Personality Specialisation in Problem Solving

Different people are good at doing different things. Lumberjacks are good at cutting down trees, and nurses are good at taking care of patients. Conversely, lumberjacks are not good at taking care of patients and nurses are not good at cutting down trees.

Such job specialisation according to a person’s predispositions, also occurs in a white collar work environment, just in a slightly more subtle way. Generally speaking, we can identify three primary character types that are suited to different areas of information work.

Analytical predisposition: People who are good at analysing things and drilling down into fine detail. For instance, anyone whose job title includes the word “analyst”.
Controlling predisposition: People who are good at controlling situations and don’t like it when processes run out of control. For instance, project managers.
Energetic predisposition: People who are good at propelling situations along and getting other people motivated and interested in what they are meant to be doing. For instance, (hopefully) the CEO.

If you read this previous article, you might be thinking that the three character descriptions above (analytical, controlling and energetic) sound suspiciously similar to the three general areas that any system can be divided into (logic, control, energy).

It would seem that this correlation is relatively accurate, at least, and it is specialisation in the various areas of the human psyche (of logic, control and energy) that gives rise to the above mentioned character types (analytical, controlling and energetic).

The Need for Accuracy

Accuracy is the most sought-after commodity in a white collar environment. For example, the most likely outcome of a software development project with ill defined requirements is promotion for the non-participants.

Doing things approximately right in a high-pressure environment just doesn’t work, for the same reason that if the clearance between a piston and a cylinder in a petrol engine is greater than ±0.0005 inches, then the engine will either seize up or run inefficiently. Since modern information systems are comprised a large number of moving parts, each of which needs an accurate input to generate an accurate output, therefore if any of the system elements operates out of its allowed variance, the impact on the system as a whole can be dramatic and, in some cases, catastrophic.

In considering how to get an accurate result, the first observation is that accurate system design and implementation can only possibly be the result of accurate thinking.

So the primary question is not necessarily, “In what way must the system be accurate?” but rather, “How do you get accurate thinking in the first place?”

Extreme Positions and Accuracy

Extreme positions in the information work environment may be very annoying, but they do tend to generate accuracy. Meticulous data management, strict adherence to work processes, and rigorous quality checks, always produce a better result than turning-a-blind-eye to “small” issues in data quality control, relaxed work process compliance, and a happy-go-lucky attitude towards quality assurance.

Additionally, extreme work-related positions, are often associated with extreme character types.

That is why:

An extremely analytical person is far more likely to drill down to the level of detailed required to make sure nothing is just “approximately right”, that someone who does not specialise in being analytical.
Somebody who absolutely has to make sure that every single detail is in the right place and under control, is likely to make a brilliant project manager.
Someone who has a fiery desire to make sure a venture is successful, is far more likely to sniff out potential pitfalls than someone who is disinterested and uninvolved.

In this way, professionals who set a high bar and produce clean, accurate results, are very often people whose character dovetails with the analytical, controlling or energising aspect of the work they are doing.

AI Chatbots and Personality Specialisation

When asking ChatGPT or Bard to respond to a question, it is possible to specify the “personality type” of the AI “respondent”.

So if I ask ChatGPT, “As Henry the Eighth, what is your opinion of the O2 building?”

I get the following answer:

Ah, the O2 Building, a marvel indeed! As a monarch of the 16th century, such a structure is beyond my wildest imaginations. In my time, the grandest of edifices were cathedrals, castles, and fortresses, each a testament to the strength and piety of the realm.

Yet here we are, discussing an arena that can house many thousands, all within a dome! A dome, I tell you, the likes of which would dwarf even the grandest hall in Hampton Court Palace. This structure, so named the O2, is said to be dedicated to merriment, music, and sport. A novel idea, I daresay, for in my time, we had to construct grandstands each time a tournament was held!

To consider that man could create such a building is a testament to the astounding progress that has been made. Indeed, the O2 Building would be a fitting symbol of majesty and power, a true wonder of the modern world. I wonder what festivities, what grand tournaments, and what feasts one might host within such a magnificent structure.

That is an amusing application of this AI capability. This AI “personality specialisation” can be used to improve the accuracy of an AI response, however.

Since (as discussed above) very controlling people are more likely to get an accurate handle on things, and very analytical people are more likely to drill down to the required level of detail, and very energetic people are more likely to intuit what could possibly derail a venture, asking an AI chatbot to respond as on of these personality types, can be used to enhance the quality of the chatbot response.

Drawing on the Experience of Industry Leaders

In this previous post, I illustrated how ChatGPT can break down a generally stated web-design problem into discrete components of logic, control and motivation.

Based on the above theory, we can do the same thing again, but this time specify personalities who are well-suited to respond to each discrete problem component. This provides a richer result.

The following question was posed to ChatGPT.

1.3 Background

Solar Based Energy, Inc. (SBE) sells state-of-the-art alternative energy systems utilizing wind and solar power. SBE customers include both individuals and businesses interested in incorporating wind or solar energy sources into either new or existing construction. SBE has identified two trends that they believe will cause explosive growth in the demand for their products.

The first is the continuing energy crisis in the western United States.

The second is the maturation of fuel cell technology which provides a feasible system for storing excess power generation for later use.

SBE sells state-of-the-art alternative energy systems utilizing wind and solar power. SBE customers include both individuals and businesses interested in incorporating wind or solar energy sources into either new or existing construction. SBE has identified several trends that they believe will cause explosive growth in the demand for their products. They include the growing consumer unease with deregulated energy markets, the potential for disruptions to energy imports, and the maturation of fuel cell technology which provides a feasible system for storing excess power generation for later use.

(plus the rest of the problem statement, quoted in the previous post)

As Jeff Bezos, describe the maximum business benefit that could be accrued from the new website.
As Elon Musk, describe the business rules that should be applied to the new website.
As Donald D. Knuth, describe what error handling and exception management should be applied to the new website.

ChatGPT responded as follows.

As Jeff Bezos:

The development of a new website will leverage digital technology to streamline operations and maximize customer reach for SBE. Firstly, by amalgamating the two existing websites into a single platform, we can offer a comprehensive experience to both existing and prospective customers. This could lead to increased customer engagement and potentially boost sales, as customers would now have direct access to product information and ordering.

Moreover, consolidating the two Oracle databases into one will help to maintain data consistency, reduce redundancy, and ultimately lead to better decision-making based on accurate data. Also, by including a language translation feature, we can cater to a global audience, further expanding our market and brand reach. Overall, these improvements could lead to an increase in sales, customer satisfaction, and operational efficiency, resulting in higher profits and competitive advantage in the long run.

As Elon Musk:

Implementing effective business rules for the new SBE website would create a more structured, efficient, and user-friendly environment.

First, there should be business rules to ensure accurate, consistent, and up-to-date product information. This could be automated updates or periodic checks by the product owners.

Secondly, access control rules should be applied to ensure only authorized personnel can modify product information or process orders.

The ordering process itself should follow a set of business rules to guide the user seamlessly through selecting a product to final purchase. This could include business rules around the shopping cart functionality, checkout, and payment procedures.

Additionally, consider business rules to manage customer interactions – such as handling inquiries, complaints, or reviews. These business rules should aim to provide a responsive and supportive customer service experience.

Lastly, remember to include business rules to comply with legal and regulatory requirements, particularly in relation to data privacy and security.

As Donald D. Knuth:

Error handling and exception management are crucial in ensuring the reliability and robustness of any system.

For user input, validation should be done at the client-side to catch and correct errors early. This can range from simple checks like verifying that required fields are filled, to more complex validations like ensuring entered data is in the correct format.

When errors occur, the system should catch them and, when possible, recover gracefully. This could mean rolling back transactions to ensure data integrity or re-trying failed operations. For critical errors, the system should log detailed error information and then alert the appropriate personnel.

To ensure proper exception management, implement a centralized error handling system that provides a unified way to handle and log exceptions. This can greatly simplify the process of diagnosing and correcting issues when they occur.

Lastly, always plan for worst-case scenarios. Implement backup and restore functionality to protect against data loss, and design the system to be resilient even in the face of unexpected failures.

Remember that users should be shielded from the complexity of the system. Any errors or exceptions they encounter should be communicated with clear and user-friendly messages that guide them on what to do next.

These responses compare favourably and are far more accurate and detailed than the answers than the vanilla-flavoured responses that ChatGPT gives if asked the same questions without specifying the responder personalities.

Conclusion

When asking ChatGPT for advice or analysis, specify the respondent persona as the best person you can think of to answer the question.

This technique can be effectively applied using a systematic problem decomposition method, allowing different “people” to address different aspects of the question.