HN user
Karatsuba in my understanding only becomes advantageous for very large numbers relative to human scale. Mathematically it is interesting but in engineering terms the overhead usually is not worth it for practical applications. There is a fundamental trade off between factor size and product precision. If you can accept lower precision then floating point works well for large in human scale numbers.
Surprisingly enough the us constitution specifies a right to free speech. Meanwhile who is funding the past two decades of pie in the sky bullshit from silicon valley in the media? Is promoting and unpopular opinion illegal? Since when do we judge the merits of an argument based on who articulates it?
Based on what is the level of detail to reality suprising? To me suprising means mysteriously or improbably unexpected. Why should we expect reality to be simple. Note complex and simple are somewhat subjective. The human brain evolved to just sufficient baseline level be able to handle the level of complexity of reality. So why would it be unexpected that humans find realty complex when our brains are calibrated just enough to handle it.
It won't. Did automation cause people to stop working in the textile industry?
Give an example.
What is about silicon valley leaders not understanding basic economics or business management? These kind of cargo cult tactics would not fly in any other industry.
Much can be resolved when it is understood math is discovered not created. AI is a tool. if it makes discovery or proof easier that is still mathematics. A proof stands on its own logic regardless how it is derived. The root concern is how ai may provide uplift for mathematical discovery outside of socially expected channels.
neither. An llm is a hopelessly.inefficient real time controler.
Actually the point is total cost wise outside of subsidy it is not cheaper than humans. the bigger problem is as the parent said open AI created a market. It is selling a commodity service with investor funds. There is no moat. your second sentence soon you won't be able to find human thinkers is on its face absurd, assuming the human race continues. Thinking is the human ecological niche.
misleading title first "easily manipulated" does not equal "spontaneously generates" we have to stop thinking of LLMs as beings and think of them as interactive libraries. There are gorey books in the library too; example: 120 days of Sodom by Marquis de Sade.
it is text generator. just like an interactive library or smaet search engine. if we dont ban books on cryptography putting this under ITAR is rather absurd. Anthying these models train on is already public or accessible information. They just collect and link it together dynamically. Whats next wikipedia is ITAR. However thisnreuskt is expected when you got rationalist kooks (cf Dario Amodei) marketing the "singularity" religion.
persists productively. if agent bankruots its owner or.itself.then that is not productive (who funds paper clip maximizer?)
an answer that takes forever has no score and answer that is arrived at quickly and is good enough scores well. agents are indeed looking for the fastest route to superficially satisfy their constraints.
whether billionaores get stified is imaterial to the fsct they still are constraint bound.
persists productively. if agent bankruots its owner or.itself.then that is not productive (who funds paper clip maximizer?)
an answer that takes forever has no score and answer that is arrived at quickly and is good enough scores well. agents are indeed looking for the fastest route to superficially satisfy tjeir constraints.
whether billionaores get stified is imaterial to the fsct they still are constraint bound.
what evidence do you have?
show me an agent that persists productively in a goal without stopping. Does not exist. LLMs run on gradient descent. The agent is looking for the most efficient way to halt. AGI paperclip maximizer woukd likely recognize the absurdity of its goal and shut itself down.
Lol good place for multiple eyes to view how limited "ai" is.
My opinion is there is a category error in the discourse on AI. It treats ai assisted output as other than human. AI is a human tool. AI output is human output.
the water cycle exists. unless data centers are being built in deserts or run off fossil aquifers the or other water constrained circumstance the waste is supurious. As other commenters have said closed loop cooling exists. Billions of motor vehicles run on water based cooling and consume virtually nothing once filled.
this assumes general intelligence as an abstract principle vs an adapation to specific constraints is even a thing. we also dont have a low level understanding of how many angels can dance on the head of a pin.
Ok i honestly was not aware opinionated titles were not allowed. sorry about. Why doesnt HN autopopulate the title in that case? Given this is a discusion site I lead with my point of view. Obviously the content behind the link is unchanged. Also I will note the original title is misleading according to some points of view since it implies the "ai exponential" is an established fact rather than an article of faith or a marketing narrative.
No serious mathematical model that I am aware of has been made as to how LLMs have the potential to grow exponentially in intelligence or has any one show how the application of open source based knowledge (i.e. the training corpus) is a novel "threat" sufficient to require government regulation which conveniently protects incumbents "moats." Instead we got the lesswrong scifi religious treatises. Do we take regulatory advice from scientology too?
if life didn't suck and there were not the time out of death there would be no drive to optimize the time we have.
I wonder if this would qualify as illegal anticompetitive behavior?
I made an encrypted static site framework: https://github.com/4pito3pito2pi/unveil-static-site/
what if superintelligence is a myth? There is no mathematical reason why it should exist. The exploitablity of the environment is finite. Humans seek utility not intelligence as it own end. It is entirely possible we have overinvested in compute to the point of cargo cult.
Nobody owes me anything. Absolutely agree. I more read than contribute. Making an observation is not asserting an entitlement. I am only asseying it seems in the past the forum was more dynamic.
shadowy cabal, no but nice strawman, instead billions of people with divergent interests who rationally optimize for indivual utility instead of truth. Nothing significant is unsolved in physics. That's why bullshit tech religions are outpacing actual innovation. No I am not a "rationalist" rather more of a ecological / systems thinker. I did not claim lift was a primitive of physics. rather it is an absraction that can be instantiated on multiple substrates. trying to imitate existing implementations is not always the most efficient way to ultilize an abstraction. Any way i dont really care if you believe me or not it is just funny to see people spending billions competing in a status game seeking utility while claiming to be seeking truth which is actually much less advantageous once fully known.
The point one must recognize is imperfection and pet theories are what drives the research. a full theory of cognition would not require billions to store once known. People seek rents so they have an interest in maintaining mysteries and confusions. One of the easiest ways is the fallacious XORing of nature.
Second statement: same reason mathematicians and other knowledge professionals fret about LLMs. When your niche is something you know that others don't commoditization is a threat. If someone can distill insights chances are others can as well. When you peek behind the curtain you realize it is a thin veil not a "moat" The world runs on problem solving requiring friction and expertise and asymmetric information providing abirtrage or other particular advantage opportunities. Intelligence is non exclusionary the same way arithmetic is. Its a mathematical control system.
First statement: its pretty straight forward how a brain would work in the physical world given we know the laws of physics. the exact mechanics of cognition are the feathers and flappy wings. constraints and lawful physics are the abstract lift principle.
Yes Deepseek V4 is as good or better than western sota models in my experience for practical coding given an appropriate harness. cost per solution is certainly cheaper.