Rob Bowley
@robbowley.net
2.2K followers 1.1K following 5.1K posts
Product & Tech Leadership Advisor, Consultant, Coach & Mentor Tech, Software Development, Science, History, Economics, Politics https://blog.robbowley.net https://www.linkedin.com/in/robertbowley https://pragmaticpartners.co.uk Manchester, UK
Posts Media Videos Starter Packs
Pinned
robbowley.net
Lots of new followers so thought I'd do an intro 🙂

This is me, without hat, in my garden, on a rainy morning in Manchester 🐝, UK

Been working in tech for 25 years - software engineer then various leadership roles. Nowadays I'm a product & tech leadership advisor, coach, consultant.

1/4
Selfi of Rob Bowley on a rainy morning in Manchester with his garden in the background
robbowley.net
@gergely.pragmaticengineer.com you were saying the other day the JLR incident deserved more visibility than it was getting. It is wider than that - all the major cyber incidents in the UK over the last 6 months have a common thread
robbowley.net
AI doesn’t mean you’ll be able to leapfrog best practices in engineering – more likely, jump off a cliff
robbowley.net
A conference with Eliyahu Goldratt memes is my kind of conference

#TheoryOfConstraints #FastFlowConf
A slide showing Eilyah Goldratt next to a signed - Warning Local Optimisation Ahead
robbowley.net
Keen to know more - i have seen multiple sources suggest GPUs under high load burn out quick (if that's what you mean), but also others refuting it
Reposted by Rob Bowley
robbowley.net
The dotcom bubble left us fibre that still powers the internet.

The AI boom might leave us a pile of short-lived silicon and silent cathedrals of compute – monuments to a forgotten era.

The web was open. AI isn’t. That may prove the real difference.
After the AI boom: what might we be left with? | Rob Bowley
blog.robbowley.net
robbowley.net
Does this make me internet famous now?
robbowley.net
Interesting waking up to 300+ comments on HN for a post I threw out in between child wrangling on a Sunday to get stuff out of my head

news.ycombinator.com/item?id=4556...
robbowley.net
The dotcom bubble left us fibre that still powers the internet.

The AI boom might leave us a pile of short-lived silicon and silent cathedrals of compute – monuments to a forgotten era.

The web was open. AI isn’t. That may prove the real difference.
After the AI boom: what might we be left with? | Rob Bowley
blog.robbowley.net
robbowley.net
2 likes here; 200+ comments on HN. I have no idea how it turned up there 🤷‍♂️
robbowley.net
The dotcom bubble left us fibre that still powers the internet.

The AI boom might leave us a pile of short-lived silicon and silent cathedrals of compute – monuments to a forgotten era.

The web was open. AI isn’t. That may prove the real difference.
After the AI boom: what might we be left with? | Rob Bowley
blog.robbowley.net
robbowley.net
Went with my son to 6th form open days. The Computer Science workbooks mentioned Waterfall, Spiral, RAD, Agile and even XP – but interestingly, not Scrum. That probably dates it to around 2003–2006, when Agile was new and RAD was still hanging on.

So 20 years out of date 😳
robbowley.net
Came to make same point you beat me to it 👏
robbowley.net
The AI boom risks leaving us with a pile of short-lived silicon and data centres that may be difficult to repurpose once the current wave passes.

6/6
robbowley.net
While some of the supporting power and networking infrastructure will remain useful, much of what’s being built is highly specialised and tightly coupled to the current generation of AI hardware.

5/6
robbowley.net
AI data centres also need to be built for extremely high power density and advanced cooling, making them harder to repurpose if demand or chip architectures shift.

4/6
robbowley.net
In contrast, much of today’s AI investment is going into expensive GPUs with a lifespan of just two to four years. These chips not only become obsolete quickly but also physically wear out faster under constant high load.

3/6
robbowley.net
However, the dotcom overbuild left durable infrastructure - fibre networks and interconnects - with multi-decade lifespans. Those assets became the backbone of the broadband and cloud eras.

2/6
robbowley.net
Some people are saying the potential overbuild because of AI might not be a bad thing because, like the dotcom era, it could leave behind infrastructure we’ll benefit from for a long time.

1/6
Reposted by Rob Bowley
edzitron.com
Premium: The AI Bubble's promises are impossible. NVIDIA's customers are running out of money, GPUs die in 3-5 years, most 1GW data centers will never get built, and OpenAI's Abilene data center doesn't won't have the power it needs before 2028 - if it ever does.
www.wheresyoured.at/the-ai-bubbl...
The AI Bubble's Impossible Promises
Readers: I’ve done a very generous “free” portion of this newsletter, but I do recommend paying for premium to get the in-depth analysis underpinning the intro. That being said, I want as many people ...
www.wheresyoured.at