iiinova — where your digital self keeps evolving

Martin Chesbrough @mchesbro.bsky.social · 04/03/2026

At the weekend I announced that the Lunatics were taking over the Asylum. So now I'm back. The first part of this post is a re-framing of the Asylum. The second part is an escape plan. If Chief Bromden can escape then so can you. martinchesbrough.net/a-way-out-of...

martinchesbrough.net

A Way Out of Lunacy

A week ago I wrote a post entitled “The Lunatics are taking over the Asylum”, which outlined some aspects of the current lunacy. The…

010

Martin Chesbrough @mchesbro.bsky.social · 28/02/2026

It took me a while but …. 🤪

000

Martin Chesbrough @mchesbro.bsky.social · 28/02/2026

I think the AI lunatics might just be right … kind of martinchesbrough.net/the-lunatics...

martinchesbrough.net

The Lunatics are taking over the Asylum

Why Should You Read This?

000

Martin Chesbrough @mchesbro.bsky.social · 15/03/2025

The argument is more Layers AND than Layers OR We need to evolve data architectures.

000

Martin Chesbrough @mchesbro.bsky.social · 15/03/2025

I open up my browser and search for “data architecture” and back comes articles talking about Medallion Architecture, Source-Warehouse-Consumption Layers and so on. I don’t see many alternatives. Are we allowed to have a different opinion?

110

Martin Chesbrough @mchesbro.bsky.social · 15/03/2025

It probably started around 10 years ago when every Big Data/Hadoop project I was involved in seemed to mimic the good ol'data warehouse architectures of the 1990s. I was frustrated ...

100

Martin Chesbrough @mchesbro.bsky.social · 15/03/2025

I've been meaning to write this for a while - it is titled "Against Layers" and is a plea for pluralism in data architecture.

120

Reposted by Martin Chesbrough

Robin Davis @robind100.bsky.social · 18/12/2024

If you work with groups and teams, the Fun Retrospectives website is an excellent resource for ideas. I enjoy using my favorite techniques and experimenting with new ones to keep things exciting. #QiSky #Facilitation

buff.ly

Futurespective | FunRetrospectives

Have fun, learn from the past and prepare for the future!

065

Reposted by Martin Chesbrough

ClickHouse @clickhouse.com · 19/12/2024

A ClickHouse query optimization guide. SQL-based observability in the wild. PIVOTing in ClickHouse, and more. It must be our final newsletter of 2024! 🎄

buff.ly

December 2024 newsletter

Welcome to the December ClickHouse newsletter, which will round up what’s happened in real-time data warehouses over the last month.

052

Martin Chesbrough @mchesbro.bsky.social · 15/12/2024

I’m not saying that building a better LLM model is not important. Of course it is. But for GenAI app developers it is less crucial. Designing the app is paramount.

130

Martin Chesbrough @mchesbro.bsky.social · 15/12/2024

When I look at the features that OpenAI, Google and Anthropocene are launching they are more dependent on app features than LLM model

100

Martin Chesbrough @mchesbro.bsky.social · 15/12/2024

My argument was that I can build an app, swap out GPT4o for Claude 3.5 and get similar results. You get better results from better prompting. Or getting another model to check your results.

100

Martin Chesbrough @mchesbro.bsky.social · 15/12/2024

I was talking to a friend who’s very into Gen AI at a recent Xmas party and I put it to home that my views is that the LLM model “wars” are now pretty well over …

110

Martin Chesbrough @mchesbro.bsky.social · 13/12/2024

- then I moved onto asking Gemini to read the code on my screen (in VSCode) and tell me what it did - it did that pretty well (after an initial hallucination) It's exciting!!!

000

Martin Chesbrough @mchesbro.bsky.social · 13/12/2024

- then I got it to explain the paper and research similar papers (ScholarGPT does this but not from a shared screen - you have to copy-paste text) - and I asked it to identify gaps in the literature and research (it did a reasonable job)

110

Martin Chesbrough @mchesbro.bsky.social · 13/12/2024

I've had a lot of fun today playing with Gemini 2.0 through aistudio.google.com - I got it to read an academic paper on my screen (Claude does this as well but Gemini does it better)

aistudio.google.com

Google AI Studio

Google AI Studio is the fastest way to start building with Gemini, our next generation family of multimodal generative AI models.

110

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

I don't agree with everything Andrea Gioia writes but I do think this book is a valuable addition to the body of knowledge on data products, data management and data architectures. Go read it!!

000

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

This is a lot of heavy stuff to get into with the first 3 chapters of a book.

100

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

I'll provide a word of warning as well - within the first 3 chapters we get a diagnosis of data platforms using System Dynamics, we get an interpretation of the organisation that builds them using Viable Systems Method and we get straight into Hexagonal Architecture as an approach to design.

100

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

Let me provide an excerpt from Chapter 1 to do with the failure of data platforms as monolithic architectures

100

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

There are too many books written that simply explain how to do things. Now that's not a bad thing - the last book I mentioned, "Deciphering Data Architectures", is such a book. These books teach you stuff which is useful. But "Managing Data as a Product" challenges you to re-think what you know.

100

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

Why should you read this? If you liked Zhamak Dehghani's book on Data Mesh then this is (in my view) a book of the same ilk. It is not overtly data mesh, although it makes plenty of data mesh references. "Of the same ilk" means that it challenges existing thinking on data and I like that.

100

Martin Chesbrough @mchesbro.bsky.social · 10/12/2024

Another day, another book recommendation. This time it is "Managing Data as a Product" by Andrea Gioia, published by Packt www.packtpub.com/en-au/produc...

100

Reposted by Martin Chesbrough

Kate Raworth @kateraworth.bsky.social · 25/11/2024

Calling univ profs & lecturers! If you want to bring concepts from Doughnut Economics into your teaching, we've just launched a webpage bursting with resources for you: 7 slide decks, reading lists, videos & activities - all open access. Dive in & pls reshare! doughnuteconomics.org/university-c...

A screenshot of slides drawn from 7 slide decks, showing the range of topics addressed on the webpage

19447186

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

My advice is that before you decide to build a data xxxx have a good think about what problem you are solving and the pros and cons of xxxx

030

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

This diagram is a good example - I like it conceptually but if you look closer it gets a few things wrong. Doesn’t change the core message though.

120

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

I will offer some words of caution. I don’t agree with everything James says. I suggest readers also do their own research and check things.

100

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

At the end of the day what “Deciphering Data Architectures” does is give you (the reader) a level of knowledge and some tools to understand the options. In my work there is too much jumping in to build a data xxxx and too little understanding of what are the different options, pros and cons!

100

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

But this is what I’d call a broad book. It covers everything at about 1 inch deep. And that is important. For example in the data ingestion section it never gets into much details on streaming ingestion. It doesn’t need to - that’s covered by other books.

100

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

It’s not a long book (275 pages) but gets into a lot of topics in those pages. James even has a chapter (chapter 8) on approaches to data modeling. I like that he includes that. He misses a few things, like semantic modeling, OWL, RDF and ActivitySchema to name a few important concepts.

100

Martin Chesbrough @mchesbro.bsky.social · 08/12/2024

Because of a discussion on misuse of the term “data warehouse” I referenced a section from James Serra’s book “Deciphering Data Architectures”. That led me to re-read the book. It’s a great book BTW and worth a read.

120

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

That perform tasks in concert with humans. Think pair programming with an infinite array of AI or software agents.

000

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

I haven't created a Wardley Map for agent-oriented software but if I did it might look a little like that one above. I'm arguing for fast experimentation with agent-based software. An infinite number of composable software agents.

120

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

Which actually gets me back to Simon Wardley in a weird way. In the same video as he explores AI as a new theocracy he also talks about his experiences with Tudor Girba and moldable development. Here's the Wardley Map he draws. What if, instead of software tools the blue area was software agents?

210

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

So, if enterprise software giants like Salesforce are rushing towards agents AND the current state of multi-agent systems is still immature. Does that mean that the enterprise software players actually have the most to lose if they don't enter this nascent space? I think so.

110

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

Well actually no. It seems enterprise giants like Salesforce have been thinking this for a little while. See github.com/SalesforceAI...

salesforce-research-dei-agents.github.io

Salesforce Research DEI Agents

Project page for Salesforce Research DEI Agents.

100

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

Is it possible that the academic community has been way ahead of us all this time and just waiting for commercial software and commercial AI to catch up? It certainly feels a bit like this.

100

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

Which also led me to think about how developing agents (whether AI or not) would change software development. I went on a little side-track with agent-oriented programming and multi-agent systems which has been a (mostly) academic topic for some time.

110

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

And start to play with building AI agents bsky.app/profile/mche...

100

Martin Chesbrough @mchesbro.bsky.social · 30/11/2024

This led me (via a rather circuitous route) to investigate the AI stack, which Andrew Ng has rather neatly diagrammed in a recent talk. youtu.be/KrRD7r7y7NY?...

100