Pavel Sokolovsky

better doing

better doing · Nov 2026

The best ideas are just lying around.

“Don’t worry about people stealing your ideas. If your ideas are any good, you’ll have to ram them down people’s throats.” The piece argues that most “original” ideas are either rediscoveries or untested dead ends, and that borrowing proven memes can feel like pushing the boulder downhill. ~ learn more

better doing · Nov 2026

The Mythical Man-Month in the age of Al.

Brooks’ Law says adding manpower to a late software project makes it later. “Adding agents to a project with unclear requirements gets you the wrong answer faster.” A neat take on why AI doesn’t magically remove coordination overhead, it just moves it onto you, the human who has to define “done” and catch hallucinations. ~ learn more

better doing · Nov 2026

Find role models, not mentors.

“I don’t have mentors, I have role models.” Hunter Walk argues mentorship has gotten “self-proclaimed, over-used and diluted,” while role models scale and do not even need to know they are helping you. ~ learn more

better doing · Nov 2026

Everything is fertile.

“In Miyazaki’s eyes there is nothing sterile about crafted nature because man cannot help but breed fertility.” It’s a shift from chasing novelty to upgrading your perception, until even a planned park or a woodpile starts to look enchanted. ~ learn more

better doing · Oct 2026

Most advice is right for the person giving it.

“Two smart people can tell you opposite things, and both can make perfect sense.” The trick is to stop copying the rule and ask what was true about their situation, then check if it’s true for you too. ~ learn more

better doing · Oct 2026

The courage to become who you already are.

This piece quotes Joseph Campbell’s Hero’s Journey, which I recently read: “We must be willing to get rid of the life we’ve planned, so as to have the life that is waiting for us. The old skin has to be shed before the new one can come.” ~ learn more

better doing · Oct 2026

How to engineer a soul.

“We did not stop having religion. We unbundled it.” Pande’s point is that we kept the parts (therapy, wellness, politics, social media) but lost the integrated system that actually forms character, so nobody is responsible for who you become. ~ learn more

better doing · Oct 2026

Better feedback with less threat.

“The brain is a threat and reward machine.” If you want people to stay open and think clearly, the talk argues you should give feedback in ways that reduce social threat and increase a sense of reward. ~ learn more

better doing · Oct 2026

I love being wrong.

“All this to say, I love being wrong. It is my favorite, because it helps me more clearly define the problem, find the right solution, learn more quickly, and solve the problem.” A simple six-step loop for product work: define, act, update your priors when new data shows up. ~ learn more

better doing · Oct 2026

Andrej Karpathy’s latest LLM tip.

Karpathy suggests forcing clearer formats like ASD-STE100 controlled language, diagrams, HTML pages, and even custom explainer videos to make model thinking easier to parse. I told my Claude to try these on stuff we’ve covered together to see for myself, and it’s quite good. ~ learn more