January 16, 2024 | Posted in News
Good technology disappears. As writers have been saying for some time, when we progress a piece of technology – or indeed any kind of service – to a point where it is at or near perfect, we tend to find that it disappears and slips into the fabric of how we live and work at a higher (or lower) level.
Very few people now light a fire to cook or keep warm, we now just turn the heating on because our home HVAC systems have become a comparatively invisible part of the fabric that makes up our homes. Good technology disappears and becomes a lot like a utility and – as current writers are saying all the time – it vanishes and becomes part of the lower substrate layer of application functions that we all use.
Of course, nobody would think that of much of the technology that surfaced throughout 2023. Such was the ‘noise’ created by new strains of generative Artificial Intelligence (AI) that many people temporarily forgot that we had been working to develop AI engines and models for many years. But the shift from traditional ‘old-fashioned’ predictive and reactive AI to new strains of generative intelligence would not quiet down; even as the year has closed out we have heard more and more about AI from the major cloud hyperscalers AWS, Google and Microsoft, all of whom tell us that their datacenter services are aligned for AI excellence like no other.
As we move towards 2024, many agree that the Large Language Model (LLM) technologies that gave generative AI its apparent breadth will give way to the use of industry-, task- and role-specific language models that are sharper, more functional and (hopefully) easier to put data guardrails around. Once again, we had seen this suggested back in the spring in line with comments relating to vector databases, but you can hold a good recurrent theme down one imagines.
Looking ahead, technology will do what it always does, it will be tuned, trimmed, tailored and tooled-up. What does that mean? It means we will take generalized applications (or cloud services, or Application Programming Interfaces, or databases, or connective network layers… or other) and hone them for more exact users, exact tasks, exact workflows, exact industries and more exact use cases at the deepest level. For example, retailers won’t just use a Point of Sale (PoS) software system to make orders, track inventories and process sales – instead, clothing retailers will use clothing retail software systems, brewers will use brewing software and golf shop owners will use golf applications.
In truth, this drive for application specificity has been around since the 1980s if not before, but the wider trend towards refinement and precision-engineering will deepen. This deepening will happen way below a retailer’s app, it will be applied to the workflow management layers that businesses in any given space run on and, crucially and fundamentally, it be applied to the exacting (there’s that work again) alignment of how every micro-component of enterprise software architecture is built, executed and managed.
What all this means is a progression where IT gets less dramatic (let’s face it, generative AI made it onto news shows that your grandparents might have seen) and a whole lot more pragmatic. Dr Scott Zoldi, chief analytics officer at credit scoring services company FICO agrees that this trend will be particularly born out in the gen-AI space.
“While its impact has been unprecedented, in 2024 generative AI will follow in the hype-cycle footsteps of other breakthrough technologies like blockchain,” said Dr Zoldi. “At their outset, both technologies appeared to be powerful novelties with great but unknown potential. As blockchain has matured and been applied in extremely useful ways beyond cryptocurrency – such as for model management governance – gen AI will find similar tributary applications that will be less dramatic, but far more pragmatic.”
While we didn’t use the term ‘workflow’ as extensively as we do today, it’s now not only useful, it is (arguably) essential. Why is this so? Because in the digital age of automated workplace functions where machines (hardware machines and software robots) are taking over more of our workplace functions, we need to be able to give part of job A to employee C and part of it to machine X and another section to software robot X. That means that AI itself is going to have to get a lot better at solving real world problems rather than showcasing its ability to create a fanfare.
“Across the technology spectrum from visual communications and design to software coding, the generative AI that wins over users and scale will have to solve real-world problems. For startups, commercially-viable use cases will win in the fierce competition for venture funding and will be the platforms to develop the fastest,” said Cameron Adams, co-founder and chief product officer at design services company Canva.
As the trends continue to play out and align, we may see new job functions thrown up. Alongside the CIO, CTO, CISO and chief data officer (CDO), the chief AI officer (CAIO) may also come to the fore. Largely in the shape of a real world human being with arms, legs and a sentient functioning upper brain system located in the head, these people will oversee AI to track for bias, hallucination, skews in behavior and the application of AI automation itself.
Richard Timperlake, SVP EMEA at data streaming platform specialist Confluent thinks that generative AI will now become commoditized (which is a lot like it disappearing as it simply becomes an internal function) and embedded (again, it’s a disappearing act) in multiple applications. Why is this happening so early on? Because, he says, LLMs and other foundational models are already becoming much easier to train and fine-tune… and there’s that word tuning again.
“As we stand today, we can certainly say that all types of AI – especially generative – will continue to dominate the agenda. In the wider world, it faces legal and regulatory challenges that need to be addressed, not least in areas such as copyright. Businesses need to introduce due diligence before data is used – and some kind of remediation if it’s used incorrectly. Closer to home, businesses need to develop processes around how they build different models that will deliver value to their organizations and their customers. This can only be successful if the bedrock of source data is secure. After all, you can’t simply build a model and flood it with data without understanding the quality and veracity of the information.”
All well and good so far then, but as we get more and more used to applying AI in the workplace, will people start to calm down and stop worrying about which jobs are going to be replaced? After all, we know that AI is not intended to replace people, it is meant to shoulder the grunt/donkey work and enable humans to focus on higher-value tasks, right?
Asking us to be more realistic, Sridhar Ramaswamy, SVP of AI at data cloud company Snowflake says that some immediate concerns will be particularly challenging in the early years of widespread AI adoption.
“For a lot of people involved in what we loosely call ‘knowledge work’, quite a few of their jobs are going to vaporize,” warned Ramaswamy. “Rapid change makes it hard to quickly absorb displaced workers elsewhere in the workforce and as a result both the private sector and governments will need to step up. Finally, advances in AI will exacerbate the digital divide that has been happening over the past 20-30 years between the haves and have nots, and will further increase inequality across the globe. I can only hope that by making information more accessible, this emerging technology leads to a new generation of young adults who better understand the issues and potential, and can counter that risk.”
But really, however this discussion starts to thread and skew, it all comes down to data and there’s a strange dichotomy developing. Organizations are navigating a future dominated by artificial intelligence and are faced with the dual challenge of harnessing the power of data that fuels AI while grappling with the complexities of data management itself. This is the opinion of senior VP of product, Cullen Childress at observability and IT management company SolarWinds.
“AI has proven to play a critical role in helping IT professionals manage high volumes of data, optimizing database performance and improving overall business outcomes. But on the other hand, AI has collected and generated a mind-boggling amount of new data from diverse sources, leaving organizations faced with the challenges of how to manage it and understand where the data is coming from,” said Childress.
He suggests that the landscape is only becoming even more intricate and the journey toward effective data management is full of complexities. The challenges organizations are set to face are not only the looming depletion of high-quality training data to support AI development… but also the fact that IT teams have to manage vast and diverse databases sprawled across on-premises and cloud environments. “With the industry increasingly turning to observability solutions, teams are provided with a comprehensive and necessary view of what’s happening inside a company’s most imperceptible and valuable systems. Amid this transformative AI era, the success of any one company hinges on adeptly navigating the evolving landscape of data management,” added the SolarWinds VP.
If an observability specialist advocating observability on the road ahead isn’t too obvious for you, consider the fact that – even in 2024 – many of our IT systems will let us down.
Because your favorite airline, hotel, restaurant or other doesn’t always apply the right number of miles to your loyalty account, we can see that major enterprises today still struggle massively with core integration tasks. It’s often why ‘the system doesn’t work’ when users are trying to perform all manner of tasks connected to technology. Generative AI will suffer from the same type of disconnects in cases where things appear to be broken suggests Massimo Pezzini in his role as head of research, future of the enterprise department at automation platform company Workato.
“At least half of generative AI initiatives in 2024 will fail to deliver the anticipated business benefits due to a lack of integration. The potential business value of generative AI can only be fully unleashed if it is incorporated in end-to-end business processes. Otherwise, its contribution will remain limited to assisting individual and isolated tasks of marginal business value,” argued Pezzini.
As right as Pezzini is about the need for any business using modern technologies to adopt a comprehensive integration infrastructure, he does work for an integration Platform-as-a-Service (iPaaS) company, which gives us our final prediction for 2024 perhaps.
As we have suggested before, the IT industry will continue to conduct surveys. Sometimes packaged in sheep’s clothing as ‘market analysis reports’ or labelled with some other euphemistically shrouded and skewed term, the surveys will still come and the projections (usually called ‘findings’) they make will continue to arrive.
The results of next year’s technology surveys will find that DevSecOps specialists advocate a comprehensive developer+security+operations platform approach to architecting truly functional, secure and performant enterprise applications. Cloud observability specialists will conduct studies into Application Performance Management (APM) because that’s what observability used to be called and it’s not a bad way of validating why APM in the modern age needs observability. Serverless specialists will study the serverless space and recommend a serverless approach, containerization firms will analyze working application scenarios and probably consider promoting containerization as a key facilitator of efficiency. Finally, perhaps, enterprise software organizations that add generative AI functions to their platforms and tools will study the market and may end up proposing the use of generative AI in live production and deployment environments.
With all that IT evolution ahead of us, it’s good to know that some things never change.