The AI Race Just Changed Direction
TL;DR
China is redesigning education for an AI economy: Universities have cut over 12,000 degree programs, favoring AI and STEM while reducing arts and language tracks to align with national tech goals.
MIT's object-memory robot points to a practical AI holy grail: The DAAAM system lets robots answer questions like "where is the part we started assembling last night" by tying 3D vision to time, place, and natural language context.
EA wants to build an ad marketplace inside games: Electronic Arts says its new in-game ad system can reach 120 million monthly players through stadium signs, billboards, branded items, and live challenges.
AI safety is shifting from abstract fear to geopolitics: Dylan highlights arguments from The Economist and Rob Mason that the US and China need verifiable AI red lines and inspections before recursive self-improvement gets out ahead of governance.
Humans still beat AI on a serious math test: In the First Proof benchmark, ChatGPT 5.5 Pro solved 6 of 10 unpublished research-level problems, strong enough to help but not enough to replace top mathematicians.
The coding race may be the real center of gravity now: Dylan argues Anthropic saw earlier than rivals that code is the main flywheel, while OpenAI faces pressure from Claude, Gemini, Meta, Chinese models, and the strategic value of tools like Cursor.
The Breakdown
China just cut more than 12,000 university degree programs to pivot toward AI, while ChatGPT fell below 50% market share for the first time and global opinion started tilting toward China as the country leading the AI race. Dylan Curious connects those shifts to a bigger theme: AI is no longer just a model contest, it is becoming a fight over education, infrastructure, coding, safety, and who gets to capture the wealth.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.
Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Playbook
The Art of Tasteful Prompting
Learn how tasteful prompting helps you move beyond generic AI output by shaping context, style, and judgment from the start.