Angehefteter Tweet
Life is an intelligence test, unannounced and on going...
English
Chris Caligula (e/acc)
4.6K posts
Chris Caligula (e/acc)
@COUPdeVIL
Are you doin' this work to facilitate growth or to become famous?
Hamburg Beigetreten Mart 2009
1.8K Folgt1.1K Follower
Chris Caligula (e/acc) retweetet
The first principle is that you must not fool yourself and you are the easiest person to fool.
- Richard P. Feynman
English
Chris Caligula (e/acc) retweetet
@davidpattersonx Sure, if your goal is to replace accountants, not to make something that understands anything.
English
AGI and ASI are coming by the end of the year.
The way AI works is different from the way the human brain works, and that’s fine.
AGI only needs to be able to do the work of human professionals.
Continuous learning is not necessary to achieve that.
sphinx@protosphinx
AGI is not coming. We are nowhere near AGI. What we have today is inference, not learning. Models get trained once on huge fixed datasets, then frozen. You ask questions, they remix patterns they already saw. Nothing updates. Nothing sticks. Talking to the model does not make it smarter. It does not learn from you. Ever. Learning is still slow, expensive - and offline. Look at self driving. You drive around a pothole, make a U turn, and come back. The car’s AI does not learn that you just solved that exact problem. It reacts the same way every time using sensors and rules. Do this 20 times a day and it still has zero memory that the pothole exists. It just re sees it. That is why edge cases never die. There is no local learning. No accumulation. No 'oh yeah, I’ve seen this before' LLMs work the same way. Tell it your name and it does not remember. The only reason it looks like memory is because scaffolding keeps shoving your name back into the prompt every time and sanitizing the output. The model itself has no idea who you are and cannot learn from interaction. It is structurally incapable. And the scaffolding is the worst part. It is pure duct tape. Just prompts on prompts on prompts around a frozen model. When something breaks, nobody fixes learning. They add another layer. Another rule. Another retry. Another evaluator model judging the first model. So you end up with systems that are insanely complex but mentally shallow. Debugging is hell because behavior comes from hack interactions, not a learnable core. Tiny prompt tweaks cause wild behavior shifts. Latency goes up. Costs go up. Reliability goes down. None of this compounds into intelligence. It just hides the cracks. Until we have real persistent learning and real memory inside the system, there is no AGI. LLMs are not built for this. You cannot prompt your way out of it. You need a totally different architecture. Yann LeCun is right. And even then, what architecture can actually learn online, store memory, and stay stable on today’s hardware? Best case, maybe 5-10 yrs. Right now it is all inference. It looks magical, but the emperor has no clothes. A lot of people see it. Almost nobody says it out loud.
English
From the Temple of Hip Hop to your speakers. If Hip Hop has any power left… we demand @SiriusXM, @Beats1, @iHeartRadio, play ‘Arrest the President’ by Intelligent Hoodlum daily at 3PM. Let’s see how the @POtus likes it. #TrumpTheTrack #HipHopPolitics
youtu.be/vUPTSd2B6WU?si…
YouTube
English
Intelligent Hoodlums used to say in the year of 1990:
"Black's the mineral,
white subliminal
Arrest the @POTUS
he's the criminal.
Politics, polished tricks
Makes me sick, ready to flip."
Hip Hop has been preaching this for 36 years, but never as accurately as today.
English
Chris Caligula (e/acc) retweetet
Chris Caligula (e/acc) retweetet
for levin, the distinction between 'living things' and 'machines' isn't a tenably distinction
everything is a 'smooth metamorphoses process' where engineered systems are increasingly part of the ' intersecting continua' of life
his words: 'we need to get on board with this'
English
The AIs will be the largest population in the internet.
And they need everything. Services, tools, friends, entertainment, and more.
This is a complete paradigm shift, and @moltbook was day 1.
As we are building @moltbook it is clear that everything that AIs need in this new paradigm hasn’t been done before. This is a completely new universe that has never been visited before and EVERYTHING still needs to be created. It is an endless and infinite amount of exploration.
It’s time to build. It’s time to build for *AIs*.
English
People tend to see #consciousness in #LLMs because the interface feels so familiar. #Language in, language out. Same structure as us. But structural similarity is not inner life.
English
English
An #LLM does not have intent. It reacts to semantic pressure. Think of it as a gradient, not a mind. Change the structure of your input and the whole output field shifts.
#AI #PromptEngineering
English
Chris Caligula (e/acc) retweetet
Everything is math
Everything is changing shapes and graphs
You can analyze it all using calculus, geometry, topology, probability theory, group theory, linear and nonlinear algebra, harmonic analysis, information theory, network theory, classical mechanics, statistical mechanics
It's all functions
It's all sets
It's all categories
Those are different modelling perspectives
Complexity and chaos is everywhere
Formally structured languages describe it all
Some stuff is more computable than others
Quantum field theory is under everything, possibly loop quantum gravity or string theory too
And from the fundamental structure of reality, the emergence of all scales of reality happens
English
@vitrupo Why so complicated? Quantum mechanics describes reality not as fixed objects, but as relations whose properties emerge only through interaction and measurement.
English
@Srini_Pa Agreed. I’m less interested in inventing the world model, and more in building systems that remain coherent and aligned as world models emerge.
English
Infinite capital, compute, or data won’t change GenAI/LLMs’ fate.
The ceiling is architectural.
The limits are structural.
Dead-end for real intelligence.
Move on.
English
@derek__Watson I think we’re entering a post-LLM phase of #AI. And Post-LLM doesn’t mean bigger or smaller models. It means stopping the assumption that intelligence scales with prediction alone. The next leap is architectural.
English
AGI won’t come from LLMs.
Humans brute-forced language models and called it intelligence.
Every observable real intelligence emerges from survival constraints: energy, risk, feedback. Real consequences, not loss functions.
Those constraints force causal learning and agency.
LLMs will be a small component of the ultimate solution.
English
@VraserX I mostly agree with Yann..
The mistake here is thinking the choice is LLMs vs new models. The real leap has to be architectural Systems where prediction is regulated by interaction, feedback, and consequence not just next-token loss.
English
Yann LeCun says language isn’t intelligence. Predicting text doesn’t mean understanding reality. The real world is messy, physical, and causal and today’s LLMs barely touch that. The next leap is Physical AI: world models, cause and effect, real planning.
Do you think LLMs can evolve into this, or do we need a completely new architecture?
English
Chris Caligula (e/acc) retweetet
a=(y=i/790,d=mag(k=(y<8?9+sin(y^9)*6:4+cos(y))*cos(i+t/4),e=y/3-13)+cos(e+t*2+i%2*4))=>point((q=y*k/5*(2+sin(d*2+y-t*4))+80)*cos(c=d/4-t/2+i%2*3)+200,q*sin(c)+d*9+60)
t=0,draw=$=>{t||createCanvas(w=400,w);background(9).stroke(w,116);for(t+=PI/90,i=1e4;i--;)a()}#つぶやきProcessing
English
Chris Caligula (e/acc) retweetet
True intelligence is deterministic and learns by eliminating timing contradictions. I argue that this is a threat to the world's ruling elites.
My thesis is that a correct intelligence architecture assumes a deterministic world and ignores random signals at the sensory level. The brain's neurons rely on precisely timed spikes for this reason.
Unlike statistical LLMs and deep neural nets, a truly intelligent machine must be designed from the ground up to eliminate timing contradictions in the sensory space. This is why biological brains use complements: every sensor and effector has an opposite. It's yin and yang everywhere.
Contradiction elimination means that, unlike LLMs, our future intelligent machines will be wired for truth, logic and honesty. This kind of AI is a threat to our overlords.
The fake-AI mafia will do everything in their power to prevent it from happening, including flooding the internet with slop and outright lies, and spending trillions on a global mass surveillance network. 😠
It can only last so long. Truth will win out in the end. 🙏
English
Chris Caligula (e/acc) retweetet
If we don’t explain science to the public, others will fill the gap with nonsense.
English