[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

[@ifeelbig](/creator/twitter/ifeelbig)
"@dearmadisonblue But if what we really needed was more raw analytical intelligence there'd be a huge run on say math majors. We'd be plucking every promising math student out of e.g. India and investing millions in training them. We don't do anything like that. Why not"  
[X Link](https://x.com/ifeelbig/status/1979302715877380346) [@ifeelbig](/creator/x/ifeelbig) 2025-10-17T21:45Z XXX followers, XX engagements


"@dearmadisonblue Because most scientific and technological breakthroughs are not the result of someone thinking really hard about something. Rather they require extremely costly (both in terms of time and money) investments in experiments situated in the real world. That and social capital"  
[X Link](https://x.com/ifeelbig/status/1979303145500217684) [@ifeelbig](/creator/x/ifeelbig) 2025-10-17T21:46Z XXX followers, XX engagements


"@dearmadisonblue I think the main thing people miss is that the subset of valuable tasks for which "intelligence" is the gating factor is actually quite small. It's not zero and as AI gets better (even accounting for the limitations of the LLM architecture) there will be some major developments"  
[X Link](https://x.com/ifeelbig/status/1979302171272225210) [@ifeelbig](/creator/x/ifeelbig) 2025-10-17T21:43Z XXX followers, XXX engagements


"@wispem_wantex @AndreTI Disagree - combination of reasonably consistent ability to process web pages and TTC-driven planning made the models from o3 on much more useful. They're often better than Google now for basic questions. In terms of actual analytical thinking very marginal gains since X agree"  
[X Link](https://x.com/ifeelbig/status/1979658951005462850) [@ifeelbig](/creator/x/ifeelbig) 2025-10-18T21:20Z XXX followers, XX engagements