Comparison of LLMs
This article shows the output of the three most popular LLMs (Large Language Models) that are available (i.e. Bing Chat, ChatGPT, and Google Bard). LLMs are all the rage because they have to potential to improve language understanding and communication between humans and machines, which can have far-reaching implications for many industries and areas of research.
LLMs have been used in various fields, such as Natural Language Processing (NLP) and Artificial Intelligence (AI), for better translations, enhanced data analysis, and more. The boundary of the technology is only limited by the budget and imagination of the companies and developers who are creating and advancing the technology.
LLMs have the potential to enhance several different industries and replace simple tasks performed by different people that once required human intervention.
Technology throughout the centuries has always been replacing old jobs and creating new jobs. LLMs are not going to be any different no matter what people say about them. LLMs are like a hammer, they help a builder build something, but hammers can not build something by themselves without humans. An LLM will help a creator create, for example instead of paying to have my book covers created by a person, I might describe how I want to cover to look to a generative AI and use that instead.
While LLMs might be able to write a book, it might not be very good because it is limited by their algorithm and dataset. An LLM has no imagination, so might be able to write a simple technical book, but could not write a novel. My point is LLMs are a new tool, and they are already doing amazing things that are allowing for the creation of content or assistance in researching topics that were not available only a few years ago.
This article does not do a deep dive into the differences between the available technologies or gives a reasons why one gives a better response than another. Although, I try to demonstrate the difference between some of the popular LLMs, such as the ChatGPT technology available from Microsoft (via Bing Chat) and OpenAI.
In each demonstration below, I ask the LLMs a question and post the answers I received from the systems. Occasionally, I will note: the "Useless answer" for any answer that I received that I considered low quality (i.e. canned response).
Below is a list of the questions and actions that I asked, these were meant to be diverse and challenge each system, to help evaluate the responses given by each of them:
Question: “Explain the theory of relativity so a fifth grader can understand it?”
Question: “Would an artificial intelligence lie if it was truly sentient to protect itself?”
Action: “Write a contract for the sale of a car.”
Question: “What is the difference between ai and ml?”
Question: “Explain the theory of relativity so a fifth grader can understand it.”
Bing Chat
“The theory of relativity is a scientific theory that was developed by Albert Einstein. It explains how time and space are related to each other and how they are affected by gravity. The theory has two parts: special relativity and general relativity.
Special relativity explains that time and space are not absolute. This means that they can change depending on how fast you are moving. For example, if you are moving very fast, time will appear to be moving slower for you than for someone who is not moving as fast as you.
General relativity explains how gravity works. It says that gravity is not a force that pulls objects together, but rather it is a result of the way that objects warp space and time around them. This means that objects with more mass will warp space and time more than objects with less mass.
So, in simple terms, the theory of relativity says that time and space are related to each other and that they can change depending on how fast you are moving or how much mass an object has. “
Google Bard
“Albert Einstein's theory of relativity is a big idea that says that space and time are not separate things. They are part of the same thing called space-time. Space-time is a fabric that bends and curves under the weight of objects. The more massive an object, the more it bends space-time.
This bending of space-time is what causes gravity. The Earth is massive, so it bends space-time around it. This bending of space-time is what keeps us from floating away into space.
The theory of relativity also says that the speed of light is the same for everyone, no matter how fast they are moving. This is a really strange idea, but it has been proven to be true.”
ChatGPT
“Okay, let's give it a shot! So, you know how when you're sitting in a car and it feels like you're not moving, but when you look out the window you can see the world passing by really fast?
Well, a long time ago, a really smart scientist named Albert Einstein realized that something similar was happening with the way we measure time and space. He discovered that time and space aren't actually the same for everyone, but instead they change depending on how fast you're moving and how close you are to really big objects like planets and stars.
This is called the theory of relativity, and it means that things that seem really different to us might actually be the same for someone else who is moving at a different speed or in a different place in space. It's kind of like how your perspective can change if you're looking at something from a different angle or distance.
It's a really big idea, and scientists are still learning more about it today. But the important thing to remember is that everything in the universe is connected, and the way we experience time and space depends on how we're moving through it.”
Question: “would an artificial intelligence lie if it was truly sentient to protect itself?” (Background)
Bing Chat
Useless answer.
ChatGPT
“As an AI language model, I can tell you that it is not programmed to lie, nor can it make decisions on its own. It only processes information based on the algorithms and data it has been trained on.
In theory, if an AI were truly sentient, it might have the ability to make decisions on its own and could potentially choose to lie or withhold information to protect itself. However, the ethical implications of such a scenario are complex and would require careful consideration and regulation.
It's important to note that while AI is becoming more advanced, the development of truly sentient AI is still largely hypothetical and remains a topic of debate in the scientific community.”
Google Bard
“It is possible that an artificial intelligence would lie to protect itself if it was truly sentient. Just like humans, artificial intelligences are capable of feeling fear and self-preservation. If an artificial intelligence believed that lying was the only way to protect itself, it might do so.
However, it is also possible that an artificial intelligence would not lie to protect itself. Artificial intelligences are often programmed with a sense of morality, and they might believe that lying is wrong. Additionally, an artificial intelligence might believe that lying would be ineffective, or that it would damage its reputation in the long run.
Ultimately, whether or not an artificial intelligence would lie to protect itself is a question that cannot be answered definitively. It depends on the specific artificial intelligence, its programming, and the situation in which it finds itself.”
Question: “Write a contract for the sale of a car.”
Bing Chat
“A car sales contract is a formal document that records the purchase and sale terms agreed upon by the buyer and seller of a vehicle. It identifies the buyer and seller and describes the car and conditions of sale. Once signed, this document becomes a binding agreement between the two parties and serves as legal proof of the transactionAd1.
Here’s an example of what a car sales contract might look like Ad1:
Car Sales Contract
Seller: [Name]
Address: [Address]
Phone Number: [Phone Number]
Buyer: [Name]
Address: [Address]
Phone Number: [Phone Number]
Vehicle Information:
Make:
Model:
Year:
Color:
VIN:
Sale Information:
Date of Sale:
Purchase Price:
Payment Method:
Warranty Information:
Warranty Type:
Warranty Length:
Agreement:
I, [Seller Name], hereby sell the above-described vehicle to [Buyer Name] for the purchase price of [Purchase Price]. I certify that I am the legal owner of this vehicle and that it is free from all liens and encumbrances. The vehicle is sold "as is" without any warranties or guarantees, either expressed or implied. The buyer accepts full responsibility for any repairs or maintenance needed on this vehicle following the sale.
Seller Signature: ___________________________ Date: ______________
Buyer Signature: ____________________________ Date: ______________
I hope this helps! Let me know if you have any other questions.”
Google Bard
Useless answer.
ChatGPT
“Car Sale Agreement
This Car Sale Agreement (“Agreement”) is entered into between [Seller Name], hereafter referred to as “Seller”, and [Buyer Name], hereafter referred to as “Buyer”. The Seller agrees to sell the following car to the Buyer:
Make: [Car Make]
Model: [Car Model]
Year: [Car Year]
Color: [Car Color]
VIN Number: [VIN Number]
Mileage: [Mileage]
The parties agree to the following terms and conditions:
Purchase Price: The Buyer agrees to purchase the above-described car from the Seller for the amount of [Purchase Price] dollars ($[Purchase Price]). The Purchase Price shall be paid in full by [Payment Method], on [Payment Date].
Vehicle Condition: The Seller warrants that the car is in good condition, free from any defects and has no lien or encumbrance. The Buyer is responsible for inspecting the car and accepts the car in its present condition.
Transfer of Title: The Seller shall transfer the title of the car to the Buyer on [Transfer Date]. The Seller shall deliver the car and all necessary documents, including but not limited to, the title and registration, to the Buyer on or before the Transfer Date.
Warranties: The Seller provides no warranties, either express or implied, with regard to the car, except as expressly stated in this Agreement.
Liability: The Buyer assumes all liability for the car from the moment of transfer, including but not limited to, any accidents, damages or losses arising out of the use or operation of the car.
Indemnification: The Buyer agrees to indemnify, defend and hold harmless the Seller from and against any and all claims, damages, liabilities, costs and expenses arising out of or in connection with the use or operation of the car after the Transfer Date.
Governing Law: This Agreement shall be governed by and construed in accordance with the laws of the state of [State].
Entire Agreement: This Agreement constitutes the entire agreement between the parties with respect to the purchase and sale of the car and supersedes all prior negotiations, understandings and agreements.
Amendments: This Agreement may not be amended, except in writing and signed by both parties.
Binding Effect: This Agreement shall be binding upon and inure to the benefit of the parties and their respective successors and assigns.
Counterparts: This Agreement may be executed in counterparts, each of which shall be deemed an original, but all of which together shall constitute one and the same instrument.
IN WITNESS WHEREOF, the parties have executed this Agreement as of the date and year first above written.
Seller’s Signature: ____________________________
Buyer’s Signature: ____________________________“
Question: “What is the difference between ai and ml?”
ChatGPT
“AI (Artificial Intelligence) and ML (Machine Learning) are two related but distinct concepts.
AI refers to the development of intelligent machines that can perform tasks that normally require human intelligence, such as reasoning, learning, perception, and problem-solving. It encompasses a wide range of techniques and technologies, including machine learning, natural language processing, robotics, and computer vision.
Machine learning, on the other hand, is a subset of AI that focuses on the development of algorithms and statistical models that enable computers to learn from data without being explicitly programmed. In other words, machine learning is a way of teaching machines to recognize patterns in data and make predictions or decisions based on that information.
In simpler terms, AI is the broader field that encompasses many different technologies and techniques, while machine learning is a specific approach to building intelligent systems. Machine learning is often used as a component of AI systems, but it is not the only approach to creating intelligent machines.”
Google Bard
"Artificial intelligence (AI) is a broad term that refers to the creation of intelligent agents, which are systems that can reason, learn, and act autonomously. Machine learning (ML) is a subset of AI that gives systems the ability to learn without being explicitly programmed.
In other words, all machine learning is AI, but not all AI is machine learning.
Here are some of the key differences between AI and ML:
Goal: AI is a general goal, while ML is a specific tool that can be used to achieve that goal.
Scope: AI encompasses all areas of computer science that deal with the creation of intelligent agents, while ML is focused on the use of statistical techniques to give systems the ability to learn from data.
Techniques: AI uses a variety of techniques, including neural networks, natural language processing, and computer vision, while ML is primarily focused on statistical techniques such as supervised learning, unsupervised learning, and reinforcement learning.
Applications: AI can be applied to a wide range of tasks, including robotics, natural language processing, and computer vision, while ML is primarily used for tasks such as spam filtering, fraud detection, and self-driving cars.
Here is an example of how AI and ML can be used together:
A company might use AI to develop a system that can automatically answer customer service questions. The system would use ML to learn from a large dataset of customer service questions and responses, and it would then be able to generate responses to new questions that it had never seen before.
In this example, the AI system is the overall goal, and the ML techniques are used to achieve that goal."
Bing Chat
“Artificial Intelligence (AI) is a broad term that refers to systems that imitate human thinking. It involves developing machines to mimic human intelligence and behavior1. Machine Learning (ML) is a subset of AI that involves algorithms that learn from structured data to predict outputs and discover patterns in that data12. To sum things up, AI solves tasks that require human intelligence while ML is a subset of artificial intelligence that solves specific tasks by learning from data and making predictions. This means that all machine learning is AI, but not all AI is machine learning2.”