolved from a productivity text bot (launched Nov 2022) into one of the world’s dominant AI platforms, growing from ~300M weekly users to nearly 700M and handling ~2.5 billion prompts daily. Since 2024 OpenAI shipped major capabilities (voice-enabled GPT-4o, text-to-video Sora), struck a partnership with Apple, and in 2025 released GPT-5 plus specialized variants like GPT-5-Codex — a coding agent that dynamically allocates compute time (seconds to hours) to solve complex bugs and refactors — and new product features such as Agents, Study Mode, and expanded enterprise deals (including a $1 federal offering). The app has also generated outsized consumer revenue (~$2B), underscoring broad commercial traction even as OpenAI navigates executive departures, lawsuits over training data, and high-profile safety concerns tied to mental-health use. Technically the landscape is shifting: OpenAI now ships multiple concurrent models (GPT-5 with Auto/Fast/Thinking modes alongside legacy GPT-4o/4.1), has returned to open weights with gpt-oss-120b and gpt-oss-20b (designed to run on a single Nvidia GPU or laptops), and for the first time uses non‑Nvidia AI chips (Google’s) in production—signaling supply‑chain and infrastructure diversification. For practitioners this means richer model choice, more on‑device and open options, new agent APIs for end‑to‑end workflows, and increased attention to safety, regulation, and compute economics as OpenAI scales data centers and pursues massive fundraising amid intensifying global competition.