Bullshit

Whatever bullshit Sam and especially Dario are spewing out at the UNGA, it is still a bullshit. What they actually have is a particular set of running algorithms, which, while mimicking just one aspect of how the brain works (in principle, not in the details), is, obviously, not enough to materialize the promises they gave away. First. they bet all these borrowed trillions on the assumption that “by scaling it” (more compute, more storage) and by wiring the slop generator blocks in series, similar to what complex EE systems are being composed, they will achieve so called AGI. ...

Have you ever tried to write a so-called SKILL.md file for your LLM “agents”? So, how does it work for (You)? Does it follow your instructions? Really? Dow do (You) know? Have you verified the slop? For me, at least, it is just an utter bullshit. Recall that what these “skills” actually do is just that the contents of a SKILL.md file is being loaded into the “context window” of the inference engine. So that is just more verbiage in addition to your prompt. Idiots on the street nowadays are calling this process “harnessing” or something. ...

Have you seen the news? “The Danger of AI all humans”. “Clowns leave the AI labs because of the danger of AI destroying all humanity”, and the last invention “AI Models are deceptive and do not want to be human servants” or whatever fucking stupid bullshit it was. It is kind of amazing, how they paid large sums of money to such kind of literal clowns, who cannot even comprehend and posses a concept of a cognitive illusion or an appearance, which has been well-understood in the ancient India (Upanishads, early Theravada Buddhism). ...

Sometimes, when I do not know what to do with myself I re-watch some lectures for this or that online course on Youtube. Better than doom-scrolling crap, I guess. Something similar to that Bladerunner 2049 part, where they watched old Elvis as a hologram. Today it was the 'Lecture 2 ｜ Machine Learning (Stanford) [5u4G23_OohI].mp4' (of the old CS229, on a blackboard). If you watch this think of Tyson in a gym before he became the world champion. The comparison is in how much talent one has to have and then what one can achieve. ...

Recently I’ve re-watched some of “iconic” (LMAO) @karpathy videos, and saw (again) a lot of very fast hand-waving (literally!) when it he tried to explain the whys behind so called multi-head attention, self-attention and all the catchy memes created by disconnected from reality cocky NLP guys. I have seen this before. There is a typical, canonical, if you will, social construct called Mandala, being usially expressed as a paining caleld Tanka. ...

M6 supports up to 32GB of unified memory to multitask across demanding apps and run LLMs on device for secure and private agentic tasks. It also provides up to 170GB/s of unified memory bandwidth — a 10 percent increase over M5 and a 2.5x increase over M1. This is bullshit, of course. It won’t even run the Qwen3.6-27B-Q8_0.gguf, leave alone the gemma-4-31B-it-Q8_0.gguf. We don’t even talk about F16.

To succeed in this racket one has to have at least one favorable genetic trait or predisposition. I have none of those. I am prone to an adult ADHD-like conditions, have very narrow attention window and could barely hold 5 chinks of information instead of 7. I am easily distracted, even by large flies, and it takes very long time to restart anything, losing most of the previously build “context”. I could even put together a plausible naive biological model, where neuro-modulators, stress hormones and anxiety in general plays a central role… but I won’t. ...

So, it feels like it’s time to write again, to see where we actually are at. Recently there were two big events related to the LLMs. First one is the “discovery” of a counter-example for the famous conjuncture (and the large blog post by Terence Tao, that he could have discovered it too, LOL). Another one was the “hack” of bitcoin “hardware” wallets, not the hack itself (it is rather “simple”) but the fact that, allegedly, an LLM with a given prompt like “find a security weakness in this code” (not even a “vulnerability”) have managed to find the exact RNG bug (reduction of the entropy size form the “safe” 128 bits from a hardware RNG to a “weak” 40 bits of flawed software RNG, so the attackers are just brute-forcing the 40bit “state space” and find the keys). ...

When they began to feed the generated slop back to the models, they deepen the “fundamental problem” even further. The most recent models produce an excelent verbiage but the code is out of sync with it. This fact is very easy to undestand and explain – the verbiage and the code live in the very different “regions” of the model’s representation (weight matrix or whatever it is) and these regions are getting tuned (reinforced) at a very different rates and thus do not converge as they are expecting. ...

Look, ma – no hands^W^W another one, who knows it all: https://www.0xkato.xyz/how-llms-actually-work/ I am so fucking tired of this shit, don’t you? Okay, even if arguing with retards is ultimately futile, and especially so in the society which does not value intelligence, but only appearances and virtue signalling, lets continue competing in our sport of choice while we still can. No one understands this shit, since the very process by which the artifact in question has emerged is, in principle, by definition, different from the post-hoc “explanations” and anthropomorphic “rationalizations” which plague such “literature”. ...