Відео

The bane of picking off the low hanging fruit has left us with the need to think about tackling the real problems

This is a challenge on more than one front. 1) The knowledge retained in the minds of individuals, rather than being documented or shared in a Wiki, is a significant loss to the organization. When these individuals leave, they take this valuable knowledge with them, creating a gap that is hard to fill. even if they haven't left the organisation, just gone elsewhere. 2) As complexity increases, there is a move to make domains smaller, such as a move from Project to Product, Big Data to Data Mesh, Monolith to Microservices. But all these approaches make technology estates considerably larger taking more resources to manage. But now AI has appeared and the obvious strategy from many organisations wanting to capitalise on this advance is to 'feed the beast' but with what? Where was that expertise captured? I think capturing knowledge is going to become our next hurdle, and what is damaging is that we think we are all doing it so well!? It's somewhere a simple 'Google-like' search can find it right?! IDC Surveys - Employees spend an average of 2.5 hours per day, or roughly 30% of the workday, searching for information. McKinsey Global Institute - Knowledge workers spend about 19% of their time searching for and gathering information. Forrester Research has noted that poor information management can lead to a loss of up to 20% of an employee's productivity, which translates to about one day per week.

Another sage observation… the herb, that this!

In October 2022, I wrote a blog post for the Ordnance Survey Tech Blog on using the Obsidian Note-Taking app with Open AI GPT3 Large Language Model to generate text. Back then, GPT3 was only accessible through their API, and I remember expressing my wish to my colleagues for a more user-friendly approach. Little did I know, a month later, we were introduced to ChatGPT, a less complex front-end to their API, launched by a group of researchers at OpenAI. The speed of progress was astounding. Ordnance Survey was a great place because it was AI-crazy; well, you can't manually process 20,000 changes per day; that ship sailed long ago! Far from an IT backwater, As AI permeates every business sector, it's crucial for professionals like architects to stay abreast of these advancements. However, with the constant flux, it's a challenge. So, how do I manage? I've enlisted the help of a couple of AI enthusiasts who keep a vigilant eye on Gen AI, using their journalism skills as my filter. I use PKM tools (Affine.pro is my tool of choice) to work things out and see how they link together, and I use infranodus.com to explore ideas, which is excellent for finding holes. what is happening? just one word .... Adaption.

You're absolutely right.

Fully agree and well put Paul!

Totally agree. Thanks for publishing this. Super important that architects think about this

Good points, as usual, Paul. So far, I’m not seeing many AI architects. Is there a plan to grow that capability?

I love this! I'm playing around with the idea of a 'Multibase' which is a patchwork of data from all sorts of organisations linking it together, so interconnecting queries are possible. Now with the sophistication of Large Language Models adherence to a strict set of standards may not be as important because it would provide not only translation but interpretation reducing the amount of complexity involved. With a Multibase we could ask it to help us solve big questions such as climate change and sustainability. Thanks to the Internet, creating something like this is possible if we have the imagination, which more and more is turning into the only barrier.

This is one of the conversations I was having with @GeekyAdams the other say

Crowdsourcing, shadow IT or empowering the business to find its own solutions. Its a tough one this - but of course in heavily regulated industries there are serious consequences - like the fine you talked about at the start! We've been meaning to do some content on pragmatic governance for a while - thanks for the reminder / inspiration!

Understanding the legacy architecture can be useful to understand why something 'is'... And perhaps avoid / reduce the risk of making the mistakes of the past!

Totally agree Paul. Business architecture doesn't get the recognition and attention that it deserves! Perhaps because it involves the mushy component between the keyboard and chair... are Enterprise Architects simply more comfortable in the other domains? Or are they not invited to contribute?

Love these "rants" Paul. Business Architecture is important.

Nice video - having some humility and playing to the strengths of the team - like it. BTW your audio is a little quiet - I'd recommend a Rode Mic if you are using a proper camera or a HyperX quadcast if you are using your computer to capture these.

Can you explain why the opposite of a principle has to be a sensible course of action? I'm struggling to follow the logic for that (maybe the semantics of the statement have been lost on me). This video immediately led my thoughts down the path of wondering why you wouldn't just go with the opposite if it's sensible. Or, even more fundamentally, why would you even bother having a principle if it's apparently sensible/ fine to go in completely the opposite direction. What makes the principle more favourable than the opposite? Maybe the problem is with the words "sensible" and "opposite". Is this like when someone approaches you with an implementation they're proposing and you realise it'll work, but there's a better way? Something like: "Both options A and B will deliver the same observable outcome to a user, however, when you consider knock-on effect X, Y and Z (which might be non-functional impacts such as maintainability, clarity, etc), it's clear option A is better"? Otherwise, the "principle" is unhelpful in that it states nothing more than common knowledge; there's no reason anyone would ever consider doing anything other than the principle so it doesn't need to exist. Even then, common sense isn't always so common. I'd echo Architect Tomorrow's point. Maybe it's more "Architectural principles should be used to provide guidance towards a preferred solution where there are multiple functionally viable/ feasible _alternatives_ that may result in subtle, suboptimal outcomes that aren't immediately obvious".

Totally agree - as the saying goes if you can't explain it simply you don't understand it well enough!

Thought provoking Paul. I'd argue feasible rather than sensible? But perhaps I am being pedantic or obsessing over semantics?!