Thinking for AI

Coding Necessary?

Metamonkeys has always asked, ‘Is coding necessary knowledge to engage in the digital world of the future?’. Not necessarily because we have an entire sector of business dedicated to solving that problem for us — Software Developers. Even then, we can not seek Developers’ help for solving problems of small scale, and generally we will not spend time learning a new skill for small issues. This is not Knowledge inadequacy, just a bridging difficulty. That will be the heart of discussion in this article and to present that doubt more clearly and move to the other side of the discussion, to the AI part, we can assume two scenarios:

Case 1:

A mechanical engineer working on a prototype needs help on a complex calculation.

Case 2:

A small business owner who heavily relies on the digital products like Google Sheets, Gmail, Canva and others to self-manage his business but finds it difficult to track the Courier orders.

These two cases require minimal scope of external expertise that can be solved by spending some time on that skill. What if ChatGPT, an AI based language model, can assist us? If we look into the two problems, solutions are simple for the machine equipped with these basic skills :

Case 1 solution:

Feed the equation from the text book along with the variables, and then it will produce the answer for it.

Case 2 solution:

Create a Google Apps Scripting to detect the mails that contain the Courier Booking number and record them in Google Sheets. (ChatGPT can do it.)

So, I asked ChatGPT:

Write a Google Apps script to extract the third column from one Google Sheets to a new Google Sheets document

It Replied:

function extractThirdColumn() {
  // Replace "SourceSheetId" with the actual ID of the source Google Sheets document
  var sourceSheet = SpreadsheetApp.openById("SourceSheetId");

  // Get the third column of the source sheet
  var sourceRange = sourceSheet.getRange("C:C");

  // Create a new Google Sheets document
  var newSheet = SpreadsheetApp.create("NewSheet");

 .......
 (Some 50 lines of code in between)
 .......
 
  // Copy the third column of the source sheet to the first sheet of the new document
  sourceRange.copyTo(newSheetFirstSheet.getRange("A1"), {contentsOnly: true});
 }%3E)

“To use this script, you will need to:

1. Replace “SourceSheetId” with the actual ID of the source Google Sheets document. You can find the ID of a Google Sheets document by looking at the URL when the document is open. The ID is the long string of characters between “/d/” and “/edit”.
2. Run the script by going to the “Tools” menu and selecting “Script editor”. Then, click the “Run” button or select the “Run” option from the “Run” menu.

I hope this helps! Let me know if you have any questions or need further

assistance.”

Now, into more about the NLM and AI:

Conventional vs Externalised Knowledge Bridging

The digital world is specifically not for ‘all-knowing ones’ and the technological advancements have relied on automation to take leaps in the past, undeniably. But before the mundane and monotonous parts of the job were automated and now with the language model getting better it has steered through the logical queries to address even the creative/subjective query.

Now, how far can this technology go?

Machine — Language — Human

Natural Language Processing is teaching the computer to understand human language. Programming languages like C++, Python and Java are the languages that a machine is designed to understand whereas NLP enables the machine to understand and communicate in human language. It uses Artificial Intelligence as its core to gain that ability, by simply building a network of patterns. In simpler terms it does two things:

Data Preprocessing: It’s like learning the alphabets and knowing what is what with the data that is feeded into the machine. Ex: If I say ‘A,B,C…’ is the sequence of alphabets to the machine, it will believe what is instructed to be the ultimate truth.

During the Data Preprocessing stage — Learning before mimicking.

Algorithm Development: This is the stage where the data(instructions) will be tied to form a pattern so it can recognise a user’s input. Ex: If you ask ‘what is the third English alphabet?’,it will say ‘C’.

So, it is simply a complex network of the instructions that is being executed by the machine to answer a single question. But it only mimics what has been taught, the intelligence that is artificially created is the ability to seek the patterns, not the ability to think. Because if we teach the machine ‘D’ comes after ‘B’ it will still say that the third english alphabet is ‘D’, not ‘C’.

After Algorithm Development — the network of instructions is relayed.

Thus, a simple but painstakingly long process is behind AI development.

Executing logic and executing vision:

In his article, Stephen Moore asked a very common speculation about the ChatGPT as a prompt “When will artificially generated writing take over?”. It was honest and bound to the understanding of this rising technology,”…Ultimately, the role of human writers will still be important in producing unique and original content”, said ChatGPT. Even when looking at the foundational structure, this intelligence is only mimicking, even the feared creativity is only mimicked. It can write a review about an author because it can read huge volumes of other reviews and weave a writeup out of it.

Still an unfair advantage for the AI, but creativity is just unachievable for an AI to generate. This technology is still going to be a cornerstone in the advancement of every field because it has the speed, versatility and accuracy to gear us to even take leaps. Maximising its use cases is also dependent on the choices we make, because be it logic or vision it’s still executing it, not producing it.

Metamonkeys YT: https://youtube.com/@metamonkeystv

Twitter: https://twitter.com/metamonkeystv

Discord Server Link: https://discord.gg/Bt67UtsuF6

Instagram: https://www.instagram.com/metamonkeys_tm/

Mastodon: https://mastodon.social/@metamonkeystv