Generative AI - The impact on the future of Solution Architecture and the Cloud industry.

“I skate to where the puck is going to be, not where it has been”

Wayne gretzky

Introduction

The one thing that I love to do is to take time to think about the impact of disruptive technologies on the world, should they reach their full potential. I have spent the last year rambling on how I believe social commerce will disrupt our buying habits and bring our social and purchasing worlds into one. Just as the rise of the cloud upended traditional IT infrastructure, I believe that the rapid evolution of Generative AI(GAI) has the potential to upend how we deploy infrastructure to the cloud, and ultimately, how Solution Architects, like myself, do our job.

GAI is not an incremental advance; rather, I believe that it promises to be a tectonic realignment in how we architect, procure and deploy our technology infrastructures. At the heart of my theory lies the proposition that GAI, unhindered by the inconsistency of human emotion and biases, will deploy raw, data-driven decision-making that will fundamentally change how software and the underlying infrastructure is architected. Void of emotion, GAI will have the ability to create technology architectures with precision and efficiency previously unseen. However, with every seismic shift, the aftershocks are equally impactful - they promise to dramatically reshape not only our methodological approach to technology selection and deployment but also demand a re-evaluation of the very role of Solution Architects (SA) within this rapidly transforming landscape

The question on my mind that was the driver behind the motivation to create this essay is what happens when I can simply describe the problem I am facing, and an AI model understands enough about my business and the technology landscape to develop an end-to-end architecture, and seamlessly deploy it to the cloud. How will this affect the vendors I use when AI, with no emotional attachments, can choose the most cost-effective, easiest to operate, and most performant solution tailored to exactly what I want to achieve? Then optimise or replace it should a better solution arise. While this seems like something out of a Star Trek and far from the GAI of today, if Sam Altman, CEO of Open AI, truly believes AI can bring about an end to human society, you would have to think that creating and deploying AI architectures will well be within its capabilities, a lot sooner I would imagine also! 

This shift, while promising unprecedented efficiency and optimisation, will reverberate throughout the industry, leading to an upheaval in traditional roles and design patterns. Customer-facing roles will need to evolve. Notably, SA’s, of which I am one, will find ourselves at an interesting crossroads. With the strategic deployment of GAI, our role will transform, becoming less focused on technology selection and more attuned to unlocking business value. We will need to cultivate a deeper understanding of the industries we serve, with a huge shift to helping customers see around corners, unlocking new business revenue, and utilising AI to deploy solutions to test our hypothesis rapidly. If the GAI industry succeeds as forecasted, my view is that I should be able to explain the problem I am trying to solve, and it should choose the appropriate stack, technology vendors, and products to best meet my needs. Then deploy it. If it turns out to be the wrong solution, it can fix that, also.

There are a lot of problems that need to be solved before this end-to-end seamless integration with AI becomes a reality, the GAI of today has little more than the capabilities of a junior engineer and I spend more time fixing its messes than I do designing a solution but as I embark on this exploration of future possibilities, I will delve into how we make technology choices today, how this significant transition to GAI may play out, and what it means for those of us who architect our digital futures as well as those vendors who build products for it. I will finish with a look at what the future of the cloud industry may look like as we transition into a future of Infrastructure as Prompts rather than code.

II. The Emotional Side of Technology Decision-Making

Anyone who knows me knows that I am a huge proponent of emotional intelligence and the impact our current emotional state has on everything we do. Decisions, at their core, are psychological endeavours, and technology selection is not exempt from this rule. The influence of emotions and biases can often be traced to the decisions we make, consciously or unconsciously. As humans, we cannot effectively make decisions without emotion; we use ‘somatic markers’ in our bodies to help us filter data sets of information. Understanding this emotional side of decision-making is vital to appreciating the transformative impact generative AI could have on technology and architecture selection.

Let's start by unpacking the psychology behind decision-making. According to Prospect Theory, developed by psychologists Daniel Kahneman and Amos Tversky, people often make decisions based on the potential value of losses and gains rather than the outcome. What's interesting is that losses are felt more intensely than gains - a principle known as loss aversion. How does this translate into technology selection? Well, for one, the fear of making a 'wrong' choice and the potential loss associated with it could drive organisations towards sticking with familiar technologies or vendors - the proverbial 'better the devil you know' approach. It can also lead to ‘the paradox of choice,’ a term developed by Psychologist Barry Schwartz, where when consumers are faced with multiple options, they often freeze as they begin to consider hypothetical trade-offs, increasing the time taken to deliver a new solution to the market.

Another emotional influence is loyalty, particularly towards vendors. Like in any relationship, time and experiences create a bond between organisations and their technology providers. This bond can sometimes sway objective decision-making, pushing businesses to choose a vendor's solution even when there might be better fits available in the market. How many enterprises choose a technology because they are locked into contracts or because the pain of working with procurement for a new vendor outweighs the benefits of the new product?

Lastly, the Fear of Missing Out, commonly known as FOMO, is another potent factor. As technological innovation accelerates, there's always something new on the horizon. This can create psychological pressure to adopt the latest technology, often without proper consideration or analysis. I have certainly been guilty of this, specifically when I started at AWS, and I had access to all of the new technologies in an instant. The rise of Kubernetes in the cloud is a case in point. Despite its complexity and steep learning curve, many organisations jumped on the Kubernetes bandwagon early, driven more by FOMO than a clear need for its capabilities. Does this sound familiar in the current AI madness we are all in right now?

The influences of loss aversion, loyalty, and FOMO on technology and architecture decisions are not merely theoretical. Numerous studies, including 'Conceptualising improvisation in information systems security' by Njenga & Brown (2012), highlight how emotional influences can lead to suboptimal technology choices.

The Drawbacks of Emotion-Driven Technology Decisions

The emotional influences that drive technology decisions aren't benign factors. They have real-world consequences for operational efficiency, innovation, and the bottom line.

Consider vendor lock-in as an example. Loyalty to a vendor, driven by familiarity and loss aversion, can sometimes lead to myopic technology choices. The comfort of staying with a known vendor can overlook the potential benefits other technologies could bring. However, this attachment can become a double-edged sword, with the organisation becoming so entwined with a particular vendor's ecosystem that migration becomes a daunting, if not impossible, task. This lock-in stifles innovation, hampers flexibility, and can potentially result in significant inefficiencies. 

Moreover, FOMO-driven decisions, which push organisations to chase after the 'next big thing' without proper evaluation, often lead to suboptimal technology choices. I remember back in my networking days, working with customers to replace old Catalyst 6500 switches with the new Nexus devices from Cisco, even though the old machine hadn’t dropped a packet in about ten years. Technology choices are driven more by vendor-led market hype than a clear need for its capabilities resulting in many businesses grappling with complexity without reaping proportionate benefits. This has been the case in vendor-driven architecture patterns, driven in some cases by the motivation to increase the amount of said vendor products purchased in a solution.

These choices aren't merely operational issues; they translate into tangible costs. As per a Gartner report, globally, organisations waste $1.8 trillion annually due to poor technology choices and inadequate resource management. This is not just a massive financial drain but also an enormous opportunity cost, as these resources could be better invested in driving strategic initiatives and innovation.

These potential pitfalls underline the necessity for objective, data-driven decision-making in technology and architecture selection. Emotion-driven decisions, while inherently human, often lead to subpar outcomes. As we delve into the potential of GAI, it becomes clear that it can help us navigate these biases and bring objectivity to our decisions, maximising the chances of successful digital transformation and opening up a whole new level of rapid experimentation. It enhances the true potential of the elastic, software-driven, cloud industry.

Generative AI: An Emotionless Decision-Maker

When we imagine the dawning era of GAI, we often daydream about notions of streamlined efficiency, improved productivity, and unprecedented potential for innovation. Yet, one aspect that is less remarked upon but arguably as transformative is its capability to make choices free of emotional bias and politics.

Decisions about vendor selection or technology stack choices are rarely as binary as one might think; indeed, as we spoke about above, they can be influenced by emotional factors such as FOMO (Fear of Missing Out), political or contractual persuasions, or the psychological concept of loss aversion. 

Enter GAI. With its capability to process enormous amounts of data and, more importantly, its inherent lack of emotion, GAI is primed to revolutionise the technology decision-making process. By analysing data based on specified parameters and objectives, it can provide choices devoid of the emotional baggage that often taints human decisions. If I skate to where GAI is today, I wouldn't trust it to deploy a single EC2 instance in my account. However, if I take Wayne Gretzky's advice and skate to where it is going, at the current rate of improvement I should be able to trust its decisions more than I trust my own in 2-3 years. This is because the introduction of an emotionless 'decision-maker' into the process could drastically reduce the likelihood of suboptimal choices driven by my own cognitive bias.

GAI, therefore, isn't just a technological game-changer; it's poised to transform the very fabric of decision-making within technology architecture. By uncoupling choices from emotional biases, we can open up a future where decisions are more rational, more aligned with business needs, and ultimately more beneficial in driving value and innovation. It can open up a way of looking at solving business problems that has never existed for us. Just look at when Google’s AI, AlphaGo, beat World Champion Go player Lee Sedol. It played a game that most human Go players would have considered bizarre, yet the AI looked at the challenge in a way we never have done before and created a whole new best practice.

The Future Impact of Generative AI on the cloud industry

Building upon the realisation of GAI as an emotionless decision-maker, we now turn to the transformative potential of the technology vendor market. By making data-driven, vendor-neutral decisions, GAI might not only change how we select technologies but could upend the dynamics of competition and innovation within the industry itself.

In the realm of technology choices, the longstanding open-source versus proprietary software debate, cloud vendor selection, or the feasibility of adopting a multi-cloud strategy are aspects that will inevitably be impacted. These decisions, traditionally influenced by a mix of rational considerations and emotional biases, will be handed over to an analytical entity capable of crunching vast amounts of data to deliver the most suitable choice. Where emotional biases may have once favoured familiar platforms or the latest market trends, GAI would prioritise choosing a solution purely based on feature compatibility, operational costs, performance efficiency, and specific business requirements.

For instance, let's consider the selection of cloud vendors. The conundrum of choosing a single provider versus diversifying across multiple platforms often swings between the need for convenience and integration versus the demands for resilience and flexibility. However, under the analytical prowess of GAI, these deliberations would transform into an evaluation of quantifiable parameters: pricing models, performance metrics, compatibility issues, and business-specific needs.

The implications of such a shift for competition and innovation in the cloud market could be profound. Vendor selection, once perhaps influenced by marketing prowess or market dominance, would be subject to the cold, hard facts of data analysis. It levels the playing field, giving smaller, innovative niche players like Nvidia a fair shot, thereby fostering competition and encouraging innovation, which is a great thing for the long term prospects of our industry and our customers as a whole.

As we gaze into the future of cloud architecture, one thing is clear: the influence of GAI on the cloud industry will be wide-reaching and transformative. It will not just be a tool, but a game-changer, leading to a cloud market driven by data, not emotion, and characterised by heightened competition, innovation, and customer-centricity.

How the SA role may evolve

As GAI reshapes the cloud market and alters the decision-making landscape, some people, including SA’s question the role of the SA in this new era. Rather than seeing this as a threat, I believe it presents a golden opportunity for SA’s to reinvent their roles and to add even more value in the decision-making process.

Historically, the journey of an SA has taken us through a variety of roles and responsibilities. We often start our career in the trenches, hands-on, building and implementing solutions. As we gain experience and ascend the ranks, our perspective broadens from the granular to the more strategic. Moving from implementation to design, ultimately placing us at the intersection between business and technology.

However, as GAI increasingly takes on the task of selecting and implementing technology solutions, the role of the SA is primed for a significant evolution. Assisted by GAI, SA’s can transcend their traditional role of mediating between business and technology, instead, transitioning into becoming industry experts, utilising our human creativity and our research expertise to help customers see where the puck is going.  This is where the true potential of the SA-AI partnership lies.

While GAI may be able to turn business problems into technology solutions, the customers need experts who can help them see where the puck is going from an industry and technology perspective, so they can identify these problems or business revenue opportunities. Our role will be to help businesses foresee potential opportunities, identify how emerging technologies can be used to create new revenue streams, and then harness the power of GAI to design and deploy these solutions in the cloud, rapidly.

In essence, GAI offers SA’s the opportunity to ascend to the role of strategic advisors, guiding organisations through the complex landscape of technology innovation. It's a shift from being a tech expert to an industry expert, from being reactive to proactive, from providing solutions to foreseeing opportunities. In this world, Architects are not threatened by AI but are empowered by it, ready to unlock the next frontier of digital transformation.

Conclusion

In my research for this paper, it became more apparent to me that the rise of GAI signifies not merely an incremental advance, but a paradigm shift in how we comprehend and employ cloud computing and solution architecture. This transformative technology, in its emotionless, data-driven decision-making approach, holds the potential to eradicate human bias and achieve unprecedented precision and efficiency in creating and deploying technology infrastructures.

I tried to begin unravelling the pervasive, yet often overlooked, role of emotional influences in technology decisions. From loss aversion to vendor loyalty, and the infamous FOMO, I learned more about how these biases often skew our decision-making, leading to suboptimal choices and missed opportunities. GAI, in some future state (the precise date I have no idea of), presents the capacity to bypass these human tendencies, bringing objectivity and informed decision-making into the picture.

The conversation focused on how GAI can help us overcome the deep-seated, yet often unnoticed, impact of emotional and political influences in technology decisions . I for one have been guilty in the past of choosing specific technologies or vendors based on my comfort level or affinities. However, as I learned, vendor lock-in and the chase for the latest technology, sometimes fueled by our emotional inclinations, often result in tangible costs to customers and a stifling of innovation. By integrating GAI into the decision-making process, I can open up a new level of rapid experimentation, leveraging the truly elastic nature of the cloud, and maximising the probability of successful digital transformations reducing the expense associated with failure.

The potential impact of GAI extends beyond individual organisations to influence the larger cloud market and the dynamics of competition and innovation within the industry. GAI's data-driven, vendor-neutral approach can lead to more balanced competition. Potentially levelling the playing field for niche players who specialise in specific technology especially if these providers can integrate with the major cloud providers who offer the other services needed by enterprises.

Finally, I explored the implications of the GAI revolution on the role of SA’s. Far from rendering the SA role obsolete, the advent of GAI paves the way for a meaningful evolution of the role. SA’s can transition from tech generalists to strategic industry analysts and technology experts, leveraging their knowledge in conjunction with GAI to foresee opportunities and create new revenue streams. Thus, GAI empowers architects to ascend from solution providers to strategic advisors, guiding organisations through the labyrinth of technology innovation.

As we stand on the precipice of this revolution, the potential of GAI in cloud computing and solution architecture is immense. However, the realisation of this future demands us, particularly those of us in roles such as SA’s, to embrace this change. It's crucial that we evolve in tandem with these technological advancements, leveraging GAI as a tool to enhance our roles and deliver more customer-obsessed solutions.

As we step into this promising future, let this not be viewed as a threat but a call to action. Let's harness the power of GAI to augment our skills, add more value to our roles, and unlock new frontiers in the digital transformation journey. In the words of Ben Thompson, "The most successful companies of the Internet era have all been predicated on the power of one zero marginal cost resource or another; it's time to harness the power of AI, the next frontier."

What is my number one takeaway on AI in this paper? Don’t try to use the Grammarly AI app on mobile, it is terrible!