Something Topical - ChatGPT

Status: March 2023 - Musing published

Given this posting is on "soapbox", the preface "Something Topical" could likely apply to any post. I am keeping it as, as this thread came to mind while working on something else. So better to "compartmentalise" this rather than let it bleed into another post.

The Context

I have been writing an article on the value of architcture "models", using some work I contributed to the LFN/GSMA Anuket Reference Models as examples. One of "topics" was about the value of models. This turned my mind to questions around: "value", "work", "effort" and "reward". These are perennial favourite questions when deciding to "do something".

Following the thought:

• the next questions was "is this a question" worth answering ?
• leading to, "Is this valuable to (1) me (i.e. I just enjoy these musing and so will persue on the basis of intrinsic satisfaction) or (2) am I doing this to satisfy some external or indirect goal (i.e. I will get paid to complete this research or I need this as input into some further endevour)"
• the first (1) would be for fun, while the second (2) might be seen as paid "work".
• another "meme" popped up on "work", "work smarter, not harder" or "could I get paid to have fun"  so I would get (1) and (2) done by "killing two birds with one stone". The problem with having fun at work is that it can lead to others in your family being unhappy, as you don't leave the office (or study in post COVID world). Maybe the word "balance" should come into this conversation ?
• So I needed to ensure that I worked "smarter and not harder" and "delegated more".  In this context (I have been doing an AI proof of concept ("an5 - Intelligent Network Design")), it made me think, "can I delegate this question to ChatGPT?".

So I delegateed to "ChatGPT":

So this was a pretty well structured response to a simple question, though it was heavy on "stakeholders" (a sign that this is something one does for others to consume) and said the model was ("acts as") a "visual representation" which is a misunderstanding evident in lots of architecture material, where "model" and "diagram" are used synomously. It also failed to mention "prediction". A model can do "prediction" but a "diagram" cannot.

NOTE: Here "do" means the model directly creates the prediction/s, rather than provides input into the person (stakeholder) consuming it, who uses model to gain some insight and makes a prediction/s based on this (so model is supporting the activity rather than completing it).

And more broadly

Now my quick summary of the implications of "ChatGPT" was that:

• in the short term "asking" questions was more valable than "answering" questions, as ChatGPT has commodised the "answering"

My next thought was, that my daughter has a great future ahead of her as she is always always asking me question, some of which are not so easily answered ;-) .

So following on:

• being able to interpret the "answers" is also valuable, irrespective of whether it was provided by a chat bot or not and
• applying and using the answer for a targetted purpose was also high value.

So what does ChatGPT think ?

Oh and just to be be sure, I should get affirmation on this approach to "working smarter not harder".

The Conclusion ..

An illusatrative semantic web of my ChatGPT session:

Did I kill two birds with one stone ?

Did I get my work answer and enjoy myself ?

ChatGPT was fun, initially, but it is clear that using it would very quickly become work again, once the novelty had worn off.  The responses provided a framework for consideration and elaboration. The structure is essentially what you learn in secondary school and issues based consulting.

Ultimately ChatGPT responds and acts like a robot, so beyond the initial pleasant surprise, it is a tool. There was no evidence of humor (though there are joke writing ChatGPT trails out there) or subtle consideration, which are characteristics of human responses.

My kids will still ask questions and I will continue to try to answer them and when unable direct them to either do some research or invite them to discuss with someone else.

As to "working smarter", I find that formulating the strategy is relatively quick, but documenting and using the result has much longer latency (ie "lag" in model above). With ChatGPT the "lag" can be shortened by providing structure on the supposition being addressed.

Having considered one question, the next question from my daughter that comes to mind was, "Dad would you rather be the funniest person or the smartest person in the world?".

My response: "Usually smart peope have a good sense of humor.".

Yes, a gross generalisation to be sure, so maybe she will have to dig into that one.

For example, "What if the smartest person in the world was living in circumstances   of massive inequality and difficulty?".

Another "soapbox" subject.  

References & Links:

"What is implication of chatGPT ?" - I seleced this response,  "ChatGPT Is a Tipping Point for AI" by Ethan Mollick ,  as it come from "Harvard Business Reivew" (i.e. past the credibility sniff test), but did not really address my question...

"work smarter not harder" - yep we can all try to achieve this management speak "nonsense" result, but the problem is that if it is "work" then it is generally "hard", maybe we should start saying "more fun less work" ;-) . One of the suggested tips in this link from "Forbes" under the tag of "leadership/careers" is to "delegate more".

"kill two birds with one stone" - an idiom, that we like as it makes us productive

an5 - Intelligent Network Design - an effort to use AI to work smarter not harder and delegate to machine

ChatGPT - a "chat bot" by a company called "OpenAI", where GPT == Generative Pre-Trained Transformer" which I think mean that you feed the result of its "generation function" into its "training" function (or putting things into my AI context, becomes" use AI "results generator" to generate lots of outcomes and then apply a scoring mechanisms to these which in turn can be used for hueristic pruning of the search space...). The result is that the learning is done "offline" using huge result space, while the "question" can be done "online" (ie conversationally with user) by apply search tree pruning to allow fast response.

Anuket Reference Model - The Reference Model (RM) specifies a virtualisation technology agnostic (VM-based and container-based) cloud infrastructure abstraction and acts as a “catalogue” of the exposed infrastructure capabilities, resources, and interfaces required by the workloads.

Appendix A - ChatGPT Dialog

What is the value of architecture model ?

What is the value of architecture model ?

An architecture model is a representation of a system's structure and components, which helps stakeholders to understand the system's design, functionality, and behavior. The value of an architecture model is significant, as it provides several benefits, including:

1. Communication: Architecture models help stakeholders to communicate and collaborate effectively. The model acts as a visual representation of the system, which allows stakeholders to discuss and exchange ideas, ensuring everyone understands the system's design and behavior.
2. Analysis: Architecture models help stakeholders to analyze the system's design and behavior. By examining the model, stakeholders can identify potential issues, bottlenecks, and performance problems, which can be addressed during the development phase.
3. Decision-making: Architecture models help stakeholders to make informed decisions. By examining the model, stakeholders can evaluate various design options, trade-offs, and alternatives, and select the best approach based on their needs and requirements.
4. Maintenance: Architecture models help stakeholders to maintain the system over time. The model acts as a reference point for the system's design, ensuring that any changes made to the system are consistent with the original design.
Overall, an architecture model is a valuable tool for stakeholders involved in the design, development, and maintenance of a system. It helps stakeholders to communicate, analyze, make informed decisions, and maintain the system over time.

What is the implication of

What is implication of chatGPT ?

ChatGPT is a large language model developed by OpenAI that uses deep learning techniques to understand and generate human-like text. The implications of ChatGPT are significant, as it has the potential to revolutionize various industries and domains. Some of the implications of ChatGPT include:

1. Natural Language Processing: ChatGPT is a major breakthrough in the field of natural language processing. It can understand and generate human-like text, making it useful in various applications, such as chatbots, customer service, and language translation.
2. Automated Content Creation: ChatGPT can generate high-quality content on various topics, including news articles, product descriptions, and marketing copy. This has implications for content creators, publishers, and marketers, as it can automate the content creation process, saving time and resources.
3. Personalization: ChatGPT can be trained on user data, allowing it to personalize its responses based on individual preferences and behaviors. This has implications for personal assistants, virtual agents, and other applications that require personalized interactions.
4. Research and Development: ChatGPT can be used to generate hypotheses, simulate experiments, and analyze data. This has implications for scientific research and development, as it can automate and accelerate the discovery process.
Overall, the implications of ChatGPT are significant, as it has the potential to revolutionize various industries and domains, including natural language processing, content creation, personalization, and research and development.

Does this mean that answering questions became a low value activty ?

Does this mean that answering questions became a low value activty ?

No, answering questions is still a valuable activity, especially when it requires critical thinking, expertise, and context. While ChatGPT can generate human-like responses, it is not a substitute for human intelligence and understanding. ChatGPT can assist in answering questions and providing information, but it is still important to have human input to validate the accuracy, relevance, and reliability of the information. In some cases, answering questions may even become more valuable, as ChatGPT can generate a large amount of information, which may require human input to distill the most relevant and useful insights. Therefore, while ChatGPT has significant implications for various industries and domains, including natural language processing and research, it does not devalue the importance of human intelligence, expertise, and context.

Is asking questions more valuable than answering questions ?

Is asking questions more valuable than answering questions ?

Both asking and answering questions are valuable activities, and they play complementary roles in the learning process.

Asking questions is valuable because it helps to stimulate curiosity, promote critical thinking, and drive the search for new knowledge and insights. Asking questions can also help to clarify confusion, fill knowledge gaps, and deepen understanding.

Answering questions is valuable because it allows us to share our knowledge and expertise with others, promote collaboration and learning, and build trust and credibility. Answering questions can also help to develop our communication skills, improve our problem-solving abilities, and foster a sense of community.

Ultimately, asking and answering questions are interconnected and interdependent activities, and both are important for individual and collective learning and growth. Asking questions can inspire new ideas and perspectives, while answering questions can contribute to the dissemination and advancement of knowledge. Therefore, neither asking nor answering questions is inherently more valuable than the other; they are both essential components of the learning process.

Who is best placed to  pose questions ?

Who is best placed to  pose questions ?

Anyone can pose questions, regardless of their background or expertise. Asking questions is a natural part of the learning process, and it is an effective way to gain knowledge, insights, and understanding.

That being said, the best way to pose questions depends on the type of question and the context in which it is being asked. For example, if you have a specific question about a particular topic, it may be best to pose your question to a subject matter expert. On the other hand, if you have a general question, you can pose it to a search engine or an online community.

In general, the best way to pose a question is to be clear, concise, and specific. It's important to provide enough context for the person answering the question to understand what you are asking and to provide an accurate and relevant answer. Additionally, it's helpful to be open-minded and receptive to different perspectives and insights.

Overall, anyone can pose questions, and there are many resources available to help you find the answers you are looking for. The most important thing is to be curious, ask questions, and continue learning.