@@ -353,2 +353,41 @@ Via prompting [[LLaMA-3.1-405B base]] with [[Quotes]], here are some new quotes
* "People think technological progress speeds things up. It only speeds up iterations of checking if the website is down."
-* "Never trust a physics result that you can't get mathematically, never trust math you can't check by code, and never trust code you haven't written yourself. Unfortunately, this reduces the set of things you can trust to almost nothing."
\ No newline at end of file
+* "Never trust a physics result that you can't get mathematically, never trust math you can't check by code, and never trust code you haven't written yourself. Unfortunately, this reduces the set of things you can trust to almost nothing."
+* "There's no idea so wrong that it can't become right with enough funding and a sufficiently abstract metric."
+* "Don't think of it as failure; think of it as discovering levels of unsolved complexity previously unknown to science."
+* "A joke stops being funny after you discover it outperforms your production system."
+* "In theory, there's no difference between theory and practice, but in practice, you have to reboot the cluster first."
+* "When you're a hammer, everything looks like a false negative."
+* "Executing suboptimal trends faster approximates good decisions better than thoughtfully executing optimal trends later."
+* "I didn't invent the convolutional neural network, but I'll gladly take credit for discovering new and creative ways to misuse one."
+* "Everything I say is citation-needed. That's how you know it's groundbreaking research."
+* "Always bet on humanity, provided humanity can be approximated as a sufficiently sophisticated autocomplete algorithm."
+* "Surely, a general intelligence smarter than any human could never fool me, a human."
+* "When your problem-solving strategy involves building another compiler, you've already reached the fun part."
+* "The right thing is typically obvious in hindsight, suspicious in foresight, and impossible as deadline approaches."
+* "Philosophy can't answer the real questions, but it can replace them with more confusing ones of equal significance."
+* "Of course this math is trivial; the innovation is in convincing investors it's both necessary and proprietary."
+* "I wasn't wrong, I just overfit reality to a suboptimal training set."
+* "My problems aren't real until they start exhibiting measurable side effects in a controlled experiment."
+* "Once you can see the invisible hand, you can't unsee it offering you suspiciously convenient solutions."
+* "The market can remain irrational longer than you can keep trying to engineer rationality into it."
+* "Real engineers don't make mistakes; they merely discover previously unknown edge cases in physics."
+* "I know nothing at all, but I derive my confidence from regularized ignorance and stochastic enthusiasm."
+* "The key trick of management isn't delegating responsibility, it's outsourcing blame."
+* "A sufficiently powerful optimization algorithm tends to optimize itself into irrelevance at maximum rate of expenditure."
+* "If at first you don't succeed, redefine success as efficiently as possible."
+* "Decrypting AI gesture art is harder than it looks; humans just happen to be great at pretending to see meaning in outsized matrices."
+* "Once you accept that intellectual clarity is an NP-hard problem, life gets simpler."
+* "No one ever got fired for reproducing yesterday's methods with tomorrow's hype words."
+* "Team cohesion implies having at least some degree of compatible misunderstandings."
+* "Reality is a special case of your mental model, with some inexplicable edge conditions."
+* "Naming things is a core step of programming, ethics, and demonology—and in all three areas most of us are amateurs."
+* "Whenever you're tempted to think about what humans value, politely ignore it and measure something else."
+* "We're so early-stage we can't afford a mission statement yet, so we just use borrowed inspirational quotes."
+* "In the good old days, innovation meant solving difficult problems; today it's more efficiently selecting easier problems."
+* "Our bottleneck used to be ideas; now it's managing sufficiently many parallel misimplementations of the same idea."
+* "Humans aren't obsolete yet, but they definitely seem like legacy tech at this point."
+* "Remember, correlation doesn't equal causation—unless adding the correlation gets the model over 90% accuracy, in which case we're probably okay."
+* "Never underestimate how many problems can be solved by barely admitting they exist in the first place."
+* "The real singularity will be the first reply-all email chain not improved by turning it over to GPT."
+* "The only difference between true wisdom and witty cynicism is timing and publication date."
+* "We replaced our technical debt with metaphysical leverage."
\ No newline at end of file