There is No AI but Man-Machine Intelligence and Learning (MMIL?)
What is to be intelligent?
What it means to be intelligent. Is it the ability to cognize, conceive, classify, conclude and communicate? To learn, retain and apply information? To recognise oneself in a mirror? To use tools? To assume things? To be creative? To think or reason? To be logical? To have critical thinking? To understand and deal with situations? To adapt to change? To solve complex problems? To make decisions?
There is a triple of features that any intelligent entities are assumed to possess:
Cognizing and Understanding, Continuous Learning and Search for Knowledge, Critical Thinking and Rational Interaction.
What is Machine Intelligence?
What is to be intelligent for machines, devices, or algorithms? It is to interact with the world rationally/effectively/causally/intelligently, knowingly processing information, making decisions and taking actions to achieve its goals in a wide range of environments.
Real AI is Machine Intelligence and Learning Machinery (MILM) which is not mimicking humans but encoding and making sense of reality to rationally/causally/effectively interact with the world/environments/machines/humans to complete human intelligence.
Now, to estimate some empirical risk and optimize the performance of some statistical algorithms on a known set of training data, a sample of IID training data points, to calculate correlations patterns in large datasets, to generate mathematical models to interpolate, even not extrapolate, data, like in today's AI/ML/DL systems, have a little with real intelligence or true learning.
IID random variables are used as an completely unrealistic assumption to simplify the underlying mathematics, statistics and probability theory of AI, predictive or generative versions, its machine learning or neural networks algorithms.
The world's data is not "a sequence of independent, identically distributed (IID, i.i.d., iid) random variables, data points."
The world' data is rather a causal multi-graphic hypergraph network of "dependent, distributed (DDD) random variables", representing real-world systems as complex networks (graphs), physical, biological, brain, mental, narrative, social, technological, neural, computer, etc.
In mathematics, formal ontology, science and network science, an undirected multi-graph hypergraph of causal variables is a generalization of all conceptual, scientific structures and mathematical structures, such as measures,?algebraic structures?(groups,?fields, etc.),?topologies,?metric structures?(geometries),?orders,?graphs,?events,?equivalence relations,?differential structures, and?categories.
The human brain or MIL is designed to make assumptions to search for patterns, ‘mental models’, worldviews or world models, to make it a more efficient intelligent machine.
Or, something is as intelligent as its internal capacities to assume right big and small things, to take something as true, accept, conclude, consider, estimate, expect, guess, infer, presume, speculate, suspect, think, or understand, encoded as learned algorithms or behavior. Our assumptions make our feelings, reasoning and actions.
Real AI as MIL makes basic assumptions about reality by means of large world models, embracing large language models to make it a more efficient intelligent machine.
A?MIL causal hypergraph network?is representing the world of complex systems as the universe of complex sets of symmetrical interactions or relations between and among various entities or objects. It is consisting of a set of?nodes?connected by?multiple, parallel looped edges/arks/lines/links/connections/relationships joining any number of nodes or vertices.
Real AI integrates AI, ML, DI, ANNs, LLMs, AGI with human intelligence in terms the causal multi-graphic hypergraph networks of the reality modeling and simulating, learning and inferencing, understanding and interaction machines, technologies, platforms.
Converging AI Models with Human Intelligence
True AI follows the strategic development formula:
Real/Integrated/Total/Hybrid AI = Man-Machine Intelligence and Machine Learning Machinery (MMILM?):
MIL +
AL +
ML/DL/NNs +
GenAI/LLMs +
NLU/NLP +
Causal AI/ML/AGI +
Robotics/Automation/IoT+
Human Intelligence…
The Dawn of a Real AI Era
It's clear that an Integrated/Unified/Hybrid AI represents a disruptive technological advancement—it's a paradigm shift in how we approach the real-world problem-solving with MMIML. By combining the best of various AI models, we're creating systems that are real intelligent, reliable, and suited to handle the complex challenges of our time.
NVIDIA has recently announced Project GR00T, Generalist Robot 00 Technology, a general-purpose foundation model for humanoid robots, acts as the mind of robots, making them capable of learning skills to solve a variety of tasks, designed to further its work driving breakthroughs in robotics and embodied AI.
“Building foundation models for general humanoid robots is one of the most exciting problems to solve in AI today,†“The enabling technologies are coming together for leading roboticists around the world to take giant leaps towards artificial general robotics.†Jensen Huang, founder and CEO of NVIDIA. NVIDIA GTC 2024 Keynote. Don’t Miss This Transformative Moment in AI
NVIDIA is building a comprehensive AI platform for leading humanoid robot companies such as 1X Technologies, Agility Robotics, Apptronik, Boston Dynamics, Figure AI, Fourier Intelligence, Sanctuary AI, Unitree Robotics and XPENG Robotics, among others.
领英推è
Will robots have the ability to think? Nope.
They never never could think or reason or feel or perceive or learn or plan or understand or create as humans.
Humans are humans, machines are machines, they only could complement each other acting in their own effective ways mutually reinforcing.
There is NO AI Technology but MILM
There is no AI as imitating human intelligence in existence, as well as NO AI-powered chips in actual existence and there is NO AI tools for any use cases.
It exists only in your mind, as any abstract or fictional or imaginary objects, like the neutralizer from MIB, the death star or lightsaber from Star Wars or the predators from the Predator trilogy or Skynet as a ANN-based conscious group mind and artificial general superintelligence system of the Terminator franchise.
AI chips are a hype, lie, confusion, and fraud, all in one, led the the biggest fraudster named Nvidia, controlling 80% of the global market share in GPUs, violating US antitrust laws, and accused in patent theft, and simulating its GPUs, as H100 and A100 processors, as AI processors, powering FF AI data centers, produced by Taiwan Semiconductor Manufacturing Corporation (TSMC).
What Is a FFAI Chip?
A FFAI chip is a specialized integrated circuit designed to handle AI tasks. Common graphics processing units (GPUs), field programmable gate arrays (FPGAs) and application-specific integrated circuits (ASICs) are all considered FFAI chips. Central processing units (CPUs) can also be sold as AI chips, effective for AI workloads and training FFAI models,
Again, parallel-processing GPUs vs. FPGAs vs. ASICs vs. NPUs refers to a microchip, which is an integrated circuit unit that has been manufactured at a microscopic scale using semiconductor material.
Etched transistors-switches power computing functions, such as memory and logic, whwre memory chips manage data storage and retrieval, logic chips are behind the operation that parallel processes the data.
Modern FF AI simply would not be possible without these specialized FF AI chips, be it:
Generative AI, LLMs, chatbots, digital/virtual assistants, content generators
Edge AI, smart devices — watches, cameras, kitchen appliances, smart homes or smart cities
Autonomous Vehicles
Robotics, humanoid robotics
If you jump from general-purpose chips to narrowly specialized chips, having parallel processing capabilities, more energy-efficient, more accurate results and better customized, such new special design attributes, you still have regular chips.
Again, there is no Real AI hardware in existence, YET, be it AI chips, processors or computers, platforms or data centers, but FFAI computing hardware used in the development and deployment of FFAI systems.
The future of real artificial intelligence largely hinges on the development of true, causal AI chips.
PS. FFAI vs, Real/True AI (RTAI)
FFAI is a wide-ranging branch of computer science that aims to build machines capable of performing tasks that typically require human intelligence. FFAI allows machines to match, or even improve upon, the capabilities of the human mind.
Or, FFAI refers to computer systems that are capable of performing tasks traditionally associated with human intelligence — such as making predictions, identifying objects, interpreting speech and generating natural language.
FFAI aims to provide machines with similar processing and analysis capabilities as humans, making it a human replacement.
TRAI systems work by using the world modeling engine, causal rules and algorithms, knowledge and data. Smart small data is collected and applied to causal mathematical models, or real algorithms, which use the information to recognize patterns and make predictions in a process known as training.
The primary approach to building AI systems is through causal machine learning (CML) where computers learn from causal datasets by identifying real-world patterns and causative relationships within the real-world data.
CML is using causal NNs, a series of algorithms that process real-world data by mimicking the structure of reality, its domains, processes and phenomena, systems and objects.
Resources