News of the day
1. OpenAI launches GPT-5.2, featuring advanced reasoning, coding, and agentic capabilities with a massive context window, aiming to lead professional AI applications. → Read more
2. Google launches managed MCP servers, simplifying AI agent integration with services like Maps and BigQuery. This move aims to make Google's ecosystem 'agent-ready by design'. → Read more
3. Disney invests $1B in OpenAI, licensing 200+ characters for Sora AI video app. Aims to responsibly extend storytelling via generative AI. → Read more
4. Shane Legg, co-founder of Google DeepMind, discusses AGI's potential impact on society, defining it as an agent capable of human-like cognitive tasks and highlighting the challenges in achieving it. → Read more
Our take
Hi Dotikers!
OpenAI just released GPT-5.2, and let's be honest: this is a survival update, not a revolution. Announced on December 11th amid an internal "code red" triggered by Sam Altman following Gemini 3's stunning launch, this new model comes in three versions (Instant, Thinking, Pro) and promises 38% fewer errors with improved reasoning performance. On paper, it's solid. In market reality, it's catch-up.
Our reading is unambiguous: OpenAI has lost its comfortable lead. The numbers speak for themselves. Anthropic now captures 40% of enterprise LLM spending versus 27% for OpenAI, a spectacular reversal from 2023 when those positions were exactly flipped. Claude Opus 4.5 dominates coding with 77% on SWE-bench Verified. Gemini 3 was the first to break the 1500 Elo barrier on LMArena. King ChatGPT finds itself surrounded.
What strikes us is the pricing: GPT-5.2 costs more than GPT-5 ($1.75 vs $1.25 per input). OpenAI is betting that quality will justify the premium. A risky bet when Meta offers Llama 4 at a fraction of the price and CIOs are scrutinizing every budget line.
GPT-5.2 is an excellent model that consolidates gains without creating disruption. For enterprises already invested in the OpenAI ecosystem, it's a welcome improvement. For those still on the fence, the market has never been more open. The model wars are entering their most intense phase, and for the first time since GPT-4, OpenAI is no longer the obvious favorite.
M.
