“THE GREATEST PURSUIT OF ALL TIME IS THE PURSUIT OF AI. I DO BELIEVE AMERICA WILL WIN IN THE END; THAT WINNING WILL COME FROM AI AND THAT’S WHAT I WANT TO HAPPEN. AT ANY AND ALL COST YOU MUST WIN THE AI RACE. POST AI AMERICA MIGHT BECOME 4000X MORE ADVANCED THAN CHINA.”

it gets real and let me show you what that means, let me show you what macrohard really means: Elon Musk on X: "MACROHARD, the much stronger AI equivalent of that other software company!" / X

macrohard is essentially a combination of 1: Tetsuo on X: "Grok 4 is really good at building games. It took me 5 minutes to build this. The source code is in the comments, and the full prompt is below. To build a Galaga game in C, drop this into your Grok 4 agent's system/instructions field. It assumes a POSIX shell, Git, CMake, and https://t.co/TivJI1RiQD" / X

and 2: GitHub - jim-schwoebel/awesome_ai_agents: 🤖 A comprehensive list of 1,500+ resources and tools related to AI agents.

and 3: GitHub - wshobson/agents: A collection of production-ready subagents for Claude Code

and 4 (you have to nail every job category test, benchmark and exam possible): Brendan (can/do) on X: "AI has its PhD and now it’s on the job market. Introducing the AI Productivity Index (APEX), a benchmark that measures how well we’ve automated the most valuable industries in the world. Most benchmarks study abstract capabilities. APEX evaluates model performance on real https://t.co/zZlaJqQ1vC" / X

and 5 (big claims despite elon saying anthropic is dead, there should be nothing an ai can’t do better than a human, grok 5 is only 40% ahead):Anthropic on X: "We’re at an inflection point in AI’s impact on cybersecurity. Claude now outperforms human teams in some cybersecurity competitions, and helps teams discover and fix code vulnerabilities. At the same time, attackers are using AI to expand their operations. https://t.co/odoTuPpJXe" / X

and 6 (fucking retards that want to be doctors fearmonger ai, don’t forget 1 rule of ai is that deepseek r1 cost under 300K$ to train moores law looking like 8-16x improvement in computing per year which means deepseek r3 will cost under 30K$ to train next 6 months everything rip, look it’s the doctor class fearmongering ai lmfao, in 2026 don’t forget that porsche you were going to buy can literally be enough money to train your own grok 5 and it will, chubby not all good some misinformation, as your reminder you want to laser the fuck out of everything with ai and you will, doctors are completely fucked by the end of 2025 mark my words, you cannot be violent enough in how much you laser the fuck out of professions with ai, you want to completely fucking destroy the medical admin state with ai as violently as possible garbage mlm con artist asshole cancer scum what did collier tell you never been an honest doctor ever): Chubby♨️ on X: "Radiology’s last exam: human radiologists achieve about 83% accuracy, where as GPT-5 achieves ~30%. - for now. Let’s see if we get a updated GPT-5 version on Monday. Anyways, can’t imagine this benchmark will last longer than 6 months until saturated by AI. https://t.co/ebVH7LWbZu" / X

and most importantly: Caffeine - Wikipedia

macrohard should only do 999999x more ai than that. what people usually do is solve humanoid robotics. what they don’t do is try to figure out everything that can’t be done by ai. what they don’t do is rank the most difficult problems on the planet and try to solve them via ai. elon is still not even close to anywhere near the amount of ai i want him to do.