ChatGPT (OpenAI Conversational AI Product, 2022)

ChatGPT is the flagship consumer-product instance of the OpenAI Foundation Model Lab (machine:openai-foundation-model-lab-2015) — the DM-Day conversational AI product that compressed the adoption arc of AI from expert-tool to civilizational-scale discourse infrastructure. Launched November 30 2022 as a conversational fine-tune of InstructGPT (GPT-3.5 base), ChatGPT reached 100 million monthly active users by January 2023 — the fastest-ever consumer-product adoption rate (TikTok took ~9 months; Instagram ~2.5 years for the same milestone). Motor is flywheel (DM canonical): user prompts → RLHF + behavioral data → better fine-tuned model → more users → more usage → ChatGPT Plus ($20/mo) + API revenue → more training compute → better model. This differentiates ChatGPT from its parent class (motor=push): the user-acquisition flywheel is the instance-specific dynamic. Key milestones: GPT-3.5 launch Nov 30 2022; ChatGPT Plus subscription $20/month Feb 2023; GPT-4 March 14 2023; Microsoft $10B+ Bing/Copilot integration Jan 2023; Italian Garante temporary ban March 2023; GPT-4-turbo November 2023; GPT-4o May 2024 (multimodal voice); EU AI Act August 2024; GPT-5 family late 2024/2025; ~200M+ weekly active users end 2024. Competitive responses 2023-24: Anthropic Claude (launched 2023), Google Gemini (launched Feb 2024 as Bard→Gemini), Meta Llama-2 (July 2023), Mistral (Sept 2023). Legal pressure: NYT v. OpenAI (Dec 2023); Authors Guild v. OpenAI; image-generation rightsholders. Substrate: incorporeal (model weights + API + web interface) + cognitive (attention-capture; task-completion engine; epistemic dependency formation) + semiotic (conversational form structures knowledge-seeking behavior) + social (communal use norms; prompt-engineering culture). NOTE: [STUB-substrate-enum-gap] institutional substrate not in enum; ChatGPT's role as a quasi-institutional actor in higher-education and knowledge-work is recorded here. SCHEMA GAP: emergence_subtype for conversational-RLHF-alignment is narrower than the crowdsourced/meritocratic_hierarchy enum; recorded as custom subtype (rlhf_conversational_ alignment) pending v0.2 schema extension. V0.2 GAP: no enum slot for 'institutional' substrate (dialogue as institution); flywheel motor with RLHF feedback loop has no per-loop state variable in v0.1 schema. Intra-era couplings: parent class OpenAI-2015, sibling Claude-2024, sibling Gemini-2023, Microsoft Azure (compute dependency), Google DeepMind (competitor), Meta AI (competitor), Chinese AI Labs (competitor), arXiv (training corpus), Stack Overflow (knowledge-base scrape), x-twitter/X (cultural distribution), GitHub Copilot (adjacent product). Cross-era: parasitic_extraction → MM newspaper-broadcast (scraped news corpus); parasitic_ extraction → MM post-humboldtian university (PhD talent + research corpus); zombie_ dependency → MM industrial-era patent system (NYT lawsuit; copyright doctrine applied to LLM training data); hostile_inheritance → LM ontological-doubt-infrastructure (creates the need it will partly be replaced by); sublimation_coupling → MM Baconian scientific method (peer- reviewed publication pipeline rendered less critical for knowledge-retrieval by conversational AI); adapted_inheritance → LM capture-resistant-ai-infrastructure (OpenAI PBC conversion as partial hedge); sublimation_coupling → LM llm-public-good-coop-class (ChatGPT commercial lock-in model is the counterpoint that heretic-public-good entities organize against).