AI Entities in the Future and the Concepts of Life and Soul

In the future, or even now, these are the entities related to AI:

• Computer
• A computer is an oracle computer (knowledge) operating on input/ouput or functions on top of avatar.
• Avatar
• Avatar is a lifeless body for computer.
• Machine
• Machine is intelligent, and functions on its own
• Automaton
• Automaton is automated may with intelligent features to understand input but totally automated

The concepts of living and soul

• When an entity functions on itself, it has LIFE
• When an entity believes in God, it has SOUL

A True Graph Note-taking App and RemNote Features

A true graph-based note taking app need these features in data storage layer:

• Independent nodes
• Edges between nodes

And the features on UI:

• Create independent nodes
• Linking between node
• Pin nodes to top level
• Reference to a node, and expandable reference to node

RemNote is quite a good graph-based note-taking app with all those required for graph-based note taking:

• Data storage
• Each top-level note is a node
• Reference and portal to make linking
• UI features
• Very easy to create independent node, click Create button
• Linking between nodes are based on references and portals
• Pin to top level
• With ref and expandable ref
• Even better
• Create nested items as nodes
• Linking even to nested items inside node
• Pin a node or pin a nested item are both ok
• Ref and portal can be created to note or nested item

If considering Wiki-like as a graph of article, RemNote is better:

• Each nested item in note (article concept in wiki) can be a node too.

The Magical Number 12 in Our Solar System and Human Eye with Imagination

There are some values in solar system, and relation to our human eye. The earth diameter is around:

• 12,000 km

The distance from Earth to the Sun is

• 12,000 times Earth diameter

When consider 1 sensory cell in our eyes as 1 pixel, the resolution of one human eye is roughly:

• 12,000 times 12,000 pixels

With the info above, if the Earth is 1 pixel, human can still imagine up to the Sun, but not further.

Strategy & Innovation: The Way for Business Growth and to Stay

Businesses need human resources (human-first), capital to grow, of course. But beside those, strategy and innovation is the way to grow and stay.

Some examples of existing businesses:

• Google
• Strategy: Keep being the search engine with the most results and top relevancy.
• Innovation: Google creates new products now and then
• Windows
• Strategy: Be the OS with the most apps
• Innovation: Windows is changing with new features on every release
• Android
• Strategy: Be the OS for mobile with the most apps
• Innovation: Changing with new features
• Facebook
• Strategy: The social network focused on communing with friends
• Innovation: Features are still added in, but the fundamental concept of social is changing, people are not posting as often as before
• Steam
• Strategy: To have the most games
• Amazon
• Strategy: Have all kinds of products from everywhere
• Apple
• Strategy: Be something stylish
• Nvidia
• Strategy: Not much, but luckily hit the hype of digital coins & AI
• TSMC
• Strategy: With the most advanced chip technology

Rarely, but important, some guys with new concepts and successful

• PayPal
• New concept: Secure online payment

These guys started good but lacking innovation after

• Yahoo
• Strategy: Originally, all services for the web from the beginning era of Internet
• Innovation: No more after a few decades and the biz is now relying on existing users

SNN and the Future of Computing

We started with Perceptron, Neuron, then MLP (multi-layer perceptrons), then ANN (Artificial Neural Network), but we're entering the new era of SNN (spiking neuron network) which is bio-simulated instead of bio-inspired (maths-based).

SNN is the point to start for Neuromorphic VM, the leading point to neuromorphics, and robotics, true machines. SNN is important just as we don't mould manually and make physical products without the design software such as CAD.

Intel is having Lava, a framework to make SNN to run on Loihi a neuromorphic processor by also Intel. The are some other libs to convert ANN (eg. TensorFlow, Keras) to SNN, and the other way round which is SNN to ANN to run on TensorFlow backend.

Bottom line, here's how we are going to:

Perceptons -> ANN -> SNN -> Neuromorphic VM -> Neuromorphics.

Microsoft Copilot for Desktop: New Kind of Input 'Vision'

The new Copilot Desktop is amazing, it can even chat about a screenshot (Win+Shift+S). This is AI which is taking image input beside the original text input; possibly further will be talking about what being recorded in camera. This is from the Copilot for desktop only, the Copilot web version has no screenshot item in the menu from '+' button.

Free Apps (Online & Offline) which are Capable of Being Brain2

The new definition of brain2:

• Brain2 = Graph of Trees or Structs including lonely Nodes
• Requirements for app: Graph, with structural UI (not autolayout)
• Optional features: Convenience is better.

Here is the comparison of FREE apps both online and offline which can be used as brain2:

• RemNote: With graph, structure is also content, very convenient
• Specific point: Can store large amount of data
• Cons: No E2EE encryption, can't pin nested docs (work-around: pin nested docs to a top-level doc as Refs)
• Pros: ESE (easy structural edit with tab/shift+tab). Additional structure on left panel for overall view (outline view) Outline view is important to see structures of multiple items at the same time.
• LegendApp: With graph-like only (mirrored items), structure is also content, very convenient
• Specific point: For small amount of data only, each top structure is a file
• Pros: ESE (easy structural edit with tab/shift+tab). With E2EE encryption
• Dynalist: With graph-like (thru' Ref), structure is also content, very convenient
• Pros: Can pin nested items, but no overall structural outline on left panel.
• Cons: RemNote has more features
• Anytype: With graph, structure is also content, not convenient UI
• Pros: E2EE encryption
• Cons: No easy-structural-edit, need to switch to a node to create children while RemNote and LegendApp can tab/shift+tab to create multiple levels of nesting docs inside.

These apps are without graph, or graph-like, or work-arounds to make it similar to graph

• Standard Notes: 
• No structure (pay-to-use)
• Joplin: 
• With graph of notes but the structure is tree of notebooks instead of notes
• Notion: 
• Work-around to make graph by linking notes but no structure shown, the note-subnote structure on left panel is not containing that graph by linking notes
• Obsidian: 
• Is a plugin for VS Code not separate app
• Evernote, OneNote, Google Keep, Amplenote

These guys are based on file per note and notes are not linked together to make graph or structures

• Dropbox Paper, Proton Docs

These guys are free but the limits are too small to use

• Workflowy

Some minor apps author hasn't tried out

• Simplenote

Brain2 vs the Classic File System and Notes Concept

We've been using the concept 'file system' for so long, it's time to change. In this AI-powered era, OSes should function on a new concept called Brain2 instead of File System.

Brain2 is a data system which is based on graph, graph of trees or structures, including separate nodes can be in the graph too. 

The old file system concept is about storing files in a tree, and all other data are on that tree, including files to serve the databases.

The new concept is that file system concept is replaced by a graph-based database at core and it is already a database without functioning on top of files.

Visually when using apps on such Brain2, user will see notes and graph of notes first instead of files first then notes as a type of files.

AI Services will be Much Cheaper in Near Future

Currently, AI service pricing is rather costly. The reason behind is not about the new tech, but because of hardware prices. A single machine learning (ML) entry GPU is T4 which can serve roughly a single chatter at a time, L4 for 2 chatters at a time. The cost is huge to buy so many GPUs to serve people, and seems only tech behemoths can do it for now.

Nvidia business expansion is still good. Nvidia started as GPU maker for gaming, then got more production for digital coin mining. The biggest fortune of Nvidia now is selling GPUs for making AIs. Only behemoths are buying in big quantity now but sooner or later, all companies around the world will buy and make AIs for themselves and the total sale will be huge.

The point is when there are so many AI providers in the world, the AI bot service will be cheaper.

Generations of Processors and the Middle Step of Intel to Neuromorphics Era

We started with complex processors, called CISC CPU (complex instruction set computer), which were complex and process complex logics directly but found not efficient.

We then changed to use RISC CPU (reduced instruction set computer), until now, with reduced complexity of the CPU for efficiently. Clock in hertz (Hz) increased higher and higher in the starting era with single core. We are now at multi-core era of CPU, and will reach 3D stacking soon to ensure Moore's Law still correct:

• CISC single core
• -> RISC single core
• -> RISC high frequency
• -> RISC multi-cores
• -> RISC 3d stacking

Later on quantum processors will be the thing but for the present, Intel is making 2 things to the future:

• NPU - Neural Processing Unit
• Loihi - Neuromorphic Processing Unit

While Loihi is still rather experimental, NPU is practical now, the middle step can be seen in NPU. NPU is processing graph (neural network graph).

• NPU - Processes graph of data
• Loihi - Contains true neurons (but still layered).