The Intelligent Parrot

6 min Read

LLMs like ChatGPT continue to demonstrate how over-rated human intelligence is!

 

For the longest time, we humans have assumed that we have vastly superior intelligence. In fact, we feel that we are far superior to all other living beings – because of our higher level of intelligence.

The Turing Test of Intelligence (1950)

Seventy-five years ago (1950) Alan Turing proposed The Turing Test* -to  test a machine’s ability to exhibit intelligent behavior which is  indistinguishable from that of a human being. The test involves a  human tester who asked the same questions in natural language  simultaneously to a human & a machine, both concealed adequately. The human had to finally guess which was human & which machine. 

*Refer “Computing Machinery and Intelligence”, circa 1950.

 

A decade later (1960) John McCarthy from MIT first developed a language based on List processing & called it LISP – a language which is still popular in AI & self generative software. Unfortunately, the  computing power & data storage available 75 years ago were inadequate to support processing of statistical routines as well as store vast amounts of data that is required.

Role of Computing Power

But for a long time, even after the advent of AI, it was restricted in specific use-cases like attribution & predictive analytics. One of the major (if not sole) reason for the huge popularity of AI is the ability of GenAI to understand complex questions in natural languages and respond from our vast resources of available data.

 

Today’s LLMs (Large Language Models) have demonstrated that (artificial) intelligence is based on access to humungous data along with statistical tools* for multiple analysis of data patterns combined with superior computing power. By the way, sharp memory with super-fast & high accuracy of response is a general description of an intelligent person!

 

LLMs have effectively & efficiently mimicked this facet of human intelligence. With a data set of 30+ years of humongous Internet data  & the superior computing power available, LLMs are able to train & retrain their algorithms in finite time & use them to answer queries in real time – something that was quite impossible till a few years ago. 

 

The result: LLMs mimic the behavior of “understanding” the patterns of natural language prompts and make authentic replies in the same natural language.

 

And all our  assumptions of superior human intelligence have suddenly under question.

Intelligent Quotient (I.Q.)

Until the 19th century, it was common to send all mentally challenged children to mental institutions in most of Europe’s “advanced civilizations”! It was only In 1908, that the Binet-Simon scale was designed to identify those  intellectually-challenged children who could be managed with special education without “institutionalising” them. This noble thought was the starting point for measuring intelligence.

In 1912 German psychologist William Stern coined “Intelligenzquotient” (IQ) as the ratio of “mental age” to chronological age x 100. (If an individual of 10 years of age has a mental age of 10, then the IQ will be computed 100. And as the mental age became greater than the chronological age, the IQ value would increase to “more intelligent”.)

Intelligence Equals Memory

LLMs, AI, Machine Learning have convincingly demonstrated that most of what we call “intelligent behavior” can be achieved with access & ability to process huge amounts of data. LLMs used specific statistics to extend this to  converse  in natural languages. 

 

From past experience, we know that when the data is really, really large & complex, a simple query will take far too long and yield too many answers. So we train statistical models on this data & use them to approximate the best answer (even if it’s approximate.) In fact, using statistics is the only way to make sense of huge data.

Memory = Access to Data using Multiple Pattern Recognition Softwares.

Armed with the above, GenAI is available to mimic human intelligence & appear startlingly “human” – both by understanding complex natural language queries (prompts) and generate complex content both text, graphic & video.

A Parrot Who Responds Intelligently!

This refers to the pic at the top. Sachin Bhosle is our Senior VP, Business Development & the proud caretaker of a rescued parrot. During a recent meeting, I had the chance to interact with this intelligent parrot – a pleasant surprise!

 

To demonstrate his parrot’s intelligence Sachin  pretended to argue with his wife in a loud tone in their Marathi language. The parrot immediately replied in near-legible Marathi “What’s the matter? What happened?” And when Sachin’s wife (name) switched off the AC in the bedroom, the familiar voice piped in again “Feeling hot. Please switch on the AC!”

 

Quite clearly, the intelligent parrot was not only able to recognize & store the set of voice patterns (prompts); he had also saved the matching voice patterns which got him the best response for each prompt! If his memory developed further, the parrot could start resembling an intelligent member of the human race!

Questions, Questions

Our Hindu scriptures define the intellect as a tool which analyses past data & uses it to predict/worry about the future. So our formidable “human intelligence” is a combination of access to humongous data preinstalled statistics software. (Statistics remember, is mandatory to make sense of any big data.)

If the value of “human experience” is proportionate to the amount of data that is available (via statistics software), then is only pattern recognition enough “make judgements” of what is likely to work in any situation?

And is limited or malfunctional internal statistical software the medical explanation of dullness, dyslexia, hallucinations, bipolarity & even insanity?

 

Btw, what will be our role in the workplace & home when everything is “suggested” by LLMs? And what will substantially occupy us human beings  when we actually give permission to LLM based software to sign agreements, issue invoices, make payments?

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts

Ad Tech

Insure your Ad agency’s Profitability..

Every agency is a hotbed of fast-paced advertising operations or ad ops.

Read More
Ad Tech

How many Excels are reqd. to run your Ad agency profitably?

Every Excel sheet is a “partial version of the truth.” Because every Excel sheet is the viewpoint

Read More
Ad Tech

Can Advertising Passion Co-exist with Commercial Discipline?

How would it feel if you could run your ad agency super-smoothly – like a commercial machine?

Read More
Rocket
How aDintelle came to be
Back in February 2017, we launched the first version of aDintellé (then called aDintell). With Print, TV & Radio modules along with a basic Virtual...
Read More
Single Platform for All Your Agency Business Needs dark (1)
Single Platform for All Your Agency Business Needs
In the fast-paced world of advertising, where deadlines are tight and clients are demanding, agencies need to stay ahead of the game to ensure...
Read More
image (13) 1
Tectonic shift of Entertainment consumption in India: The streaming revolution
India has witnessed a tectonic shift in cricket consumption.

Read More
AdIntelle-Blog2 image 1
Interesting But Predictable Trends in India's Advertising Landscape
Year on year digital ad spend has been on the rise, and 2023 is no exception. Meanwhile TV, Print, Radio & Outdoor media continue to shrink /...
Read More
Social search web adintelle 1
FUTURE OF SOCIAL WEB 2025-26
Have you ever wondered how the landscape of social web search will evolve in the coming years?

Read More
Agency Management blog image 1
Using Accounting Software Alone for Your Agency? Think Again.
The complex ecosystem of ad agency operations involves constant balancing of financial management with creative endeavors.

Read More
Adintelle_Blog image
A Silent Ad Agency Revolution
Since the past decade or more, digital agencies have been increasingly offering full services, with added focus on creative strategy

Read More
adintelle _last_blog_image 1
Transforming TV Ad Operations: How aDintelle Revolutionized Billing for India's Premier Agency
In the fast-paced world of agency conglomerates, where every second counts and accuracy is paramount, billing for TV advertisements poses a significant...
Read More
Pilgrim's Progress
A Pilgrim’s Progress
It's been 3 decades since we established Mediaware as a software service for ad agencies. And here we are, still relevant after 3 decades!...
Read More
Blog1
How will ad agencies operate in the gig economy, driven by LLM?
It's been 3 decades since we established Mediaware as a software service for ad agencies. And here we are, still relevant after 3 decades!...
Read More

End to end agency operations suite

Looking for multiple solutions?

A crazy bunch of marketers, techies and product enthusiasts trying to build the next big thing!

Our Story

Driven by a passion for empowering advertising agencies, aDintelle is on a mission to redefine the way agencies work and supporting them in achieving their fullest potential.

About Us

Our team of dedicated professionals is committed to delivering excellence by providing cutting-edge software solutions tailored to agencies' unique needs and enabling them to achieve remarkable results.

Join Us

We are a dynamic team of young and experienced members where talented individuals can thrive and make a significant impact while also having some fun!