Last week I received a newsletter from Conor Grennan – The Infinite Focus Group: AI In Marketing he explained how to use AI to create product feedback and it's a fascinating concept. He frequently writes about how to approach AI with the right mindset, rather than learning a multitude of prompts. I believe he's onto something, but I also find it incredibly useful to observe how others are utilising AI and then consider how we might adapt those ideas to the world of finance. This week, I take a lesson from Conor's approach and created a focus group of Chief Executives asking them to provide candid feedback to the Prudential Regulatory Authority Dear CEO letter UK Deposit Takers Supervision: 2024 priorities. To be clear these are ai created CEOs. Method behind the madness To write this letter I worked with ChatGPTin a step-by-step process and thereby combined human and AI skills together. The steps were: Identifying the Task: The initial request was to explore how CEOs would respond to a letter from the PRA outlining its 2024 priorities for UK deposit takers. This exploration was aimed at understanding the potential private sentiments and strategic considerations of these CEOs under the umbrella of Chatham House rules. Generating CEO Responses: To realistically simulate these responses, ChatGPT dissected the PRA's letter, identifying key themes such as governance, risk management, credit risk, and operational resilience. Then together we created three distinct CEO profiles—small, medium, and large deposit takers—each with its unique set of challenges and perspectives. Simulating a Roundtable Discussion: Building on an initial response, we expanded the scope to include a broader array of voices, simulating a roundtable discussion involving 15 CEOs. This step was designed to capture a diversity of opinions and strategic priorities reflective of the entire sector. Analysing Themes and Sentiments: From the individual responses, we identified common themes and underlying sentiments. This analysis revealed concerns about resource constraints, the need for more tailored guidance, the challenges of operational and technological resilience, and a desire for closer engagement with the PRA, among others. Drafting the Collective Letter: Using the insights gained, we drafted a letter from the collective of CEOs to the PRA. This letter candidly articulated the sector's feedback on the 2024 priorities, highlighting areas of concern and proposing constructive suggestions for how the PRA could further support banks in meeting these challenges. This summary not only highlights the steps taken to create a dialogue between the banking sector and its regulator but also underscores the potential of using creative simulations to gather insights and foster constructive conversations within the financial industry. If you have a ChatGPT subscription this is my full discussion This is the collective synthetic response Draft Letter from CEOs to the PRA Dear David Bailey and Laura Wallis, We, a collective of CEOs from across the UK deposit-taking sector, appreciate the opportunity to respond to the Prudential Regulation Authority's (PRA) 2024 priorities. This letter reflects a unified stance, capturing diverse perspectives from small, medium, and large institutions. We commend the PRA's comprehensive approach and share the commitment to maintaining a resilient, robust financial system. However, we wish to provide honest feedback and constructive suggestions on areas where we believe the PRA could further support the sector in meeting these priorities. Resource Constraints vs. Regulatory Expectations There is a palpable concern among smaller and medium-sized banks regarding the challenges of meeting regulatory expectations with limited resources. The detailed and complex nature of the requirements, especially around operational resilience and technological innovations, presents significant hurdles. We suggest the PRA could offer more tailored guidance and support mechanisms, such as forums or shared resources, to help smaller institutions navigate these challenges more effectively. Operational and Technological Resilience The focus on operational resilience is crucial, yet the scale of compliance varies significantly across the sector. While larger banks have the resources to invest in sophisticated IT infrastructure, smaller banks struggle with the same expectations. More practical steps and examples of how to achieve these goals, considering the size and scale of operations, would be immensely beneficial. Additionally, a clear framework for assessing and mitigating cyber threats, applicable to banks of all sizes, would be appreciated. Credit Risk Management in an Uncertain Economy The economic landscape presents a universal challenge, with the management of credit risk at the forefront. Enhanced dialogue around balancing risk management with supporting growth would help banks navigate these uncertainties. The PRA could also facilitate peer sharing of best practices in credit risk management to foster a more resilient banking sector. Climate-related Financial Risks While there is a sector-wide commitment to addressing climate-related financial risks, there is also a request for more specific guidance. Detailed case studies, best practice sharing, and workshops could help banks at various stages of integrating climate risk management into their frameworks. Talent and Succession Planning For smaller banks, attracting and retaining executive talent is a significant challenge. The PRA could support by highlighting pathways for leadership development within the sector and facilitating mentorship programs that ensure the continuity of strong governance and risk management practices. Data and Model Risk Management The pace of change and the volume of data present ongoing challenges. We propose the PRA could provide clearer guidelines on data governance and model risk management, perhaps through a series of workshops or consultation sessions, to help banks better navigate these areas. Engagement with Regulators A recurring theme is the desire for more open, ongoing dialogue with the PRA. Regular roundtable discussions or feedback sessions on emerging regulations could enhance understanding and compliance across the sector. Such engagement would be particularly valuable in addressing novel risks and ensuring alignment on regulatory expectations. Conclusion In closing, we wish to emphasize our collective dedication to meeting the PRA's priorities and contributing to the stability and resilience of the UK financial system. We believe that through constructive dialogue and collaboration, we can address the challenges identified and work together towards our mutual goals. We look forward to the PRA's support and guidance in these areas and to establishing a more interactive and supportive regulatory relationship. Yours sincerely, [Signatures of the CEOs] This letter aims to constructively articulate the collective feedback and suggestions from the CEOs of small, medium, and large deposit-taking institutions, focusing on areas where the PRA could provide additional support to facilitate compliance and promote a resilient banking sector. The Synthetic versus The Real World The approach of using AI to simulate responses in a focus group setting, like the one we've created with synthetically generated CEO perspectives, offers a unique and efficient way to anticipate and analyse complex reactions to regulatory changes or other significant industry developments. However, when considering how closely these synthetic responses might align with real-world feedback from CEOs, several factors come into play. Real CEOs bring a depth of experience, intuition, and personal judgment to their decision-making processes that are shaped by years of industry involvement, personal experiences, and a nuanced understanding of their specific business context. While AI can generate responses based on a vast dataset and identify patterns or sentiments that are likely to emerge, it might not fully capture the individual complexities and the depth of strategic thinking unique to each human CEO. Real-world feedback is influenced by the current economic climate, geopolitical tensions, recent market events, and emerging industry trends. Although AI incorporates a broad range of data up to its last training update, its ability to factor in the very latest events or subtle shifts in market sentiment is limited by its training cut-off point. CEOs' decisions and feedback are not solely based on data and strategic considerations; they also reflect ethical values, emotional intelligence, and personal leadership styles. These aspects can profoundly influence how CEOs perceive and respond to regulatory changes. AI-generated responses may not fully encapsulate these human elements, which add significant depth and variability to real-world feedback. CEOs operate within the context of their organisations' cultures and are influenced by a wide range of stakeholders, including employees, customers, shareholders, and regulatory bodies. Their responses to regulatory changes are often crafted to align with these varied interests. While AI can simulate responses based on generalised scenarios, capturing the unique corporate culture and specific stakeholder dynamics of each firm would be challenging. AI-generated responses can be remarkably insightful, especially in identifying broad trends and general sentiments. They can provide a useful approximation of how CEOs might respond to certain stimuli, acting as a cost-effective and efficient way to gather initial insights. However, the predictive accuracy of these simulations in matching the exact responses of real CEOs under similar conditions would inherently have limitations. While AI-generated focus groups offer valuable perspectives and can serve as a powerful tool for scenario planning and strategic insights, they should be viewed as a complement to, rather than a substitute for, engaging directly with real-world executives. The unique value of these exercises lies in their ability to highlight potential areas of concern, generate novel ideas, and stimulate further discussion among actual decision-makers, rather than precisely predicting specific feedback from individuals. Regulators Could Benefit The Prudential Regulation Authority could significantly benefit from utilising a feedback process similar to the one we discussed before finalising and issuing guidance letters. This approach presents several key advantages. By simulating responses from a diverse array of banking sector CEOs before finalising their letters, the PRA can gain insights into how their guidance might be received across the spectrum of institutions they regulate. This could lead to adjustments in their recommendations to ensure they are not only ambitious in terms of regulatory standards but also practical and achievable across different scales of operation. Pre-emptive feedback can highlight areas where additional clarity is needed, ensuring the final letter addresses real-world applicability and implementation challenges more effectively. Simulated feedback processes would enable the PRA to understand better the current economic, operational, and competitive pressures facing banks. This nuanced understanding can make the regulatory guidance more responsive and sensitive, ensuring that expectations are aligned with what is feasible and reasonable under current conditions. Engaging in a synthetic feedback process, even internally, underscores a commitment to a collaborative approach to regulation. It signals to regulated entities that the PRA is considerate of their operational realities and is seeking to issue guidance that supports not just compliance but also the sector's overall health and competitiveness. This can strengthen the relationship between the regulator and the regulated, promoting a more open, trust-based dialogue. Is this of Use to You? Synthetic focus groups, constructed through advanced AI simulations, represent a new approach for businesses seeking to improve their decision-making and innovation. Synthetic focus groups enable companies to test scenarios and strategies in a controlled environment, providing valuable foresight into potential challenges thereby offering the potential to enhance strategic planning and decision-making processes. For product development, synthetic focus groups can offer a wealth of insights into consumer preferences, expectations, and aversions. By generating diverse consumer perspectives on new products or services, companies can refine their offerings, tailor their marketing strategies, and innovate with greater confidence in alignment with market needs and trends. Synthetic focus groups can also simulate customer reactions to changes in service, pricing, or policies, providing businesses with a nuanced understanding of what drives customer satisfaction and loyalty. This insight allows companies to address issues, personalise customer interactions, and build stronger relationships with their target audience. For businesses looking to enter new markets or expand their presence, synthetic focus groups can simulate the responses of different market segments, including potential customers, competitors, and regulators. This enables businesses to anticipate challenges, understand competitive dynamics, and design market entry strategies that are informed. Companies facing stringent regulatory environments or aiming to enhance their social responsibility efforts can use synthetic focus groups to gauge the impact of their compliance strategies and CSR initiatives. This helps ensure that their actions not only meet regulatory requirements but also resonate positively, reinforcing the company's commitment to ethical practices and societal well-being. What’s happened is that AI has provided us with new tools that can handle problems differently. In this case, we can craft policy before it’s written so it’s more in tune with the target audience and therefore more effective. What is striking is the speed and low cost of undertaking detailed research, it has the potential to improve aspects of business that in the past would have needed expensive consultants. Thank you, Conor, for your insight that triggered my interest here. I’m a great believer that AI is about experimentation and practice, you don’t need to be an IT expert just curious and willing to give it a go.

In this article, I ask whether Claude, the large language model from Anthropic, should be given a seat in the boardroom. After all, Claude is well-versed in worldly knowledge and is ready to engage on any topic that comes his way. So let's jump in and see what Claude brings to the table. The story is like this: the treasury wants to widen its ability to place deposits within the banking system, primarily because it believes that the interest paid by bank counterparties is higher than that offered by the Bank of England. The argument is that given the regulatory environment we currently operate in, the major banks are all safe and offer very little in the way of credit exposure, and therefore, this is more or less a free lunch. There's little in the way of executive pushback because the perception is that, in the event of a bank running into trouble, the Bank of England will have no choice but to provide support. The proposal goes through without a lot of discussion. Common enough, but is it sensible? It depends, and the answer lies more in the way of thinking rather than the decision itself. Let's prompt Claude and see what transpires: Prompt: The Board of a small Building Society has been asked to approve the extension of credit lines to five banks. Specifically, the credit lines will be for simple deposits with the five largest UK banks, and utilization will be by the Treasury. The proposal is for a maximum of £10m per bank. Currently, the Society holds all its liquidity at the Bank of England. The Society has a balance sheet of £1,000m and capital of £100m, and they are required to maintain a 10% capital ratio. Their calculations show that there is £3m of capital surplus. Focusing only on the credit aspects of this proposal, is it sensible? Following a lengthy discussion, I asked Claude to summarise things in a way that would be useful for the board, and here it is: Extending Credit Lines to the Five Largest UK Banks: Risk Assessment and Regulatory Considerations The proposal to extend credit lines to the five largest UK banks appears to be relatively low risk, but there are still some important considerations The five largest UK banks are generally considered to be financially stable and have strong credit profiles, with high credit ratings from major rating agencies, reflecting their low probability of default. However, concentration risk remains a concern as the exposure is limited to the banking sector and the five largest players. A systemic issue affecting the UK banking sector could expose the Building Society to increased credit risk. The proposed credit lines total £50m, representing 5% of the Building Society's balance sheet. While this may seem relatively small, it could become a significant source of stress during a severe crisis. The Building Society's current capital surplus of £3m above its required 10% capital ratio may not be sufficient to absorb potential losses from the proposed credit lines in a severe stress scenario, especially if losses are incurred simultaneously on other parts of the balance sheet. Considering the stress scenarios of 1974 and 2008 The 1974 secondary banking crisis suggests that the proposed credit lines to the five largest UK banks could be relatively resilient, as the largest banks were not as severely affected as smaller institutions. However, the 2008 global financial crisis highlights the potential for significant credit risk in a severe banking crisis. Even large, systemically important banks faced difficulties, with some requiring government support. Since the 2008 crisis, UK banks have been required to hold more capital and liquidity to improve their resilience to market stress. The regulatory environment has also been strengthened to reduce the likelihood and impact of bank failures, which could help mitigate the potential impact on the Building Society's credit lines in a stress scenario. The proposed credit lines to the five largest UK banks could potentially lead to a serious strain on the Building Society's capital position, if such severe events were to occur once or twice in a 50-year period Historical evidence suggests that severe banking crises can occur once or twice in a 50-year period, as seen with the 1974 secondary banking crisis and the 2008 global financial crisis in the UK. In a crisis similar to 2008, the potential for default by one or more of the five largest UK banks could lead to substantial losses for the Building Society. The Building Society's current capital buffer of £3m may not be sufficient to absorb the potential losses from the proposed credit lines in a severe stress scenario, especially if losses are incurred on other parts of the balance sheet simultaneously. Adopting lower limits for the proposed credit lines could be a prudent approach Lower limits would reduce the Building Society's overall exposure to the five largest UK banks, minimizing potential losses in the event of a default during a crisis. By reducing the size of the credit lines, the Building Society would be better positioned to preserve its capital buffer. A phased approach, starting with lower limits and gradually increasing them over time, would allow the Building Society to monitor risks and adjust its strategy as needed. The Building Society could consider setting initial limits of £5m per bank, resulting in a total exposure of £25m (2.5% of the balance sheet), which would provide some diversification benefits while reducing the potential strain on the capital position in a stress scenario. Five questions the regulator may ask How does the Building Society define and quantify its risk appetite for credit risk, specifically in relation to exposures to other financial institutions? What is the process for setting credit line limits, and how are these limits aligned with the Building Society's risk appetite? What risk management processes and systems are in place to monitor and control exposures against the established limits? What is the Board's role in overseeing the risk appetite framework and the setting of credit line limits? How does the Building Society ensure compliance with regulatory requirements and guidelines related to credit risk management and large exposures? The really important questions for the Board How do you ensure that the Building Society's risk appetite for credit risk is properly aligned with its strategic objectives and capital management plan? As NEDs, how do you challenge and validate the assumptions and methodologies used in setting the credit line limits, and how do you satisfy yourself that these limits are appropriate given the Building Society's risk profile? What level of oversight and scrutiny does the Board apply to the monitoring and reporting of credit exposures against the established limits, and how do you ensure that any breaches or deviations are promptly brought to your attention? How do you balance the need to diversify the Building Society's liquidity holdings with the imperative to maintain a prudent and sustainable risk profile? How have you determined that the proposed credit line limits are consistent with the Building Society's risk appetite and capital management objectives, and what analysis has been performed to support this conclusion? In conclusion, while the proposal to extend credit lines to the five largest UK banks appears to be relatively low risk, the Board must carefully consider the potential impact of severe stress scenarios on the Building Society's capital position. Adopting lower initial limits and gradually increasing them over time could be a prudent approach to balance diversification benefits with capital preservation. The Board should engage in robust discussions, challenge assumptions, and ensure that the risk appetite framework and credit line limits are properly aligned with the Building Society's strategic objectives and regulatory requirements. Claude has facilitated high-level thinking, and this is different from the completion of boring, repetitive work that LLMs are considered to be good at. I would venture to say that, with the aid of an LLM, you can work through the key aspects of almost any part of your business without having expert knowledge of the topic. What you do need is an understanding of how to prompt and push the LLM, and this comes with practice. My venture with Claude has also made me question whether the fear of using LLMs is overdone. There's a lot of talk about the risk of using large language models in the context of proprietary data. That's why many firms are very reluctant to let their employees near these models, fearing that something could go wrong. But what I've shown here is that you can have a very meaningful discussion without providing any proprietary information to the model. There are a whole host of risks in financial institutions that lend themselves extremely well to this exercise at a high level. Almost every board paper that I have seen can be repurposed into something that could be put into a large language model without disclosing any information that is not already in the public domain. Senior executives who do not use these models to question and analyse the risks and strategies that their firms are engaged in are missing out on a very powerful advocate. Should Claude be in the boardroom? Well, not as a voting member of the board, but definitely at the table, discussing strategy and risks, he makes a cheap and efficient consultant.

When I was at school, I was fortunate enough to have an excellent English teacher, Tim Cornish. What made him exceptional was his ability to bring literature to life for a group of schoolboys, which is indeed a gift. One of the poets he introduced us to was Wilfred Owen. This has led me to a lifetime interest in the work of this great poet. He could articulate the graphic and horrific scenes that confronted him on the Western Front during the First World War. I'm not certain where it came from, but I had a lightbulb moment: what if I used one of Owen's poems as a prompt for ChatGPT? ChatGPT gave me a list of Owen’s poems from which I selected, and this is what I got: Dulce et Decorum Est: Exposure: The Sentry: Strange Meeting: Owen's skill with words translated into the most vivid pictures. One art form is converted into another: It's quite remarkable. It added to my impression that AI vision has come a long way. Furthermore, it made me wonder whether ChatGPT's ability to interpret images had improved too. Taking a picture of a red, amber, and green risk report, I asked ChatGPT to explain what it was seeing. My previous blog is the result: https://www.barbicanconsulting.co.uk/post/red-amber-green-gpt-4-does-risk It is odd how the mind works; my classroom experiences of Owen's poems led to prompting a picture that then led to analysing risk through vision. I never would have thought that my English lessons all those years ago would have helped me do this. It's inconceivable. It's also an example of where mental mindsets are extremely valuable in tackling problems, and AI has been able to facilitate this. That's why there will be many discoveries that are beneficial to mankind over the coming years, arising from connections that AI can help us make, a Strange Meeting.

I've taken a screenshot of the notes from my previous blog and uploaded them to ChatGPT. Then I've asked ChatGPT to write using the following prompt: "please create a long-form article from my notes. Avoid hyperbole and the use of lots of adjectives, the tone should be factual and informative. I do want an Introduction, White collar work, Communication, all Multimodel features, implications for work as themes thank you" I was initially disappointed. I was expecting something much more in my style However, this expectation is unrealistic. When I wrote my original article, I took a lot of time and I developed my thought processes. While GPT-4 is an excellent scribe, it doesn't have the same thought patterns that I have. Furthermore, I used ChatGPT to investigate some of my chains of thought. I asked it to help improve the understandability and readability of my text; we had a long dialogue, and in my opinion, it improved the clarity. As a result of this experiment, I've learned this: ChatGPT in its GPT-4 version can certainly take a picture of your notes and turn them into an article. If someone had shown me this only 12 months ago, I would have been astounded at the speed and quality of what it could do. In the future, as we rely more and more on large language models, our critical thinking skills could become, well, less critical. When you write, part of the benefit is the process of developing your thoughts. By using ChatGPT to do the job for you, this process is short-circuited and, worse, in many cases, it may lead to lazy thinking. It's a bit like the quality of my handwriting, which has deteriorated over the years I've been using a keyboard. ChatGPT has a voice of its own. It writes well, but the way it writes is not the same as the way that you write. Some people will prefer the artificial text, and others your text. What is certain is that your way of expressing things will not be presented by ChatGPT unless you do a lot of prompting, encouraging it to adopt your tone. If you need to think things through and add your tone use ChatGPT, let it help you write and do it jointly. You can think of this just like a painter who has a particular style; their picture is unique and conveys what they see. Use Chat GPT as the brush. For generic documents and when you just need to get things done ChatGPT excels on its own. Don't get hung up on its inability to write as you would. Many people are not wordsmiths, and the fact that ChatGPT can do a very adequate job in writing means that people who have trouble expressing themselves on paper are now in a situation where they can do so easily and clearly. My conclusion is that ChatGPT or another LLMs can write just from a picture of some basic notes; something you may have jotted in a meeting. It will turn them into an effective document but if you want to add your stamp it's harder. Choosing which approach to take I leave to you. Be the Judge: The article I wrote This is what ChatGPT made off my notes: Embracing Multimodal AI in White Collar Professions: A Detailed Overview Introduction In the epoch of digital transformation, white-collar work has been consistently reshaped by technological advancements. Among the plethora of innovations, multimodal artificial intelligence (AI) systems, particularly those like ChatGPT, stand at the forefront. With capabilities that extend beyond simple task automation, these systems are becoming essential tools in the professional toolkit. This exploration delves into the integration and implications of multimodal AI in the landscape of white-collar work. The Role of ChatGPT in White Collar Work The advent of AI in white-collar professions has introduced a new dimension of efficiency and capability. ChatGPT, powered by OpenAI's GPT-4, is a significant contributor to this shift. Through its design, it learns by trial and error, enhancing its ability to assist professionals in various tasks. These tasks span the spectrum from analysis and understanding to writing and planning. The practicality of ChatGPT lies in its versatility. It serves as a round-the-clock assistant, ready to support a myriad of functions that form the crux of white-collar work. While it's not flawless, the AI's continuous learning capability ensures that it is an asset that evolves in tandem with the needs and complexities of modern professions. Communication: The AI Facilitator Communication is at the heart of white-collar jobs, whether it's through written emails, reports, or dialogues. Here, ChatGPT serves as a communication facilitator. It not only aids in crafting clear and coherent messages but also enables the absorption and relay of information in a structured manner. This is crucial when it comes to learning adaptation, where new skills are applied to familiar tasks, or explaining complex ideas to others in a comprehensible way. ChatGPT exemplifies the notion of "knowledge work," where information is processed and tasks are completed more efficiently. It also excels at breaking down complex problems into manageable segments, allowing for a clearer assessment of potential solutions and outcomes. Multimodal Capabilities: Enhancing AI's Workforce Integration ChatGPT is emblematic of the multimodal AI revolution, showcasing an array of features that mimic the multifaceted nature of human job functions. This multimodality means that it can operate across various mediums and methods: Data Analysis: ChatGPT can conduct rigorous data analysis by processing uploaded datasets, identifying patterns, and making predictions, thereby streamlining decision-making processes. Web Browsing: With browsing capabilities through Bing, it can access and synthesize information from the web, effectively functioning as a research assistant. Visual Processing: By uploading images, users can engage with the AI to understand and interact with visual data, an essential function for professions relying on graphical information. Creative Generation: The integration of Dalle into the AI suite allows ChatGPT to generate images, thereby assisting in the creative aspects of white-collar tasks, from presentations to product design. These features not only expand the utility of AI in the workplace but also herald the biggest technological change in how professional tasks are approached and executed. Implications for White-Collar Work The implementation of multimodal AI, such as ChatGPT, in white-collar professions comes with significant implications. The speed of task completion and the ability to operate in multiple ways mirror the dynamic and diverse nature of human workers. These AI systems support and sometimes surpass human capability, especially in terms of availability and speed. This AI transformation facilitates a more efficient work environment, where tasks that used to take hours can now be completed in minutes. Moreover, by taking over repetitive and time-consuming tasks, AI allows professionals to focus on more strategic and creative aspects of their work, fostering innovation and driving growth. The multimodal nature of ChatGPT also indicates a future where AI could potentially replace certain jobs. However, it is more likely that AI will augment human roles, leading to job evolution rather than elimination. The change prompted by AI in white-collar work is profound, and its full impact will unfold as the technology matures and becomes more integrated into everyday professional activities. Conclusion As we stand at the brink of an AI revolution in the workplace, it is clear that tools like ChatGPT will play an integral role in shaping the future of white-collar professions. These advanced systems promise not just efficiency but also a transformative approach to how work is conducted. As we continue to harness the power of multimodal AI, we are paving the way for a new era of professional expertise, where technology and human ingenuity coalesce to unlock unprecedented potential.

Introduction to ChatGPT If you haven't tried ChatGPT and are a bit lost about where to start, the following will help you. ChatGPT is software that allows you to use artificial intelligence; it's a type of chatbot developed by OpenAI, capable of engaging in conversational dialogue through natural language processing. It can perform a variety of tasks including answering questions, composing essays, summarising documents, developing software, and recognising images. Background and Evolution of GPT Models Let's understand, in simple terms, more about GPT-4 and other models, (my apologies to the clever people who worked on it). Machine learning (ML) has been around since the late 1940s. The recent development of being able to program Artificial Intelligence (AI) through natural language has occurred as a result of combined breakthroughs in mathematics, the internet, and computer hardware, in particular, memory and processing speed. The mathematicians and engineers designed and built clever code called a Generative Pre-Trained transformer (GPT). You can think of this as a next-word predictor, like the one you use when you text. Once the algorithm (GPT) has been built, it is trained. This requires a vast amount of data typically taken from the web and books. The GPT processes this data and works out the relationship between words so it becomes a next-word prediction model. This takes a lot of computing power to do. Once complete, the model is not perfect; it needs fine-tuning, which is done by asking it questions, seeing its response, and giving it feedback. Once the process is complete, it can be used either as a closed model, like GPT-4, where you have very little control over the way it works via its weightings, or as an open-source model like LLAMA from Meta, where you can adjust the way it works by changing the weightings. Some models are small, like Mistral’s 7 billion parameter model which can run on a laptop, whereas GPT-4's 400 billion parameter model requires Graphic Processing Units. These are very powerful arrays of chips and, at the time of writing, are expensive and highly sought after. Overview of GPT-4 and why you need it and not the free version There are several versions of ChatGPT you can use. The free version is GPT-3.5. The paid version is GPT-4. I’ll describe some of the differences: GPT-4 has a significantly enhanced capability to understand and generate more text. This is because its model has 400 billion parameters*, whereas GPT-3.5 has 175 billion parameters. GPT-4's context window is approximately 25,000 words (32,000 tokens), while GPT-3.5 has a window of 6,500 to 7,000 words (8,192 tokens). This is important; the larger context window allows for more extended interactions, enhanced linguistic finesse, programming power, image and graphics understanding, and a reduction in inappropriate or biased responses. However, while GPT-4 offers more depth and complexity in its responses, it is slower than GPT-3.5. GPT-4 has a limit of 25 to 50 messages every 3 hours; GPT-3.5 is unlimited. From experience, it is quite feasible to hit this cap, and you have to use 3.5 or have a coffee break before proceeding further with GPT-4. GPT-4 is multimodal; it can do more than generate text; it can code, read documents, examine pictures, generate pictures, write computer code, and a lot more. If you want to use the best model with all the functionality and are serious about using ChatGPT, then you need GPT-4. At $20.00 a month, you will soon find, as I have, it’s good value. I do not get paid to write this by OpenAI. Practical Applications of GPT-4 in Knowledge Work Many of us are employed in what we call knowledge work, which involves many separate tasks. You could describe these as writing, talking, analysing, explaining, presenting, and persuading. Information comes to us, we process it and in turn, add value to it before passing it on to others via conversation, emails, texts, spreadsheets, graphs, and presentations. All of these tasks have one thing in common; they come under the heading of communication. To improve our communication, we are learning new skills and adapting old ones to make sense of the world and convey messages to others. When you use GPT-4, you realise that, in its multimodal form, it can integrate into what you do. It has tools and features which can be used in various situations to help you. GPT-4 is the Swiss army knife of knowledge work. The tools that it has when used properly, can not only speed up the way you do things but can also improve what you do. And if you are doing things faster and better, this is a “good” thing. In a market-based economy, it makes you more efficient and therefore less prone to obsolescence. How GPT-4 helps When you consider all the skills we use at work, you realise the development of GPT-4 has not been a haphazard process. Its multimodal functionality helps us undertake what we do. Let's have a look at some of these features: Writing: Through the chat or prompt interface, you can communicate with ChatGPT and ask it to do anything related to text. It will write an email, a thank you note, a memorandum, a letter; this list is endless. It will write in a particular tone or style or enter a conversation with you. Information retrieval: GPT-4 has a training database drawn from a mass of information from the web. You can ask it for facts, information, and opinions on just about any topic. The model that it currently works on has a cut-off point of April 2023. If you want reliable information after this date, you must ensure that GPT-4 is enabled to browse the web. Browsing: GPT-4 does this through Bing. When you converse with GPT-4, it can search the web for more up-to-date information and use that in its response. With the browse function enabled, GPT-4 will decide when to browse. I’ve found it helpful to ask it to browse, particularly when you want events post-April 2023 to be included in your conversation. Data analysis: GPT-4 has a powerful data analysis capability, which works with both numeric and text data. For example, you can upload a spreadsheet and ask it to analyse trends or upload text-based customer feedback and ask it to comment on sentiment. With trial and error, GPT-4 will help you gain insights quickly, easily, and cheaply. Vision: You can upload pictures to GPT-4 and work with these. For example, you can upload a handwritten note and get it transcribed, and then rewrite the contents using GPT-4. You could upload a picture of a report, for example, a technical medical paper or diagnosis, and ask GPT-4 to explain the contents to a layman, and it will do it. DALL·E: This is Chat GPT's picture-generating model. You can ask GPT-4 to generate a picture by giving it instructions in the prompt window; it will then generate a picture that mirrors your request. You influence the output by altering your prompt. Writing computer code: GPT-4 enables non-programmers to write executable code, it not only generates code snippets that run directly but also offers features like code cleanup, code changes, and insights on coding practices. It can write in languages such as Python and HTML, making coding accessible to individuals with varying levels of technical expertise. Furthermore, GPT-4's autonomous code-writing capabilities are closely linked to the data analysis function in that it writes code to undertake the data analysis tasks; prompting the model leads to code writing and execution and as a non-programmer, this is how I use it. Custom GPTs: These represent an advanced yet user-friendly feature that allows you to create tailored AI models based on specific prompt configurations. Imagine the process of interacting with GPT-4 by crafting detailed prompts to guide its responses or actions. These carefully constructed prompts can be transformed into custom GPTs, enabling you to automate repetitive tasks or specialised queries without the need for continuous prompting or retraining of the base GPT-4 model. These custom models can be saved for personal use, shared with select individuals via a direct link, or made publicly available through the GPT Store. The GPT Store functions similarly to an App Store, offering a platform where anyone can discover and utilise custom GPTs created by others. I find the custom GPTs very helpful as they let me use GPT-4 in a way that would normally require developer skills. Accessing GPT-4 There are several ways to access GPT-4; the best way depends on what you need it for: For individuals seeking a straightforward and user-friendly way to interact with GPT-4, a personal subscription to ChatGPT is the best option. It provides easy access to the AI's conversational capabilities without the need for technical knowledge or development skills. This subscription is well-suited for a wide range of personal use cases, from learning and entertainment to productivity and creative assistance. For business users, there is either API access, which is direct access to GPT-4 through OpenAI's API, (encrypted key), this is suitable for developers and businesses looking to integrate AI capabilities into their applications or services. Or Team GPT which provides collaborative features and more confidentiality (you can stop OpenAI training on your data). It's designed for enterprise use, requiring a more expensive business subscription due to a minimum of two seats. You can also use OpenAI’s sandbox, which is another environment primarily for development where you can adjust the parameters and see how various prompts behave. Finally, you may be using GPT-4 without knowing it through third-party websites or applications integrating the API into their offerings. This method can vary in user experience and functionality, depending on how the third party has utilised the GPT-4 capabilities. The World of Work and GPT-4’s Role In the knowledge economy, our work involves reading, writing, and arithmetic (if I can use that as a description of coding, analysis, and strategy), and the faster that we can obtain information and convert it into something valuable and then communicate with others, the more efficient we become and the more valuable we are. When you consider GPT-4, it has all the things that we do built into it. It can write, browse, analyse, study pictures, make pictures, write code, and enter dialogue, it can save your prompts and store them for you; it is multifunctional. When you see this, you realise GPT-4 and similar general models are massive steps forward in our ability to deal with data and convert this quickly and intelligently into something usable and helpful. Because these models are the worst models that we're going to use and they are getting better and better very quickly, they will soon be on par with, or better than, human experts in a given field. Where your work is not at an expert level, then it is very likely that these models will replace what you do because they are cheap. This means economic disruption for many. The only way to take advantage of these changes is to embrace the use of these models in the work you do. The best way to do this is to see them as helpful graduate trainees who have a very good basic understanding of how the world works, but they do not currently have in-depth expertise that has been gained through experience. By working with them, much of the day-to-day work can be delegated out, and you can concentrate on using your experience to take your work to a higher level. Summary Models like GPT-4 have changed things. In white-collar industries, we are paid for our knowledge and ability to use information, process it, and complete all sorts of tasks. I’ve called this communication. At first sight, GPT-4 is a chatbot, but once you explore all the things it can do, there is a whole suite of tools like the proverbial Swiss army knife. There is no training manual because how you use it depends on what you want it to do, your curiosity, and your willingness to accept change. Because this computing power is cheap and effective, the workplace will change; if your job can be done by GPT-4 or its offspring, your bargaining power is likely to decline. The rate of change will be much faster than, say, web usage as the infrastructure and incentives are in place, and the advancements are speeding up, not slowing down. You should consider bringing these models into your work to future-proof what you do. Use them as helpful assistants, treat them in the same way you would a co-worker who you are instructing, and you will be amazed at what they can do for you. This is truly where you can learn. Using GPT-4 is a very practical skill that is largely gained from practice. You can get the basics in a few hours but “intuition” about its capability needs a lot more time and experimentation. By working together, you can see what you can do that it can’t, this is where you are adding value and it’s where the future is. Critical thinking, emotional intelligence, and complex problem-solving are increasingly valuable in a GPT-4 augmented workplace. This is both a great change but also a massive opportunity to do new things. *A note on parameters: Think of the parameters in a GPT model as the internal parts of an engine that you can adjust to optimise performance. These parameters are akin to the nuts, bolts, timing belts, spark plugs, and fuel injectors in a car's engine. Parameters as Engine Components: Adjustable Settings: Just like you can tweak the timing of the spark plug or the air-to-fuel ratio to get the best performance out of an engine, in a GPT model, the parameters (which are essentially weights in the neural network) are adjusted during the training process to make the model more accurate. These adjustments are based on the data the model is trained on, similar to how you might tune an engine based on the type of fuel it will use or the conditions it will operate in. Complexity and Performance: The number of parameters in a GPT model is similar to the complexity of an engine. An engine with more components (like a V8 versus a V4) can be more powerful if tuned correctly, but it's also more complex to manage. Similarly, a GPT model with more parameters can process language more effectively, capturing nuance and complexity in text, but requires more data and computational power to train and tune. Optimisation for Efficiency: In the same way that you might use a turbocharger to increase the efficiency and output of an engine without making it bigger, optimisation techniques in machine learning adjust these parameters to improve the model's performance without necessarily increasing its size. This involves fine-tuning the model's parameters to ensure that it can understand and generate language as accurately and coherently as possible. Fine-Tuning for Specific Tasks: Imagine you have a basic engine setup that you can then fine-tune for different purposes, like racing, towing, or fuel efficiency. Similarly, once a GPT model has been pre-trained on a large dataset to understand language generally, it can be fine-tuned by making smaller adjustments to its parameters for specific tasks, such as writing in a particular style, translating languages, or answering questions. This fine-tuning process adjusts the model's internal settings to specialise in the task at hand, much like adjusting an engine to optimise it for a specific type of driving. In the workshop of machine learning, parameters are the internal components of the GPT "engine" that determine how it processes and generates language. By adjusting these parameters through training and fine-tuning, similar to how a mechanic would tune an engine, we can optimise the model's performance for specific tasks or improve its overall efficiency and accuracy. Just like with engines, the skill lies in knowing how to adjust these parts correctly, balancing power, efficiency, and usability to meet the desired outcome.

Every time you use ChatGPT, you enter a prompt, which is a request for what you would like it to do. At a basic level, prompting is straightforward, provided you have the mindset to think about the task at hand. However, this means that each time you need to create a new prompt to get ChatGPT to work for you. There is an easier solution, and that is to make custom GPTs. These remember prompts and related actions, allowing you to complete tasks with ChatGPT without having to remember the details yourself. Creating a custom GPT can be done by anybody, and no special coding skills are required. Here's an example: Let's suppose I want a selection of thumbnail pictures that I can use for a blog post or article. I would like these to be related to the article and draw the reader to the content. If I was only doing this once, then I would just use the normal prompt box in ChatGPT. But if it is a task that I am doing on a regular basis, then it would be helpful to have a custom GPT. The first thing you do is click on "Explore GPTs" on the left-hand side: Then in the top right , you will see a button labeled "Create.": Click on this, and it opens a window. Complete this with the task you would like the GPT to help you with. Here, you can see my instructions for the thumbnails that I wish to create. On the right-hand side, you can test the action of the GPT and make any necessary changes: Once you are happy, you can save the GPT either for your own use or for those who have the link. Alternatively, you can open it entirely and put it in the store. And that's about it. This is the first step to creating GPTs, and the more you try them, the more creative you become. Good luck! P.S. I posted this blog into my custom GPT and this is what it came up with: ------------------------------ Thank you for sharing the details of your blog! Based on the summary, it's clear the blog discusses the concept of using ChatGPT for various tasks, focusing on the ease of creating custom GPTs to streamline and personalize the interaction process. The key themes include the simplicity of prompt creation, the advantage of custom GPTs for recurring tasks, and the accessibility of GPT technology to users without coding skills. Let's create four distinct thumbnails to encapsulate these themes: Painting Thumbnail: A vibrant, impressionistic painting of a person sitting at a desk, surrounded by floating, colorful thought bubbles. Each bubble contains symbols or icons representing different tasks (like a camera for photography, a pen for writing, etc.), with a larger bubble showing a simplified, friendly GPT interface. This image will symbolize the creativity and variety of tasks that custom GPTs can assist with. Photo Thumbnail: A high-quality, realistic photo of a diverse group of people gathered around a computer screen, visibly engaged and excited. The screen shows a simplified, user-friendly GPT interface, hinting at the collaborative and accessible nature of custom GPTs. This photo will convey the idea of inclusivity and the broad appeal of GPT technology. Sketch Thumbnail: A detailed, black and white sketch showing a close-up of hands typing on a laptop, with abstract lines and shapes emanating from the screen to represent the digital workflow and the seamless interaction with GPT. This sketch will focus on the personal interaction between the user and the technology, highlighting the ease of creating custom prompts. Abstract Thumbnail: An abstract design featuring geometric shapes and lines converging towards a central point that represents a digital brain or AI concept. The use of bright, contrasting colors against a dark background will create a sense of innovation and the futuristic potential of custom GPTs. This thumbnail will emphasize the advanced technology and creative potential behind GPTs. I'll begin with the painting thumbnail. Please hold on while I create the image. Here's the impressionistic painting thumbnail that captures the creative and versatile use of custom GPTs. It visually represents a person engaging with a variety of tasks through thought bubbles, with a focus on the GPT interface. Next, I'll create the photo thumbnail. This will be a realistic depiction of people around a computer, highlighting the collaborative and accessible nature of GPT technology. Please hold on The photo thumbnail is ready, showcasing a group of people engaging with GPT technology, reflecting the theme of collaboration and accessibility. Moving on, I'll work on the sketch thumbnail that will focus on the personal interaction between the user and the technology, emphasizing the ease of creating custom prompts. Please hold on. The sketch thumbnail is complete, highlighting the intimate interaction between the user and GPT technology through the act of typing and the flow of ideas. Finally, I'll create the abstract thumbnail focusing on the futuristic potential of custom GPTs with an emphasis on innovation and creativity. Please hold on. The abstract thumbnail is ready, capturing the essence of innovation and the futuristic potential of custom GPTs through dynamic shapes and vibrant colors. You now have four distinct thumbnails to choose from for your blog: The impressionistic painting thumbnail, emphasizing creativity and task diversity. The photo thumbnail, showcasing collaboration and accessibility. The sketch thumbnail, highlighting personal interaction and the ease of use. The abstract thumbnail, representing the innovative and futuristic potential of GPTs. Each thumbnail has been designed to draw the reader in and accurately reflect the content and themes of your blog. Please review them and select the one that best suits your needs.

Here's an experiment with ChatGPT. I've asked it to re-write my blog in the vernacular. Is it easier to understand like this? The clarity it brought to the message was surprising, leading me to wonder: has our shift to digital communication altered our language standards? If it has then using ChatGPT to write like this may be more effective than expected. Here's the blog: You might think you gotta be some kind of genius to get the hang of this AI stuff, but that's way off base. You don't need to be some kind of "Prompt Guru" or anything, just gotta be curious about what it can do for you. I wrote this especially for folks who haven't had the chance to mess around with ChatGPT or anything similar. Sure, the tech behind it is pretty complex – I'll be the first to admit I don't get all the ins and outs – but that hasn't stopped me or anyone else from using it. If you're good at firing off emails or texting, you're already set to get good at using AI in no time. Here, I'll walk you through some simple tricks to find what works best for you. You can ask ChatGPT to whip up something for you. Just type in what you're thinking and let it do its thing. Usually, you'll get back something pretty decent, even if it's not exactly in your style. Want something tweaked a certain way? Ask ChatGPT to redo it – maybe in a lawyerly way, or like a journalist, or even in slang. Check out what it gives you and see if it hits the mark better than your first go. What you've got might already be spot on, but maybe you wanna make it more your own. Just copy what ChatGPT gives you into a Word doc and tweak it to your heart's content. Here's a trick I find super useful: say what you wanna say into a Word doc, then copy that over to ChatGPT. You can ask it to just clean up the spelling, grammar, and syntax without changing anything else. Or, you could ask it to jazz it up – make it simpler, more advanced, or just better in general. As you can see, it's not rocket science. Just chat with ChatGPT like you would with a coworker. You can't mess it up. The more you use it, the easier it gets. It's a real time-saver, and I reckon you'll be pretty chuffed with what it can do if you give it a whirl.

One of the misconceptions about generative AI is that you need to be a rocket scientist to use it, but nothing could be further from the truth. You do not need to be a "Prompt Engineer" but just curious about what it can do for you and how it can help you. I've written this primarily for those who have not had the chance to use ChatGPT or similar tools. While the software itself involves a huge amount of complexity that I confess I don't fully understand, it doesn't stop me or anyone else from using it. If you can write emails or send texts, then there is no reason why you can't become proficient at using AI in a short space of time. Here, I will explain a few of the simple techniques that you can use to explore what works for you. You can ask ChatGPT to write something on your behalf. Just type into the prompt what you want and let it go ahead. What comes back is normally satisfactory, although it may not be exactly to your taste or style. https://chat.openai.com/share/1861bc73-ad0f-4e18-a1d6-edc873474bff Ask ChatGPT to rewrite the piece in a particular manner, for example, as a lawyer or as a journalist. Alternatively, ask it to write in a particular style, like the vernacular. You can then consider the results and see if they are an improvement on what you originally had. https://chat.openai.com/share/9fdeb055-50f2-4b98-94f2-dc7a6793c987 What you have done so far may already be enough, but perhaps you want to personalize it a bit more. You can cut and paste the output from ChatGPT into a Word document and make the alterations as you wish. Something I find particularly helpful is to dictate what I wish to say into a Word document and then copy it straight into ChatGPT. You can then ask it to correct the spelling, grammar, and syntax, leaving the rest unchanged, and it will give you the new version as requested. Alternatively, you can ask it to improve and make changes that make the document easier to understand, at the level of a 10-year-old, at the level of a graduate, or just ask it to make improvements as it sees fit.https://chat.openai.com/share/bd218094-9d15-415e-9c76-5bc3712ecf5a As you can see, there is very little to it. Just converse with ChatGPT as you would a work colleague, and you can't really go wrong. The more you use it, the easier it becomes. It will save you a lot of time, and I think you will be impressed with the results if you've not tried it.

At 97, my mother faces the challenges of old age, including complete hearing loss, but she still enjoys reading. I've observed that elderly individuals often take pleasure in reminiscing about the past. This led me to experiment: could I develop a GPT model that would bring joy to my mother? The result was "Mum's Diary," a customized GPT designed to generate a hypothetical diary page from exactly 50 years ago, incorporating people and events from that time that she knew. It creates a narrative of what she might have done, including a small picture related to the diary content. How It Works With this custom GPT, you simply enter the prompt "let's go," along with a picture, and a diary entry that my mother could have written 50 years to the day materializes. It's remarkably straightforward. Building It Is Simple To create your custom GPT, you need access to GPT-4 (the paid version) and to follow the provided instructions: No programming skills are necessary. I specified that I wanted a diary entry reflecting my mother's activities and thoughts from 50 years ago, including interactions with people and typical daily activities. I requested the inclusion of relevant local, national, or global events and their impact on her, alongside background information about her job, family, neighbours, shopping habits, holidays, car, weekend activities, pets, etc: This is what it looks like (and every time you prompt you get a new page)*: My Mother's Reaction It's too early for a definitive opinion, but she is intrigued by the generated content and enjoys reading it. This project has painted a vivid picture of her past activities, proving to be a delightful experience. She has commented on various historical accuracies and discrepancies, which can be easily adjusted by modifying the custom gpt. Further inquiries into improving the custom GPT (by asking GPT-4 itself) suggested that adding more detailed information relevant to the specific period would enhance the accuracy and reliability of the output. Although some generated storylines may seem repetitive, this doesn't concern me much at the moment. I believe this can be addressed by adjusting the instructions to vary the content generation, which I am currently exploring. Potential Applications While this custom GPT is tailored specifically to my mother's life, making it unsuitable for generic use in the GPT store, its concept could benefit others. By inputting personalised details, it could serve as a unique diary for another individual, potentially aiding memory recall and mental stimulation for elderly people either at home or in care facilities. This application of generative AI could add a little something to improve the quality of life for many. Conclusion In conclusion, this project exemplifies a simple yet powerful way to utilise generative AI to enhance people's lives. If you found this interesting, please share it with others who might benefit. Should you require assistance in developing a similar GPT, feel free to contact me. ---------------------------------------------------- * The complete diary entry for the day: Diary Entry: February 16th, 1974 Today was a particularly reflective day for me, full of small but significant moments that reminded me of the delicate balance of our daily lives here in Yattendon. The morning air was crisp as I made my familiar walk down to the school, the quiet of the village broken only by the distant sound of Flynn's tractor in the fields and the cheerful greetings from Mr. Peaty as he made his rounds. In school, the children seemed more spirited than usual, perhaps the result of the recent news that had the whole village talking: the miners' strike and the resulting three-day work week had finally come to an end. Mrs. Gaskell and I had a lengthy discussion about it during our morning break, pondering over the impact it had on the local community and the country at large. It was a relief to know that, for now, the crisis was over, and there was a collective sense of hopeful anticipation for what lay ahead. After school, I stopped by Mrs. Hogg's post office to pick up a few essentials. The lively barks of her two pugs greeted me before I even stepped inside, bringing a smile to my face. Mrs. Hogg shared the latest gossip as she always does, mentioning that Mr. Behr had decided to host a small celebration at the Royal Oak to mark the end of the power cuts. The community needed something to cheer about, and this seemed like the perfect occasion. On my walk back home, I couldn't help but marvel at the beauty of our village. The sight of our blue Land Rover parked outside Pound Cottage, with Chilly lounging lazily in the sun nearby, filled me with a deep sense of contentment. Ivan was in the garden, tending to his flowers, a pastime that had become more than just a hobby for him. We exchanged stories of our day over a cup of tea, reflecting on the importance of community and resilience in times of uncertainty. This evening, as I sit down to write these words, I am reminded of the strength and perseverance that define us, not just as individuals but as a community. The challenges we face, whether they're nationwide crises or personal struggles, have a way of bringing us closer together, strengthening the bonds that tie us to this place we call home. As I close my diary for the night, I am filled with gratitude for the life we have built here, for the friends and neighbours who enrich our lives in countless ways, and for the simple, enduring pleasures of village life. Note: Chilly was a pet cat. Just typing in a different date into the prompt generated an entry from a different period.

Forward looking projections ask us to look into the future. Their uses are varied and range from planning and risk to investment and speculation. We try to eliminate guesswork by using a model. It contains the inputs and algorithms to provide answers. Subjectivity often abounds. The inputs are frequently estimates they include correlations and behaviour. However adding more and more moving parts is fashionable. It creates spurious accuracy and makes challenge harder. Uncomfortable truths are hidden. They include: KISS. It’s better to be approximately right than precisely wrong; Unknown unknowns. Beyond the very short term things go wrong; Back solving. The answers you want build the model; Unobservables. Fudge the result. Let’s be clear projections give us some feel of the issues involved. But they are not facts. They help (when taken with a pinch of salt) in assessing value, risk and future contingencies. They provide some of the analysis as to whether you are on the right track. No more than this. Next time you are told about the future as if it’s a fact you are being misled.

People, systems and external events all generate operational risk. In the words of one Rogue trader “It’s sure to go up, trust me.” Whilst we’ve been focused on solvency and liquidity this just quietly ticks away – a sort of Cinderella of risk that from time-to-time bites. It makes you ask the question “just why didn’t we see this coming?” One of the difficulties is measurement. Subjectivity in terms of ranking and severity of loss inevitably apply. However a little more could be done to at least reduce the chance of mishap. One of the friction points is where systems and data meet people. This falls into three distinct headings: Collection of information (integrity); Working with information (processing; The end result (output). Nothing new here…..garbage in, garbage out but can anything improve? I think so. Whilst firms rely on experienced management a lot of day-to-day work is undertaken by less skilled operatives their objective being task completion. This needs to change and the simplest way to do this is to “sense check” things. This does of course rely on a basic understanding of what you are doing and its importance to the business. Failure by banks to provide employees with this basic skill is asking for trouble.

Paul Lewis in an article in the FT on 18/6/16 (Cash v stocks: the winner may surprise you) has uncovered something that people need to be aware of. The fund management industry attracts savings in the form of ISAs and pensions on the basis that it can outperform. But does it offer the saver any level of comfort that this will indeed be the case? In his article, Mr. Lewis compares cash savings returns with those from shares. His tentative conclusion is that for periods of up to 20 years cash beats shares and if you do buy shares low-cost index trackers are best. This appears to be contrary to what the public get told by financial advisors. Why? Because the comparisons made that support the case to buy shares do not compare like with like. Lewis goes on to explain that cash can be deposited at the best buy rates year in year out. These rates are significantly higher than T-bill rates. But it is T-bill rates that are often used to compare cash and share returns. Furthermore, funds incur fees, commissions and dealing costs. But these costs are frequently ignored in calculating share based returns. In other words comparisons made between cash deposits and fund based investments are skewed to make funds look significantly more attractive than they really are. For many sitting down once a year and reinvesting in a best buy fixed rate account will provide the best and safest return.