this has @visakanv vibes
February 2025
17 tweets
81 likes
6 retweets
you could definitely build something like this now with LLMs!
respecting distrust is a necessary part of allowing trust to build. applies to subtle interpersonal philosophical contexts, and to a toddler learning to eat:
when was the last time you unblocked someone's work? I spent 4h today getting @Morphenius set up to code for the first time in >20y* so he can build LLM-based tools for helping ppl learn cognitive empathy (how to teach) (*my, what has changed. not just LLM IDEs but also git!)
@eshear > The bureaucracy's aversion to punishment (audits, budget cuts, bad press) leads to strong conservative biases in policy updates, just like PPO's clipping prevents large policy changes. This is why process changes usually add constraints rather than remove them - deletion is
tfw someone wakes you up and wants you to do something before you've even had a chance to read your system prompt
@DRMacIver yesssss x.com/Malcolm_Ocean/…
startup idea: real quantum coin flips a lil circuit in a coin that detects when it's being flipped, does some quantum shit, then shifts its inner weights so it actually forks the multiverse when you read it
@Morphenius have you written up "agents can't self-refer" yet?
@exgenesis it's unnecessarily narrow even on my fairly-wide screen size. margins on the page too big.
@ChrisChipMonk vibes with this
idk about a wiki page but this is the sort of thing that LLMs are fairly good at. here's Claude: --- Let me first expand on those initial examples, then add some additional crucial experiments: Those first two examples need some important clarification: • Galileo's gravity experiments (though the Tower of Pisa story is likely apocryphal) demonstrated that objects of different masses fall at the same rate in a vacuum. He used inclined planes and careful timing to show that, contrary to Aristotelian physics, a heavy object doesn't fall faster than a light one. This overturned the common-sense notion that heavier things naturally fall faster, helping establish the foundations of modern physics. • The double-slit experiment, first performed with light by Thomas Young in 1801, showed interference patterns that could only be explained if light behaved as a wave. Later versions with individual photons and electrons revealed even stranger quantum behavior - single particles somehow going through both slits simultaneously and interfering with themselves, demonstrating wave-particle duality. Here are three more pivotal experiments: • Michelson-Morley Experiment (1887): Using an incredibly precise interferometer, they tried to detect the "aether" - a hypothetical medium through which light was thought to propagate. Their null result showed no evidence of aether, helping pave the way for Einstein's special relativity and fundamentally changing our understanding of space, time, and light. • Miller-Urey Experiment (1952): They recreated the conditions of early Earth's atmosphere in a sealed apparatus and applied electrical discharges to simulate lightning. Within days, they found amino acids - the building blocks of proteins - had formed spontaneously. This demonstrated how complex organic molecules could arise from simple chemicals, providing crucial evidence for theories about the chemical origins of life. • Mendel's Pea Plant Experiments (1856-1863): Through meticulously controlled breeding experiments with pea plants, Mendel discovered patterns of inheritance that led to the basic laws of genetics. By carefully tracking traits across generations and applying statistical analysis, he showed that characteristics are passed down through discrete "factors" (what we now call genes) rather than by blending of parental traits as was commonly believed. CopyRetryClaude can make mistakes. Please double-check responses.
I'm done pretending I'm gonna grow @IntendApp. But I'm gonna keep it alive, and I'm open to the possibility that it might thrive more with someone else's energy/leadership.
every couple of weeks, I stumble onto Facebook videos. and there's this bizarre trend where many are reaaallly long and boring but with a sense of anticipation. like some imminent reveal scene in a sketch takes LIKE 4 MINUTES. why? there aren't ads mid-vid.
haven't read this article yet about the science of purpose but I am very confident I'll want to. vibes with some of my writing on exaptation which I'll put in the next tweet also maybe @Morphenius
.@praeterpropter do you and @longtermgov know about The SimPol Solution? it's got some cool proposals that either might be good inspiration for your work or good "wrong on the internet, now we shall explain" fodder
wtf image models cannot handle "a guy doing a handstand". incredible shape rotation problems this is google's imagen 3. they look okay for a beat til you realize somebody took the barbie apart and put it back together backwards
