Chanchana 🐳

@aaryan_kakad @0xSero your light mode is super broken btw

@0xSero theneuralforge.online this ui is oneshotted using gpt5.4high

This is for all the frontend/fullstack devs, it'll make your life easier. Do you want GPT-5.4 & mini to make better UIs? No amount of prompting is going to fix the issue, the model simply is not good enough at frontend design on it's own. I love it, just not fro UI. 1. Go on ui.shadcn.com 2. Go through their UI builder 3. Click "Copy Command" 4. Run it Voila, GPT now uses an expert crafted UI-component kit to build you any UI with perfect consistency. All free and open source btw.

@AI_EmeraldApple Here's the basic principle of reasoning: if you claim a thing exist, it's your burden to provide the evidence for its existence. It's called Hitchens’s Razor

@AI_EmeraldApple Sorry, but you will need more believable bulshitting to convince me that god is a thing.

Simulation theory is like "finding god" for the atheists. The problem and the reason why so many people cringe at the idea of "god" is because popular religion and mass organized religion deconstructed the idea of "God" into a literal cartoon bearded man in the sky that looks like humans and has the same psychology... and treat bible verses as concrete atomized concepts that has no relation to the chapter or to the whole. Thomas Aquinas, perhaps one of the most intelligent men to live, said it best in Summa Theologica. God is the pure spirit with zero body, zero parts, zero location, zero time... almost like a hologram that exists in all space and time at once and nothing at the same time. The essence of existence itself God isn't something human beings can comprehend, just like an amoeba can't comprehend quantum physics. We can't even ask the right questions of reality, and maybe the best we can come up with is "simulation theory"... but even that has a problem of infinite regression of the simulators themselves being simulated themselves and so on. Simpler people think that humans "look like God," but Genesis doesn't say we look like God. It says we were made in HIS image, after HIS likeness. God is the rational soul... specifically ties into our intellect: the power to know, reason, the logos... and will, the power to love and choose freely between good and evil. That's the "image", not something stupid like a mirror selfie, but a real, participatory likeness that's beyond a simple picture. In Christian theology, specifically, this is the idea that god made sub-creators as humans who can ponder creation itself and, in doing so, we humans dimly reflect the ONE who just is existence. This is why Christianity, in its sophisticated form, invites us, humans, towards discovering god's design... and this was precisely why the scientific revolution was born out of the Christian tradition and nowhere else. This is why almost all of the scientists who were godfathers of modern science... Copernicus, Kepler, Galileo, Newton, Boyle, Maxwell... were Christian... and even people like Fermi and Einstein were deeply spiritual in the same tradition, even if they weren't specifically "Christian" . Here, the implication is that the language of mathematics itself is a small fraction of god's design. This is why high-IQ people arrive at the same conclusion as the low-IQ people of the uncaused cause. High IQ people see the beauty in the logos, in reason, the rational order, and in math, and see a fractional glimpse of god's design. Low IQ people accept god as the default of existence because they intuit it without reason. It's the midwits who claim atheism because they can't fathom religious thought to be anything sophisticated... classic Dunning-Kruger effect applied to metaphysics

@doodlestein Do you have a problem planning with GPT pro since it doesn’t have context and cannot look up your private repo like Codex?

If you actually follow my process exactly, this is not an issue at all. Where people struggle is when they do a half-assed job with the planning and don't follow the strategy I outline, which is extremely intense and involves many, many rounds of iteration and the contributions of all frontier models. I highly recommend actually trying what I propose, and you'll see that it works. There are 800 people in the Flywheel Discord who can confirm what I'm saying here, too. It's not just me who has gotten this process to work so well.

The complete guide to my Agent Flywheel approach of integrated tooling, workflows, and prompts (or “How I Learned to Stop Worrying and Love Generating 1,000 High-Quality Commits a Day”): agent-flywheel.com/complete-guide

@doodlestein To give an example, let's say while you are planning, you need an answer to an experiment, so you need to let the AI code the experiment first otherwise your plan would explode in complexity e.g., the plan might say "if experiment result is X do Y otherwise do Z"

@doodlestein However, because you are not in contact with reality (tinkering with the app), it can still cause you to move slower because you are doing too much mental simulation. Think of it similar to a large learning rate in gradient descent. I like planning, but it needs improvement.

@IanISSawesome @Youclidean @elonmusk The main problem people go nowhere with their therapy work is because of their lack of true honesty. They keep circling around the surface, not brave enough to dig deeper into what's lying underneath. So of course when you keep avoiding the pain, you never understand and heal it.

@IanISSawesome @Youclidean @elonmusk This is false because there's been a lot of evidence that shadow work is effective. And shadow work looks like reinforcing negative neural pathways on the surface because it requires diving into the unconscious and the pain.

Reinforcing negative neural pathways via therapy or introspection is a recipe for misery. Don’t cut a rut in the road.

@elonmusk A good introspection is simply truth seeking. If you are afraid of truth because of the misery it might bring, you are probably confusing true introspection with rumination. Also, if your value truth over comfort, why even care whether introspection will bring you misery?

@oskargroth @grok Analyze my statement truthfulness from first principles.

@oskargroth It makes sense that it should be able to reverse engineer any app you install locally if it's smart enough because all the logic is inside the binary and that binary is executed by your PC so you cannot have any secret otherwise your PC can't execute it. Logic is just obfuscated.

Wasn't joking about this one btw You can reverse-engineer pretty much any part of Apple platform internals in seconds using Claude or Codex with Hopper MCP

@caviterginsoy @phyrooo @ChristosTzamos Dude you know the model has unreliable computation. Why are you saying things that make you sound ignorant?

@offchan420 @phyrooo @ChristosTzamos I’m not following what your argument is exactly, because models do have some calculation ability already. What you’re saying is already what’s happening now.

1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers. We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy

@caviterginsoy @phyrooo @ChristosTzamos This overhead argument is huge because it's possible that during reasoning, you might need to do back of the envelope calculation a lot. That requires specialized algo that the model should invent on the fly and discard once it no longer needs.

@caviterginsoy @phyrooo @ChristosTzamos Another reason for why you want model to have some compute capability is because there are cases when the overhead of calling tools exceed the actual computation itself. For example, let's say you want to do 2+2, are you going to call a tool for that? Probably not.

@Plinz software is literally crystalized human intelligence. it's about compressing thought patterns into a form that is repeatedly executable. the whole point is to eventually delegate all the information processing tasks to machines so that we can be free.

I deeply agree with Emily Bender's main point: LLMs are useless, unless you want to offload cognition. (The other two usecases she suggests are rare special cases of the third.) Offloading cognition into machines has always been the purpose and application of computer science and AI.

@Prathkum If you can code faster, the Theory of Constraints states that the bottleneck would shift to something else like how fast you generate ideas, how fast you get customer feedbacks, how fast you review code, etc. So you cannot ship 10x faster (whatever that means).

Everyone is talking about writing code 10x faster with AI. Very few are showing the products they shipped 10x faster.

@robinebers @lassvestergaard why are you lying? how is 2 prompts worth $100+ value? get a grip

@lassvestergaard bro it’s $20 lmao you’re complaining after getting probably $100+ out of value (probably more) if $20 is your budget, go china models

never buy claude pro, this is after using opus 4.6 with plan mode 2 times and just redo a landing page 2 times. All took max 4-6 minutes 🥲

@JShodanVR motion sickness

I don’t understand. This stuff is real, it exists, you can actually buy it, and yet the VR market is almost dead. What happened to the kids in the 90s who dreamed about this possibility? Why do modern kids now go crazy for games where you don’t really do anything, where at most you just swipe up and down, or spend their time screaming in stupid games that require no concentration, like amoebas with no curiosity? This stuff finally f***ing exists, and yet the videogame market pretends it’s invisible. F***! F***! F***! In 1998, when we were 13 playing Resident Evil 2 on the PlayStation with my friends, we would have killed to be able to play something like this. Credits: @VirtuixOmni here’s their YT channel: youtube.com/watch?v=Evyjom…