Brennan's Conversations

Brennan Colberg

Long, deep, rambling talks with friends and acquaintances. read less
Society & CultureSociety & Culture
[5] Saurav Pahadia (& Will Hoppin)
Feb 18 2021
[5] Saurav Pahadia (& Will Hoppin)
Check them out:Saurav: https://twitter.com/sauravpahadiaWill: https://vimeo.com/willhoppinThings mentioned:FRC (FIRST Robotics Competition) https://www.firstinspires.org/robotics/frcFuture Engineers Camp, a startup Saurav made https://futureengineerscamp.comthe iPad Pro, which I used for high school & sold many of https://amzn.to/3diIt3ATeacherly, another startup Saurav made https://teacher.lyStripe, a fantastic payments API https://stripe.compersonal.website, a side project of mine https://personal.websiteMystery Books, a product Saurav built"How to Win Friends and Influence People" by Dale Carnegie https://amzn.to/3bdRckVMadrona Venture Labs, where we both worked https://www.madronavl.comAI2 Incubator, where Saurav now works https://www.ai2incubator.comUW Seattle, where we both went to school https://uw.eduParkinson's Law https://en.wikipedia.org/wiki/Parkinson's_lawEdyfi, the community house we live in https://edyfi.comulti.vote, a ranked-choice voting system I built https://multi.voteTesla Today, a Tesla-related news aggregator I revamped https://teslatoday.com"The Art of Negotiation" by Michael Wheeler https://amzn.to/3bfAq4WWEEKDAYS, where I'm winding down a job https://joinweekdays.comSynthesis, where I'm winding up a job https://synthesis.isPay 4 Hate, a potential project for inoculating against rejection (note: we didn't actually end up building this, it IS a bad idea upon sober reflection) https://pay4hate.comProduct Hunt, a place to launch projects https://www.producthunt.comGoogle Domains, my preferred domain registrar https://domains.google.comMechanical Turk, an API for repetitive human work https://www.mturk.comSamson's post about the "sync gap" https://www.samsonzhang.com/2021/01/13/the-sync-gap-a-framework-for-good-conversation-writing-learning-and-all-other-creative-dialogues-internal-and-external.html
[2] Davide Radaelli (& Mattea Holt Colberg)
Feb 6 2021
[2] Davide Radaelli (& Mattea Holt Colberg)
Technological progress as along a tech tree; cheap energy as the next frontier; labor isn’t free; don’t optimize things non-mission-critical things that you can outsource; there’s no such thing as infinite resources; everything is an optimization problem; people can’t perceive reality, just your marketing of it; fundraising is a marketing problem; delegation is not a magic wand; leadership is marketing and decision making; startups have to solve their problems, not just get funded; the easiest way to market something is to be telling the truth; my vague discomfort with Davide’s priorities; the dangers of getting funding through sheer marketing (via Theranos); Holmes had to lie, or she’d get shut down; make sure you’re actually right, instead of getting a vision funded through brute force; blindly believing something doesn’t make it true; first lay out a path forward, then make sure it will work; will perfectly simulating a cell cure aging?; jumping dozens of orders of magnitude fundamentally changes the problem; Amazon is basically just a scaling problem; software complexity does not scale linearly; we can’t “train” a cell model because we don’t know anything about a cell for sure; accuracy of models depends on calibration; results of models still need to be tested in base reality; it is critically important that leaders and modelers know what they’re talking about; how do you know when you know enough; “origins of life” simulation wouldn’t replicate historical evolution; overview existing research on the origins of life (from Mattea, a biologist); evolution is random, so you can’t simulate it forwards and get the exact life we have today; if you constrain an evolutionary simulation to have a certain result, it’s no longer random or accurate; the history of science is realizing that our models are crap; you can’t find novel ground truths through models; I haven’t made a high school because I don’t have an answer yet; sometimes you should start before you have a plan, because it’s hard to validate a plan before you start; always optimize for simplicity; I’m bad at optimizing for simplicity; always keep the end in mind, and look for easy hacks; always be doubting your premises; if you can simulate a human, there’s a ton of non-aging potential; the biggest value-add to fixing big problems is often unsexy; unhelpful fame and glory vs anonymous bigger impact; results count, not labor; my failure to launch a high school; we live at 20-30% overall efficiency; how can you work at 90% of potential?; fully-committed failure is scary; quantifying wasted time.Check Davide out: https://daviderad.comCheck Mattea out: https://personal.website/matteaholtcolberg