
rk
2.4K posts








I have so much gratitude to people who wrote extremely complex software character-by-character. It already feels difficult to remember how much effort it really took. Thank you for getting us to this point.








It's not exactly like that. 20 years took me to master all the knowledge from algebra, topology, algebraic geometry and number theory to be even able to enter the field of my main expertise: arithmetic algebraic geometry. When I look in hindsight, I can perfectly see now places where I've spent lots of hours chasing through the papers and sometimes coming back with nothing. Arithmetic geometry is an extremely rich blend of maths which had huge historical impact on our understanding of connections between questions about numbers and geometry. Basically all of modern cryptography is founded on such applications. From my perspective, many things I have mastered can now be easily accessed with models. Models can help you even figure out solutions to quite complex problems in those fields (especially when computer algebra systems are involved). I was doing a research in several areas of algebraic geometry and I can generate lots of questions as by-products of my research. So the problem the GPT 5.4 solved is one such problem, curated specifically for the purpose of the benchmark. But this problem was not the essence of my scientific life. To the contrary, I am really amazed that now in many situations such advanced models can go with me side by side in my mathematical investigations and accelerate the exploration. Yet, I find it extremely puzzling that an algorithm without mind (so non-human in this particular sense) can be on par with my top-tier colleagues and feel like a research partner.














Patriarchat jest faktem! Według danych @pewresearch pracujący mężowie w USA mają więcej czasu wolnego niż żony. Różnica rośnie u par z dziećmi (z 2 do 3 godz./tydz.), a z małymi dziećmi wynosi aż 4,5 godz./tydz. na korzyść ojców.









