Simon Griffiths

What AI Changes About Building Blogs

Wed, 03 Jun 2026 09:00:00 +0100

WordPress powers around 43% of the web. For years it was the default answer to “how do I start a blog?” It was mine too.

This is not a post about why WordPress is bad. It is about something its pricing quietly depends on: the belief that leaving is expensive.

Agents Don't Replace APIs. They Expose How Weak Most APIs Already Are

Tue, 02 Jun 2026 09:00:00 +0100

There is a growing narrative that AI agents, often coupled with things like Model Context Protocol, will replace APIs.

It is easy to see why that idea has taken hold. Agents can discover tools, reason about which one to call, and assemble workflows at runtime. That feels very different from the static integrations we have lived with for years, where one system calls another through a fixed endpoint, with a fixed payload, in a fixed sequence.

AI Coding Tools Are Not the End of Craft

Fri, 29 May 2026 09:00:00 +0100

This is the first post in a short Friday series I’m calling Craft & Code: what carpentry can teach us about AI, skill and the future of software.

When I was six, my father’s workshop felt like a magical place.

The Microservice Debt Collectors Have Arrived

Wed, 27 May 2026 09:00:00 +0100

I’ve been spending a lot of time recently thinking about what happens when AI agents are pointed at real production systems — not the clean demos, but the actual accumulated architectural decisions of the last decade. What I keep coming back to is this: the software industry has been running a tab. We decomposed systems that didn’t need decomposing, distributed data that worked perfectly well together, and replaced database-enforced integrity with the optimistic assumption that well-behaved services would keep everything consistent. The tab wasn’t obviously wrong — some systems genuinely needed this architecture, and the teams building them had the discipline to make it work.

Building Your First Real GPT Is Not a Prompting Exercise

Tue, 26 May 2026 09:00:00 +0100

I recently built my first non-trivial GPT.

The interesting lesson was not about clever prompting. It was almost the opposite.

The GPT only started to become useful when I stopped treating it as something I could configure with a good instruction and a pile of documents, and started treating it as a small software project: source material, structure, behaviour, tests, iteration, version control.

The antenna, the inter-library loans and infinite patience

Thu, 21 May 2026 23:30:00 +0100

When I was about sixteen, I decided I wanted to build a VHF antenna.

The problem was that I had absolutely no idea how.

This was before PCs were common, before the internet, in the era of the ZX80 and VIC-20. If you wanted to learn something specialised, access to information was often the biggest obstacle.

Your Cloud Is Not a Backup

Tue, 19 May 2026 05:41:52 +0100

There’s a quiet assumption many of us make about our digital lives: that because our photos are in iCloud, our documents in Google Drive, and our files syncing to OneDrive, they’re safe.

The Other Logins You’ve Forgotten About — Apple, Facebook, and Who’s Still Watching

Mon, 18 May 2026 06:22:00 +0100

If you read my earlier piece on Google account permissions and went away to check your list, well done. But Google isn’t the only door you may have left open.

Most people who’ve been online for more than a few years have also clicked “Sign in with Facebook” on things they’ve long since stopped using. And more recently, if you’re in the Apple ecosystem, you’ve probably used “Sign in with Apple” too — possibly without fully understanding how it works, or where to find the list of everything you’ve connected.

At 61, AI Gave Me a Second Act

Fri, 15 May 2026 05:39:57 +0100

I’m 61. I wrote my first line of code at 15. I have a VP-level job, two dogs, a new wife, and a blog I’m somehow managing to post to almost every day.

The Keys to Your Google Account — Do You Know Who Has Them, and Why You Should Check Now

Thu, 14 May 2026 05:00:00 +0100

Go to myaccount.google.com/permissions right now. Count how many apps are listed.

I’ll go first. I found seventeen. And I say that as someone who has never been particularly keen on Google sign-on — I’ve always had a mild instinct against it, preferring to create separate accounts where I could. Seventeen anyway. A decade of occasional lapses, free trials, and moments of convenience adds up.

Why Apple Pay Actually Works — and it matters to you now

Tue, 12 May 2026 04:39:21 +0100

In a recent post I argued that you should replace your debit card and switch to Apple Pay or Google Pay for online purchases. A few people asked the obvious follow-up: why? What actually makes tapping your phone safer than tapping a physical card?

Why You Should Change Your Debit Card Now

Tue, 12 May 2026 04:09:10 +0100

There’s a piece of advice you won’t hear from your bank. Not because they don’t know — their security teams absolutely do — but because saying it out loud is uncomfortable when you’re also in the business of reassuring people that everything is fine.

NVIDIA Didn’t Miss the Data Problem — They Drew a Boundary Around It

Fri, 01 May 2026 14:41:58 +0100

I’ve been looking at some of the recent AI architecture models coming out of NVIDIA. They’re polished, coherent, and very strong on infrastructure. But something stood out immediately: they largely skip over the hardest part of enterprise AI — the data.

AI Didn’t Break Infrastructure — It Exposed Its Edges

Wed, 29 Apr 2026 20:04:21 +0100

For a long time, the industry settled into a comfortable model for infrastructure.

Two CPUs. A standard server. Add more boxes when you need more scale. Keep everything broadly interchangeable. Let software do the heavy lifting.

What’s Next for Coders?

Tue, 28 Apr 2026 09:27:19 +0100

In software development there has always been a tension between the effort invested before coding and the effort spent doing the coding. In mainframe days, the actual job of coding was extremely time-consuming — you had to wait overnight for a compilation, and a test run had to be planned days in advance.

How do you keep up with AI?

Fri, 24 Apr 2026 17:26:45 +0100

Earlier today, I looked at some work I did in ChatGPT in late February — about eight weeks ago — and it immediately struck me that the work needed to be redone. Not because it was bad at the time, but because the models have made huge steps forward in just a few weeks. The prompts I wrote then were less refined, and the outputs that seemed impressive two months ago now look noticeably rough around the edges.

AI Is About to Expose Your Security Posture — Not Improve It

Thu, 23 Apr 2026 10:28:00 +0100

There’s a growing narrative that AI can now scan codebases, find subtle bugs, and even uncover vulnerabilities that humans have missed for years.

You’ll see this framed as progress. Better tools. Safer software. “Many eyes,” but automated.

How to Build Good Software with AI — Not How You’ve Been Told

Wed, 22 Apr 2026 00:07:48 +0100

In a previous post, I wrote about the current narrative around AI agents “coding for hours” and what’s actually going on behind the scenes. If you haven’t read that, it’s here: https://simongriffiths.io/2026/04/21/ai-agents-coding-for-hours-whats-really-going-on/

That post was really about demystifying the claim.

AI Agents Coding for Hours — What’s Really Going On?

Tue, 21 Apr 2026 23:32:14 +0100

There’s a growing narrative that AI agents can now “code for hours” and build meaningful applications with minimal human involvement

I’ve spent a bit of time looking into this, and I found myself slightly unconvinced—not because it isn’t happening, but because the framing doesn’t quite line up with how software engineering actually works.

Why I Stopped Letting AI Run SQL

Mon, 20 Apr 2026 12:24:45 +0100

AI can write SQL for you and run it immediately. That’s the problem.

Much of the recent excitement around AI and databases has focused on MCP (Model Context Protocol) — tools that let you “talk to your data” using natural language. The LLM translates your request into SQL and executes it on the spot.

Auth0 + ORDS on ADB - Playbook

Thu, 16 Apr 2026 15:47:30 +0100

Step-by-Step Tutorial: Configure Auth0, Create an ORDS JWT Profile, and Secure an Endpoint

[ there is a companion GitHub to this post ]

This is the hands-on companion to the architecture write-up . If you’ve read that one, you already know the shape of what we’re doing: Auth0 issues a token, ORDS validates it, ADB serves the protected data. Here we’re actually going to build it.

Auth0 + ORDS on ADB - Architecture

Thu, 16 Apr 2026 15:29:39 +0100

Why This Pattern Works: Auth0 as Identity Provider, ORDS as Resource Server

[ also see the companion GitHub repo ]

I spent longer than I’d like to admit trying to secure an ORDS endpoint with Auth0 before I realised my mental model was wrong. I kept treating ORDS like it needed to be part of the identity story, and wondering why the pieces didn’t quite click.

Building Apps on Oracle with AI on a Mac

Tue, 07 Apr 2026 12:45:00 +0100

The most talked about way of connecting to an Oracle database from AI, whether Codex, Claude Code or another AI is MCP - it’s great and enables you to easily connect and query the database with natural language. But what if you are building an app, it’s not about ad-hoc queries, the focus is on repeatability, capturing exact commands, repetitive testing and build scripts. MCP is actually not the best approach - sometimes existing techniques work just as well, and in this case better.

Business Rules and AI

Mon, 06 Apr 2026 22:35:46 +0100

We are swiftly moving to a world where businesses will see AI agents as a primary channel for business processes. Chains of agents will push through a process from begin to end, monitor the process, chase up when processes fail and deal with exceptions.

AI Trust and Data Security

Wed, 25 Mar 2026 11:11:41 +0000

We have all heard the stories about AI agents going “rogue” and releasing information that is commercially sensitive or private and with the unpredictability of LLMs, this is a real risk for any AI systems that try to provide an organisations data for use by an agent using an LLM.

Agent SKILLS for Oracle database

Wed, 25 Mar 2026 10:18:38 +0000

While we can all building applications at light-speed with AI, ensuring that the code produced follows best practice is more difficult. LLMs are fundamentally based on public code, and code that is public is not always based on best practice. Code that is truly scalable, fully secure and efficient is not always the default we see in public. Common example code is often simplified and designed for ease of use, not for performance.

Data Lakes, Operational Reporting and Enterprise BI

Sat, 24 Jan 2026 23:33:12 +0000

There has never been a time that data hasn’t been recognised to be as important as it is today. It seems that the world has woken up to the potential value that an organisation’s data can offer - not only for external monetisation, but also for understanding the internal organisations’s workings. The discussions around machine learning and artificial intelligence using data to automatically respond to situations and events offers even greater potential for the future.

AI and Data

Sat, 24 Jan 2026 23:19:40 +0000

The use of AI to write and maintain computer code has already become the industry norm - there can hardly be any programmers out there who are no seeing some advantages of AI. While there is still dispute how much the AI is actually helping or proving value, the situation is changing every week with advancements not only in the foundation LLMs, but in the various IDEs, AI tools and AI enabled editors.

Oracle SQL and MCP

Sat, 24 Jan 2026 21:51:05 +0000

All of the main LLMs are pretty good at generating SQL, and they also have a pretty good understanding of the Oracle Data Dictionary. This means that if you can connect your LLM chat to an Oracle database, you can query, manage and even tune your Oracle database through simple language rather than needing to actually write any SQL.

Microservices and Devops ? What would Adam Smith think …

Thu, 19 Dec 2019 09:24:59 +0000

Some history

It is no exaggeration to say that the relative prosperity of the world today is fundamentally based on the division of labour . The division of labour enabled us to move from an agrarian society to build towns and cities, and then powered the industrial revolution. Adam Smith’s Wealth of Nations put the division of labour and specialisation as one of the foundations of the industrial revolution. Today’s modern society means that every one of you reading this will have already specialised and will not be subsisting on what you can grow in your own garden.

Removing the Object Relational Impedance

Sun, 15 Dec 2019 10:00:00 +0000

Four Model

In a previous article I looked at why object-relational impedance was still a thing, and why we need both ways of holding data - object model to facilitate the easy interactions needed for user-facing applications, and relational for reporting and analytics on the same data.

Why developers and DBAs just don't get on

Sat, 14 Dec 2019 19:17:00 +0000

Introduction

The enmity between developers and DBAs is both legendary and persistent. Developers get frustrated when DBAs take so long to do the simplest of tasks, and DBAs are disparaging of developers who want the database to do nothing more than be persistence for their app.

Bridging the Gap between Data Discovery and Execution

Fri, 13 Dec 2019 21:17:41 +0000

Overview

The introduction of a discrete environment dedicated to data discovery and model building (e.g. a Datalab) can substantially deal with the lack of agility of more traditional data warehouses where agility has been deliberately locked out to deliver the resilience and reliability required by a mission-critical data warehouse. However, this does introduce a new challenge - i.e. how can we quickly and effectively promote the value discovered in the data discovery into the mission-critical execution environment of the data warehouse?

How to optimise your read-only Bigdata storage model

Thu, 12 Dec 2019 22:32:34 +0000

Introduction

Discussion of data models in Bigdata environments usually centres around technology focused question - file formats, what compression to use, which columnar store is preferred, whether the file structures are splittable and so on. While these are valid and important questions, there are more fundamental questions to be asked on the higher level data structures that should be built using those various technologies.

Operational Reporting and Enterprise BI

Wed, 11 Dec 2019 23:29:18 +0000

Just what is Enterprise Architecture?

Tue, 10 Dec 2019 22:43:31 +0000

Yesterday, there was a question on an enterprise architecture discussion forum asking whether Powerpoint could have any value as an EA tool? I was interested to read the comments where there was a fairly strong viewpoint from many of the posters that any EA work done in Powerpoint could never have any value as enterprise architecture, and that Powerpoint itself offered zero value.

Data Lake Top Tips

Mon, 09 Dec 2019 22:27:38 +0000

I’ve called this Data Lake Top Tips because if I called it Data Warehouse Top Tips, then nobody would read it (because nobody is building Data Warehouses any more, right?). The reality is that Data Lakes, Data Warehouses, Data Marts and Data Science (and others) all have to work together to deliver the business need, and so the real top tips are how all those can work together.

Building a Data Driven Organisation

Sun, 08 Dec 2019 16:55:53 +0000

It’s obvious to everyone that we are currently going through a huge shift in the way that IT is used within our lives. The term “paradigm-shift” is much overused, but in this case it’s probably warranted. We all carry around a mobile device - perhaps more than one - and gradually our homes and cars are becoming digital arenas with our cars permanently connected to the carmaker and our homes even embedding devices that we can talk to and they can talk back. Even within the previously staid world of back-office IT, we are seeing the encroachment of mobile-first interfaces, deeper automation and API driven chatbots. All of these generate huge volumes of data that we are only just beginning to work out how to manage.

Data Platform

Sat, 07 Dec 2019 21:41:50 +0000

Introduction

There’s no question that data is becoming an important topic in almost every organisation. For some organisations data will be seen as a challenge - as security breaches, privacy violations and data misuse appear on the news almost every day; for other organisations, data is an opportunity as new startups look to use data in new ways and established organisations look to monetise their current data. What is common across all organisations is the recognition that data cannot be ignored - data doesn’t look after itself and the value inherent in data cannot be realised without both up-front investment and ongoing management.

Friday night moan ….. Fashion driven IT

Fri, 06 Dec 2019 23:50:07 +0000

While I’ve been researching my articles over the past week, one thing that struck me was how much we seem to have diverged from a scientific approach to IT development and switched to a world that can probably be best described as being driven by fashion. At the extreme, advocates for new methodologies or tools reach levels of passion usually reserved for religion or politics with much noise, but little rationality.

Microservices and GDPR

Thu, 05 Dec 2019 21:53:38 +0000

Well, there’s two words you don’t often see together. What has an application development methodology have to do with European data protection laws? In fact, rather a lot. Let’s start with a quick recap of GDPR.

Securing your data

Wed, 04 Dec 2019 23:55:36 +0000

We all know that cyber-crime is increasing every year and at the moment we seem to be losing the battle against the blacks out there. Clearly the solution is to brick up the cyber walls and hope for the best ? But is that really the best approach?

Microservices and Business Processes

Tue, 03 Dec 2019 16:33:15 +0000

In yesterday’s article I discussed the challenges of co-ordinating activities across a complex set of microservices, and I touched on the subject of end-to-end view of customer journeys and complex business processes, but I didn’t really cover that in any detail, so let’s look at what we need, and how that could be achieved.

Some thoughts on Orchestration and Choreography

Mon, 02 Dec 2019 12:54:15 +0000

Analogies are risky things, there’s always a chance that if you stretch one too far, it breaks and disproves the point you’re trying to make. What’s even worse is when an analogy needs to be explained, and therefore the analogy is pointless. And here with orchestration and choreography we have a case in point, not only does the analogy need explanation but it completely misses the point.

Architecture Advent

Sun, 01 Dec 2019 23:25:36 +0000

At this time of year, the demand for enterprise architecture seems to drop precipitously. Most companies have the dual challenge of meeting huge challenges, but with reduced resources as everyone tries to take some well deserved holiday. Change freezes have been known to start in November, and for retail, logistics and utility companies, Christmas planning will have started many months before. So as these efforts step up, it’s not really surprising that long term strategy takes a back seat.