AI conversational text model, ChatGPT, the end of Google?

9 Dec 2022 | AI conversational text model, ChatGPT, the end of Google? |

The latest rage sweeping the Internet circa December 2022 is chatGPT, What is chatGPT? Here I’ll let it (chatGPT) answer that question itself (The paragraphs below were created with this chatbot)

ChatGPT (https://chat.openai.com/chat ) is a cutting-edge language model developed by OpenAI. It is a variant of the popular GPT-3 model, which has been trained to generate human-like text in real-time. This allows it to be used in a variety of applications, including chatbots and conversational interfaces.

One of the key features of ChatGPT is its ability to generate text that is both natural and contextually relevant. This is made possible by the vast amounts of data that the model has been trained on, which includes everything from books and articles to social media posts and conversations. As a result, ChatGPT is able to understand the nuances of language and generate responses that are appropriate for the given context.

One of the main benefits of using ChatGPT is that it can improve the user experience of conversational interfaces. By providing timely and relevant responses, ChatGPT can make interactions with chatbots more engaging and natural. This can be especially useful for customer service applications, where users often have specific questions or concerns that they want addressed.

Another potential use for ChatGPT is in the creation of personalized content. The model’s ability to understand and generate text based on context means that it can be used to generate unique content for individual users. This could be especially useful for applications such as personalized news articles or social media posts.

Midjourney Generated Image

Overall, ChatGPT is a powerful and versatile language model that has the potential to revolutionize the way we interact with conversational interfaces. By providing natural and contextually relevant responses, it can improve the user experience and enable the creation of personalized content. As the technology continues to evolve, we can expect to see ChatGPT being used in a wide range of applications in the future.

Generating Code

Like Github copilot , chatGPT can actually generate code , you can ask it simple things like , “generate a number guessing game from 1 to 1000 in python ” and it will render the example below (change it to JavaScript, or Gobang or Pascal ) and like magic an equivalent code sample is produced.:

import random
# Generate a random number from 1 to 1000
target = random.randint(1, 1000)
# Prompt the user to guess the number
guess = int(input("Guess a number from 1 to 1000: "))
# Keep track of the number of guesses
num_guesses = 0
# Loop until the user guesses the correct number
while guess != target:
  # Increment the number of guesses
  num_guesses += 1
  # If the guess is too high, inform the user and prompt for a new guess
  if guess > target:
    print("Your guess is too high. Try again.")
    guess = int(input("Guess a number from 1 to 1000: "))
  # If the guess is too low, inform the user and prompt for a new guess
  else:
    print("Your guess is too low. Try again.")
    guess = int(input("Guess a number from 1 to 1000: "))
# The user has guessed the correct number, inform them and print the number of guesses
print("Congratulations! You guessed the correct number in {} guesses.".format(num_guesses))

It can also answer code questions for simple things or even more complex things, although there are limits to what it can generate. OpenAI has a more specific code generator called Codex.

As amazing this might seem apparently there’s been another AI breakthrough , Deepmind (The folks that brought us AlphaGo , and AlphaFold) have a new apparent model of software development that DeepCode is being hailed as a breakthrouth in writing code at a coding competition

Math and facts like wolfram Alpha vs. chatGPT

Wolfram Alpha is another online computational engine, that’s mostly designed to factual quantitative questions on a whole host of topics, from math, to engineering, money and finance, transportation, statistics etc. Below is a simple comparison of just a few topics answered by both Wolfram Alpha vs. ChatGPT:

Question asked (from Wolfram Alpha site)Wolfram Alpha AnswerchatGPT answer
Nutrition: Analyze nutritional information for a food:
2 slices of swiss cheese
produced nutritional facts boxAccording to the USDA Food Composition Databases, two slices of Swiss cheese (weighing approximately 28 grams) contain the following nutritional information:
Calories: 160
Total fat: 12 grams
Saturated fat: 7 grams
Cholesterol: 45 milligrams
Sodium: 360 milligrams
Total carbohydrates: 1 gram
Protein: 11 grams
Physics: centripetal acceleration, 30mph, 500 ftcentripetal acceleration | 118 cm/s^2 (centimeters per second squared)
= 3.872 ft/s^2 (feet per second squared)
= 9504 mi/h^2 (miles per hour squared)
= 1.18 m/s^2 (meters per second squared)
In your example, if an object is moving at a velocity of 30 mph and has a circular path with a radius of 500 feet, the centripetal acceleration would be calculated as follows:
ac = (30 mph)^2 / 500 ft = 900 / 500 = 1.8 ft/s^2
Statistics: calculate the mean for mean {21.3, 38.4, 12.7, 41.6}28.5The mean, or average, of a set of numbers is calculated by summing the values and dividing by the total number of values. To calculate the mean of the set {21.3, 38.4, 12.7, 41.6}, we would do the following:
mean = (21.3 + 38.4 + 12.7 + 41.6) / 4 = 114 / 4 = 28.5
Jokes: Tell me a physics joke.Q: What did one quantum physicist say when he wanted to fight another quantum physicist?
A: Let me atom!
Why was the physics book sad?
Because it had too many problems.
Math: Solve solve {x + y = 10, x – y = 4}x = 7, y = 3In this case, we can add the equations together to eliminate the y variable:
(x + y) + (x – y) = 10 + 4 2x = 14 x = 7
We can then substitute this value for x in either of the original equations to find the value of y:
x + y = 10 7 + y = 10 y = 3
Therefore, the solution to the system of equations is x = 7 and y = 3. This means that both equations are true when x is equal to 7 and y is equal to 3.
Personal Finance: Compute the time to pay off mortgage $150,000, 6.5%, 30 yearsmonthly payment | $948
effective interest rate | 6.697%

principal paid | $150000
total interest paid | $191317
total payments | $341317 |
Plugging in the values from your example, we get the following:
Monthly payment = (6.5% / 12) x ($150,000 / (1 – (1 + 6.5% / 12)^-30 years)) = 0.00542 x ($150,000 / (1 – (1.00542)^-360)) = $857.32
Comparison samples between Wolfram Alpha and chatGPT

Writing and Poems

Finally in addition to factual based questions which is mostly (but not 100% ) gets right , it can render prose and other written language amazingly well..

For example : “write a short rhyming poem about the beach and the sun”
It produced:

The sun is shining bright

On the sandy beach today

I can feel its warmth As I sit and watch the waves

The water sparkles In the sunlight’s gleam

And the seagulls call As they glide on the breeze

The sand is soft Between my toes

And the salt air Fills my nose

It’s a perfect day At the beach,

it’s true I’m so glad I’m here

To enjoy it with you.

Wrap up and thoughts

As I write this chatGPT is becoming quite a viral sensation, people are amazed, teachers and professors are worried, Google probably has its own variant, etc.

It’s not perfect, it makes silly errors such as when you ask it to multiply $45.07 by 2.13 =95.99 is the correct answer. Yet it replies with $95.65 , because the language model isn’t a pure calculator .

It also fails on some questions for example if you ask it what’s the fastest land bird, it says Peregrine Falcon (that’s the fastest bird) , the fastest land bird is likely the Ostrich (at 45 mph) .. again the language model has led it down the wrong path. This is what you ultimately need to be aware with this model, while the answers seem like AI they’re just. Model find the best response.

At the end of the day the amazing part is just how human like these AI models (Dalle-2, chatGPT) etc. have become… I think the next watershed moment will be when music AI models begin to break through…

2 thoughts on “AI conversational text model, ChatGPT, the end of Google?

  1. Reply Joao lourenco Feb 2,2023 12:08 pm

    I agree , Google search is going to have to change

  2. Reply Joao lourenco Feb 2,2023 12:08 pm

    I agree Google search will need to change

Leave a Reply