In The Age Of AI, The Future Of Programming Is Not More Programming

Artificial Intelligence will not replace coders, it will help users learn powerful programming.

Coders will be replaced, by the visual programming languages, we can now afford to build with the aid of artificial intelligence.

In fact, the correct way to write a computer program, will be discovered by power users, not programmers.

Moreover, the old debate between coder and programmer, is now based on the wrong premise.

In this bright new age, coders and programmers become, system architects, inventors, and digital revolutionaries.

Their job is to discover, invent or architect better system components, finer system architectures.

But their job ends here, they are not to use their inventions, they are to keep going upgrading the landscape ahead.

The inventions are for the new generation of users, who are now equipped with the artificial intelligence that will guide them.

Yes, eventually, eventually, the cycle of programmer’s life takes over, power users become programmers, and programmers retire to become users.

A system component such as the notion of an Agent or Supervisor or even something as tiny as a Signal, or a combineLatest Signal Operator…

Is to be placed in the hands of the bright new minds, under a visual programming system that smartly passes data between Agents.

With AI based expert system, or just AI providing helpful guidance, certainly with AI generated tutorials, recipes, and HOWTOs in the side pane.

The smallest possible implementation of such a visual programming language, is the Apple Automator, which is little more than Command Line on its side.

But that is a singe chain, Output Port of A, has ONE AND ONLY ONE, a connection, to the input port of B.

This allows for a wonderfully simple UI, actions or commands simply stacked one on top of the other.

But the output of an action, will frequently enough, be needed to be sent to multiple other Actions and Operators.

That is why Linux users need Shell Scripting, the single chin command line: is simply not enough.

The only user interface that can exist in this space, which also includes small mobile screens.

Is a map, overview shows zones, close up programs, and clicking on a location will just show a subway map of all the actions.

The programming ergonomics are just fine, you are just adding locations to a map and setting what actions they stand for.

A lot of people feel that Visual Programming is not enough, but they are making a mistake.

They want to write programs with loops and lists, ifs and butts, but the lines on a map represent packets of information.

Frames of a video, server requests for a file, or user actions launces from a button press event in a user interface.

The user interface, is a tree, and it works best as a tree, but to make it dynamic and have it respond to user actions.

We sens well defined actions, that get intercepted, transformed or mapped, or just trigger other actions if they are not filtered out buy some logic first.

That is what the map is for, and it cannot be made more complex, than packets over wire, because it is meant to help the users understand.

The visual programming language, the map, the simplicity of routing packets, is meant to help the users program.

For those of you who are confused how networks of packets can have logic, it is very simple:

The packet that comes from a user who is not logged in, or authorized, is stopped by a filter action, a dot on the subway map.

The filter, stands in the middle of a line that at the next location, triggers actions, that require logged in users with correct permissions.

For a user action or request not to be dropped, it must pass the test or tests contained in the filter.

In your mind you simply imagine, un-authorized user actions. will not get through here.

And you can implement action triggers that do sensitive things, at the next location.

There you could have a switch that routs packets out to different lines, based on the packet type or name.

And there you could think to yourself, “Oh only packets that deal with profile updates arrive at this location”

This passes by bus-stop test, in that you can use your phone to write software, at the bus stop, with just your imagination and fingers.

We have to learn not to be too quick to dismiss these ideas and similar, AI makes all the users power users now.

AI can explain software better than the programmers who created it, and allow me to give you an example.

The user can tell AI what functional reactive framework functions to use, Ramda, or Bacon, and explain what they are trying to do.

The AI is sure to write a solution, and explain it at every step of the way.

Times have changed, we live in a new age now, the age of artificial intelligence.

There is nothing here to dismiss, in fact, take time to marvel at the system interoperability.

A location can be within the user application, on the local server, and on a distant 3rd party server that provides paid components.

Database storage and queries, AI customization, pre-written-prompts, or creation.

The inter-op takes the user across an entire world of complexity, and services, in a consistent user interface.

Compare that to your systems of today, before you complain to me.

Better yet, start building, there are connections, and nodes that have ports, that the connections can bride, at the heart of the node is the action payload.

This is not just about Agents, do not forget the Supervisor pattern.

Who will hold a job in its queue until it is marked completed, by one the workers it recruits, again local, or remote.