“Even thyroid and milkshakes and quesadillas aren't always enough to make up for the horrors that are imposed on us and some of the natural defects of the environment.” – Ray Peat
I believe that explains his ability to synthesize ideas from a dozen different seemingly unrelated topics into a single coherent worldview which definitely lends itself towards top tier writing ability but I believe prose, structure etc. are a craft that need to be practiced separately to convey those ideas convincingly. Wondering if he ever did writing exercises, stream of consciousness writing, imitation etc to sharpen that skill. All I remember him saying that is slightly adjacent to this is that he didn’t like reading fiction.
If health is mostly understood as a thousand separate pathway failures, you can sell a thousand separate fixes. If health is more fundamentally about maintaining a coherent energetic state, many of the best interventions become broad, preventive, generic, and cheap.
@nikitabier X should add a, in addition to mute by region, mute by sentiment. It's more important than muting by region imo as introducing physical boundaries to idea space is less congruent with the spirit of free ideas.
There should be a way to, in addition to muting words, mute posts based on sentiment. To me, this makes more sense for a “free” platform compared to banning by location.
The algorithm is pretty trash for me at the moment and it’s been getting worse since the past few months.
@doodlestein Would be cool to have a tiny program that collects/streams your prompts for a particular project for others to learn from. Open source code is not enough lol
The rent seeking will never end even if we have fully local models. We won't be able to do a lifetime purchase of a model so the next best thing is to build leasable local models. These will be cheap though. Should see this in the next year or two hopefully.
I think it's wise to have many idea docs stashed up ready for the next OOM improvement in AI. Highly likely the upcoming agentic capabilities would nearly one shot many things.
Came across a brilliant analogy insight somewhere–around the invention of the personal computer, the companies that bet against the PC in favor 10 supercomputers total were comically wrong. Similarly, those betting on like only 10 GPU clusters now would be wrong in 10-15 years. Everyone will have their own hardware to run AI.