Machine Intelligence and Learning (MIL?) vs. Big Tech "Artificial Intelligence"
Why MIL instead of "AI"
Who invented the term "Artificial Intelligence" was not too considerate and thoughtful.
We might have an artificial reef. artificial sweetener. an artificial limb, or an artificial teeth.
Intelligence can not be artificial, as bogus, counterfeit, ersatz, fabricated, factitious, faked, false, hyped-up, manufactured, mock, phony, plastic, sham, simulated, specious, spurious, substitute, synthetic, unnatural.
By self-recognizing, today’s AI is made in imitation of human intelligence or as a substitute for the natural brain.
Then one might conclude that all today’s AI, like as promoted by its largest promoters, Apple, Microsoft, Alphabet, NVIDIA, Meta Platforms (Facebook), Tesla, Adobe, IBM, Palantir, Mobileye, Dynatrace, SentinelOne. Uipath, Aurora Innovation, Tempus AI, is
“bogus, counterfeit, ersatz, fabricated, factitious, faked, false, hyped-up, manufactured, mock, phony, sham, simulated, specious, spurious, substitute, synthetic, unnatural” intelligence.
It is not surprising that fake AI tools, as Midjourney, DALL-E, DeepAI, are flooding social media with false information and fake images.
You should not copycat the human brain, but rather its power to encode and make sense of the world.
This is performed by Machine Intelligence and Learning (MIL) with encodes and makes sense of the world of reality, physical, mental, social, digital or virtual.
So, the right term is MIL as technological/machine/computer/computational/cybernetic/electronic/silicon/non-human intelligence.
As such, "AI", its types and sorts, platforms, systems and tools, mimicking human intelligence, thinking and actions, has no solid reality, intelligence and human ethics. It is the travesty of human intelligence emerging as the Big Tech AI bullshit-industrial complex to cause overhype and excitement—and waste $trillions of investment capital.
We are stuck in techno-oligopolies... Our future is either an AI-caused human society collapse or a friendly MIL takeover.
The only viable, all-sustainable techno-scientific alternative is the non-human, human-complete MIL, the real-world foundation for man-machine hybrid hyperintelligence.
"AI" as the Travesty of Humans
A reality check on intelligence is reality (world modeling & simulation & understanding), science (scientific modeling & simulation) and morality (ethical modeling & simulation), not unreality, statistics, and amorality.
"AI" is the travesty of humans and our intelligence, being a false, distorted and absurd representation of human intelligence.
All the present human-like AI, as its platforms, systems and tools, as mimicking human intelligence, thinking and actions, is in reality:
non-artificial, drawing their strength from the work of real humans: scientists, engineers, artists, musicians, programmers and writers whose creative and professional output is now appropriated in the name of artificial innovations;
non-intelligent, drawing their power from substantial computational resources to manage statistical algorithms and large data sets, rote memorization and rote learning public training data sets and pattern-matching.
Its state-of-the-art statistical auto-regressive LL models, Claude 3.5 Sonnet, GPT-4o 2024-05-13, and Llama 3.1 405B, OpenAI’s o1 Series, are nothing but multimodal and large language datasets which are inherently unable of real learning, reasoning or understanding due to lacking any world modeling capacities. [Eureka: Evaluating and understanding progress in AI]
"AI" is all scam AI and AI scams, deep fakes, online abuse, AI hallucinations, data privacy and insecurity, overhype and false expectations, massive fraud and counterfeit.
Today's AI all like China fixes hyper-realistic humanoid robots with the fake robot girls acting robots for a fake "AI" marketing campaign:
Apple, Microsoft, Alphabet, NVIDIA, Meta Platforms (Facebook), Tesla, Adobe, IBM, Palantir, Mobileye, Dynatrace, SentinelOne. Uipath, Aurora Innovation, Tempus AI
[There is NO "Artificial Intelligence" but Machine Intelligence and Learning (MIL?), or why "AI" is a massive fraud and big lie...]; [AI as a disruptive innovation or deliberate ignorance, big lie, wishful thinking, self-deception, and brain-washing]
"AI" = Fake AI = Non-AI
In reality, there is NO AI but Machine Intelligence and Learning (MIL) as real-world intelligent computing that enables a technology (an algorithm or model, machine, computer, device) to interact with reality causally and rationally or intelligently.
"AI" as the idea of a "thinking machine" mimicking human intelligence is not just a a wild imagination conjecture but rather an utter nonsense first suggested by A. Turing in his "imitation game" mental experiment, extended as AI by The Dartmouth Summer Research Project on Artificial Intelligence in 1956 "that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it".
Such an "AI" is all untrue, fake/false/fabricated/fictional, which human-mimicking tools present all sorts of risks and threats from structural unemployment, privacy and security, LAWs arms race to “profound risks to society and humanity”.
It is all-unsustainable to model, simulate, replicate, emulate or mimic [the unknown] human intelligence in machines and computing systems, "to make AI systems learn and reason like animals and humans", and nothing but a waste of $trillions., instead of modeling and understanding the world itself, the universe as the total environment, physical, mental, social, digital or virtual reality.
Now, people using "AI" tools globally went past 250 million in 2023, more than doubling from the 2020 number. This growth in "AI: tool users is expected to continue, pushing past 700 million by the end of the decade, unless humans or all jobs are taken over by the human-like AI models and humanoid AI robots.
Machine Intelligence and Learning (MIL?) vs. Big Tech Fake "AI"
MIL is about developing foundational real-world models and causal algorithms and analytics that computer systems use to perform complex intelligent tasks and solve complex real-world problems.
MIL aims to program a machine how to perform intelligent/intellectual tasks and provide intelligent outputs by identifying real-world/objective/causal patterns/relationships in the world and its data representations.
MIL systems rely on computational comprehensive world models to self-learn and self-correct when provided with new world's knowledge, information or data.
As the world witnesses unprecedented growth in ML technologies, it's essential to consider the potential risks and challenges associated with their widespread adoption of fake/false/fabricated/fictional/untrue AI.
Whoever believes or follows or involves in the human-faking AI is either full ignoramus or full fraudster, intentionally or not, involved in "deceit with the intention to illegally or unethically gain at the expense of another", with all the legal consequences.
Some of the most common types of fraud involve the insurance industry, and the mortgage market and the stock market, especially, the AI stock market owned by Big Tech, Apple, Microsoft, Nvidia, Amazon, Google or Meta.
The "Big Tech AI" fraud costs the economy trillions of dollars, not mentioning all threats and risks for humanity.
To be specific, the so-called Safe Superintelligence (SSI) has raised $1 billion in cash to develop safe artificial intelligence systems that far surpass human capabilities, involving top venture capital firms Andreessen Horowitz, Sequoia Capital, DST Global and SV Angel. NFDG.
It is life-critical to understand what sort of "AI" is promoted by large techno-corporations, and what risks and threats it poses.
Put simply, in the context of the current human-replicating paradigm of building larger- and larger-scale "AI" systems, there is no "AI" without Big Tech.
Besides, every startup, new entrant, and AI research lab is dependent on the large tech firms. All rely on the computing infrastructure of Nvidia, Microsoft, Amazon, and Google to "train" their human-mimicking systems and tools to deploy and sell their "AI" products.
Again, the SEC chair Gary Gensler has warned that having a small number of AI models and actors at the foundation of the AI ecosystem poses systemic risks to the financial order.
Our public message is:
the big tech AI is to be disrupted by MIL?, identifying the real-world intelligent technologies, models, applications, products, goods or services.
MIL has programmed “world models” — machine's internal models of how the world works - having an all-encompassing architecture of real-world multi-graphic hypergraph networks, which is the most integrative or comprehensive conceptual, mathematical and data and algorithmic structures, from facts to laws, measures to categories or arrays to graphs or regression and decision trees to artificial neural networks.
As such, MIL? is invalidating the big tech "AI" systems needing computationally intensive statistical methods, millions or billions of pieces of labeled training data and millions or billions of reinforcement learning trials in virtual environments.
Replacing AI/LLMs with MIL World Models
A MIL World Model is a multi-trillion dollar question for autonomous intelligent machines, transcending human mental models, as individual or collective worldviews, conceptual models and all specific AI models, as foundational LLMs.
It is the only true path to real-world intelligence, learning and understanding, full autonomy and automation, to common-sense knowledge and reasoning, informed decisions and interactions, comprehending of the basic nature of reality and its contents, and why the world is the way it is and how it is to change.
The MIL World Models are combining the ontological and computational modeling of the world, as the totality of All and Everything, of entities, interactions and causal relationships, with a variety of conceptual models of a complex reality.
It is ranging from mental images to scientific, mathematical, digital or metaphysical models, representing a single thing, whole classes of things, the environment, the physical, spatio-temporal universe or all physical, mental, social, digital, virtual reality.
The MIL's variety and scope of world models are due to the variety of tasks, goals or purposes, from generating information content, predictions, decisions to self-driving cars to space robotics.
It is the world modeling power, worldviews, cognitive maps, spatial maps, schemas, or structural knowledge, what unites human and machine intelligence. For any intelligent "agents—whether human, animal, or artificial intelligence (AI)—need to form internal representations of the world and themselves?to plan and respond flexibly to changes, constituting internal world models (IWMs)".
LeCun proposes that one of the most important challenges in today's "AI" is devising learning paradigms and architectures that would allow machines to learn world models in a self-supervised fashion computing gradient estimates of some objective function and then use those models to predict, reason, and plan.
But it is an utter nonsense to model, simulate, replicate, emulate or mimic [the unknown] human intelligence in machines and computing systems, instead of the world itself, its reality, the universe as the total environment.
Again, one should not reduce the World Models to human mental models, like Yann LeCun?and Yoshua or Karl Friston, and many other researchers, viewing the world model as the key to evolve "AI" but in the way biological intelligence has evolved to make AI systems learn and reason like animals and humans.
So, untrue "artificial intelligence (AI) is technology that enables computers and machines to simulate human learning, comprehension, problem solving, decision making, creativity and autonomy", or the human body/brain/brains/behavior/business.
"The AI we see today still lacks common-sense, reasoning behind decisions, understanding of the basic nature of their environment, and why the world is the way it is. They never know the intrinsic understanding of the world and causality of their actions, and they are as good as the data they were trained on. Hence it was plausible that they would fail in unpredictable scenarios, and driving in the real world definitely gives the worst of it".
The application domain of MIL software and hardware, <a sphere of knowledge and reasoning, influence and activity, the targeted subject area to which the user applies a MIL program>, is not the subjective reality (human intelligence/cognition/behavior), but the total world modeling of subjective and objective, physical and digital reality
Modeling and understanding the universe or the world ;in all its complexity, scales and scopes, is what the MIL models after.
For the real-world machine/technological/computing/cybernetic intelligence and learning consists in simulating, modeling and comprehending the world and its contents, its physicality and virtuality:
real world entities (objects), states, phenomena, processes, or behaviors and relationships (patterns, rules, laws, associations), concepts (abstract objects, subjective reality, mentality) and data (quantities, variables; language, audio, images, software code, text or video).
In a domain-specific "AI", logical and mathematical models or?statistical models and neural networks are used for building?expert systems?and?knowledge-based systems or statistical predictive and generative models, as Narrow AI, ML, DL, passive, prompt-responding large language models or conversational AI chatbots.
Such "AI" is without real intelligence, learning and inference or explainability and interpretability demanding specific types of machine learning techniques or algorithms, from repression analysis and decision trees to artificial neural network and deep NNs, allegedly "modeled after the human brain's structure and function" or "simulating the complex decision-making power of the human brain" (complex human brain has nothing with NNs, in any architectures).
In the domain-general MIL, where total subjective and objective, physical and virtual reality is encoded as the real-world multi-hyper-graph networks, with multiple nodes, edges and loops, generalizing all conceptual, scientific, mathematical, concept and data models, we create general-purpose MIL systems, making a true, real, genuine, generalist AI.
Of all domain-specific AI projects, we know only one project which mission is "to understand the universe" (xAI, "a company working on building artificial intelligence to accelerate human scientific discovery. We are guided by our mission to advance our collective understanding of the universe").
As we have posted in our recent, the real, true or genuine AI as the real-world machine intelligence and learning (Trans-AI = MIL) exists as conceptual computational world models for building a general-purpose AI technology.
It is embracing conceptual, mathematical or statistical modeling techniques and methods, such as narrow AI models, DNNs and LLMs, and has not been implemented, developed, deployed and distributed, for geopolitical, ethical and financial reasons.
It was indicated that the mainstream big tech AI is nothing but deliberate ignorance, big lie, wishful thinking, self-deception, or brain-washing.
Real-world, interactive generalist MIL as being implemented as a real-life autonomous MIL, does not exist, yet..
We have developed the working world model of general MIL relying on the world hypergraph network models, to be encoded, embodied, enacted as real-world intelligence technology, which x-times costlier than the first nuclear bomb of the Manhattan Project ($2/27B).
It is only Mother Nature could cheaply produce natural intelligence in its natural reproductive ways.
Again, it is hugely expensive to engineer real-world intelligent entities; for you need as a necessary condition to model, organize and encode the most of the world learning and laws, a data source of real intelligence machines, embedded as binary codes and mathematical algorithms, causal data structures and algorithms and programs.
Or, in computer science language, you need to develop the real-world Data Structures and Algorithms fitting for MIL Programs and Applications.
Even you possess the infinite computer power for web-scrapping all the 147 zettabytes Internet data, a zettabyte is about a trillion gigabytes, to train machine learning algorithms and models for various applications, it will not produce full real-life intelligence.
Now, “Why artificial intelligence doesn’t exist?”
Because of a defective assumption of simulating human intelligence as a multiplicity of intelligences, when there is no single MIL, but a multitude of specialized AIs, designed or data-trained for narrow jobs, tasks, works. Comparing this to a human capable of performing any tasks,
Now, each specific task requires a dedicated specific AI, from board games to objects recognition to chatting to protein structures to car-driving, each of which could algorithmically/mechanically perform only what is designated, being completely unaware of its task. No more, no less.
There is no AI, but statistical software tools
This human-replacing big tech AI perspective deprecates the idea of a general-purpose (ideally, single omnipotent omniscient) MIL and pushes us to see the mainstream AI as a set of specialized software tools, as sampled below:
Now, all "AI" tools users with the latest smart phone holders promising to have "AI" built-in are frauded and cheated:
Apple and Google dominate the fake AI market with a joint total market share of 72%; for the IOS and Android operating systems control 99% of the global phone market.
As of 2021, there are about 4.2 billion devices with fake AI-powered assistants. By 2024, this number is expected to double with 8.4+ billion devices powered by false AI assistants.
The biggest concern comes from some survey that nearly half (49%) of Alpha children are already using "AI" tools corrupting their cognitive, social, and emotional development, while millennials and Gen Z have been informed by the internet and the smartphone revolution.
All in all, big tech "AI" appears as a giant scam, deliberate ignorance, big lie, wishful thinking, self-deception, and brain-washing, all to be outlawed, while MIL? is a disruptive general-purpose technology transforming the world.
Resources
Real Artificial Intelligence, Machine Learning, Deep Learning, Data, the World
The MIL is to merge Artificial Intelligence?(Weak AI,?General AI, and?Strong AI) and Machine Learning (Supervised learning,?Unsupervised learning, Reinforcement learning or Lifelong learning) as the most disruptive technologies for creating real-world intelligent systems.
Today's Machine Learning has No Mind or Intelligence, Learning or Reasoning, Consciousness or Self-Knowledge, Perception or Understanding, but just data, statistics, mathematical models, training techniques, software automation, GPUs and MPPs.
If we’re not careful, Microsoft, Amazon, and other large companies will leverage their position to set the policy agenda for AI, as they have in many other sectors.
SUPPLEMENT: from the Nuclear, Space and AI & ML Arms Race to MIL Takeover
In the 20th century, there had been a few disruptive technological projects and strategic initiatives in the search of global technological dominance, to achieve superior spaceflight or superpower or superintelligence capabilities:
The Soviet space program setting many records in space exploration, including the first intercontinental missile (R-7 Semyorka) that launched the first satellite (Sputnik 1) and sent the first animal (Laika) into?Earth orbit?in 1957, and placed the first human in space in 1961,?Yuri Gagarin, first woman in space,?Valentina Tereshkova, in 1963 and the?first spacewalk?in 1965,?computerized robotic?missions?exploring the Moon?starting in 1959: being the first to?reach the surface of the Moon, recording the first image of the?far side of the Moon, and achieving the first soft landing on the Moon. The Soviet program also achieved the first space rover deployment with the?Lunokhod programme?in 1966, and sent the first robotic probe that automatically extracted a sample of?lunar soil?and brought it to Earth in 1970,?Luna 16.?The Soviet program was also responsible for leading the first?interplanetary probes?to?Venus?and?Mars?and made successful soft landings on these planets in the 1960s and 1970s. It put the first?space station,?Salyut 1, into?low Earth orbit?in 1971, and the first modular space station,?Mir, in 1986;
NASA's Apollo program, also known as Project Apollo, of "landing a man on the Moon and returning him safely to the Earth". Apollo set several major?human spaceflight milestones. It stands alone in sending crewed missions beyond?low Earth orbit.?Apollo 8?was the first crewed spacecraft to orbit another celestial body, and Apollo 11 was the first crewed spacecraft to land humans on one. It is repeated as the Artemis Program. Artemis III (2026) is planned to be the first American crewed lunar landing since Apollo 17 in December 1972;
Reagan's "Star Wars" Plan, the Strategic Defence Initiative, involving advanced weapon concepts, including lasers,?particle-beam weapons, and ground and space-based missile systems, along with sensor,?command and control, and computer systems needed to control a system consisting of hundreds of combat centers and satellites spanning the globe;
Japan's Fifth Generation Computer Systems (FGCS) to develop "Knowledge Information Processing systems" (applied AI);
the US Strategic Computing Initiative of advanced computer hardware and AI to build full machine intelligence, from?chip design and manufacture,?computer architecture?to?artificial intelligence?software, for $1B, within from 1983 to 1993.
The 21st century has been marked with the alternative intelligent technology projects:
The global AI & ML arms race, led by superpowers, as the US, China, EU, Russia, and big tech Google, Nvidia, Meta and Microsoft, to lead "to make AI systems learn and reason like animals and humans". It implies reaching the human-like intelligence technology, the superior military AI power, superhuman humanoid robots and Lethal Autonomous Weapon Systems (LAWS).
The global MI & ML initiative implying man-machine hybrid hyperintelligence, while excluding the global AI arms race, the human-like intelligence technology, the superior military AI power, superhuman humanoid robots and LAWS.
AI That Grows Your Business, Not Your To-Do List.
2 个月AI has complex implications, raising valid concerns. Azamat Abdoullaev