1%

Time for one more email then I'll plug it in...

Tearing down the software development walls - Jon Ayre In software development, the handoff of work from Product Manager to Engineer to QA was always an anti-pattern. It's a shame it's taken AI to wake people up to that organisational religion. One size f...

In software development, the handoff of work from Product Manager to Engineer to QA was always an anti-pattern. It’s a shame it’s taken AI to wake people up to that organisational religion. One size fits all is never true.

People think software skills are linear; that someone half as good takes twice as long. The reality is that there's an ability threshold below which delivery simply never happens, no matter how long you give it or how many people and how much money you throw at it.

Also, no-one ever invaded a country that has nuclear weapons.

The PM -> engineer -> QA handoff was always an anti-pattern. Shame it's taken AI to wake people up to that organisational religion. One size fits all is never true. With methodologies, each... The PM -> engineer -> QA handoff was always an anti-pattern. Shame it's taken AI to wake people up to that organisational religion. One size fits all is never true. With methodologies, each one has i...

The PM -> engineer -> QA handoff was always an anti-pattern. Shame it's taken AI to wake people up to that organisational religion. One size fits all is never true:

"I can't program but I just got Claude to build an app/system/website for me" has Dunning Kruger written all over it.

Welcome to the long hard road of difficult lessons that leads ultimately to software engineering enlightenment.

When the CEO of Microsoft AI goes on social media to state they haven't hit an AI ceiling, my immediate conclusion is that they've hit an AI ceiling. Especially when the claim is riddled with confusing facts about different domains.

If there's one thing I've learned in my 60 years on this planet it's this. If a super rich CEO of a major corporation makes a statement, you should absolutely believe it. They are the most intelligent, informed and trustworthy individuals you're ever likely to meet.

Or maybe it's just that phones provide access to information that makes bliss through ignorance harder to hold on to than it was for previous generations?

TBF it was a pretty short trip. It's the long stays on the ISS that really showed up the issues.

And as a blog post:

I love analogies, and I also love cats. So here's a cat based analogy that attempts to explain why the apparent intelligence some people are seeing in LLMs is just an illusion. Cats like playing…... I love analogies, and I also love cats. So here's a cat based analogy that attempts to explain why the apparent intelligence some people are seeing in LLMs is just an illusion. Cats like playing with...

I love analogies, and I also love cats. So here's a cat based analogy that attempts to explain why the apparent intelligence some people are seeing in LLMs is just an illusion.

LinkedIn post or gratuitous excuse to post a picture of one of my cats? You decide 😉

Makes absolute sense tbh

And as a blog post:

I predicted Meta would fail in its VR venture, and Apple would struggle to sell its headset. I said Amazon would get nowhere quickly with its drone delivery plans if it gave the task to its logistics…... I predicted Meta would fail in its VR venture, and Apple would struggle to sell its headset. I said Amazon would get nowhere quickly with its drone delivery plans if it gave the task to its logistics ...

I predicted Meta would fail in its VR venture, & Apple would struggle to sell its headset. I said Amazon would get nowhere with its drone delivery plans if it gave the task to its logistics division, & healthcare would not go well for Amazon, Apple or Google. I knew property was not for John Lewis:

Eran is right that rewriting a system is just going to repeat previous mistakes if you don't do things differently the second time. Also, he's right that it may well be admitting defeat. Howev... Eran is right that rewriting a system is just going to repeat previous mistakes if you don't do things differently the second time. Also, he's right that it may well be admitting defeat. However, sunk...

Eran is right that rewriting a system is just going to repeat previous mistakes if you don't do things differently the second time. Also, he's right that it may well be admitting defeat. However, sunk cost fallacy is a thing. There are times when admitting failure is the right option:

LLMs are not conscious. I've been working with artificial neural networks since 1985, and in doing that I've consumes a lot of material about biological neural networks that do exhibit conscio... LLMs are not conscious. I've been working with artificial neural networks since 1985, and in doing that I've consumes a lot of material about biological neural networks that do exhibit consciousness. ...

LLMs aren't conscious. I've worked on neural networks since 1985, and in doing that I've consumed a lot of material about biological neural networks that do exhibit consciousness. And no, this isn't about human exceptionalism so don't throw that fatuous argument at me. It's about basic architecture.

Stephan is spot on with this assessment, and this summary clearly explains why software development is so bad in organisations that think writing software is just a numbers game. Writing software is…... Stephan is spot on with this assessment, and this summary clearly explains why software development is so bad in organisations that think writing software is just a numbers game. Writing software is ...

Stephan is spot on with this assessment, and his summary clearly explains why software development is so bad in organisations that think writing software is just a numbers game.

Writing software is not a mass production problem.

Technology and the barriers to virtual worlds - Jon Ayre The Metaverse is presented as the next big thing to impact our daily lives, but before that can happen there are some big hurdles to overcome.

Surprisingly, amidst all the AI noise, a few people are also noticing that Meta is winding up it's VR venture (despite renaming itself after said venture). Here's something I wrote back when they embarked on this venture that explains why it was always going to fail:

Money, money, money, or burn baby burn? - Jon Ayre The AI cost issue is already rearing its head. People are using workarounds, and AI companies aren't even charging a fraction of what it costs to run the models. What happens when they run out of inve...

And as a blog post:

The AI cost issue is already rearing its head. People are using workarounds, and AI companies aren't even charging a fraction of what it costs to run the models. What happens when they run out of…... The AI cost issue is already rearing its head. People are using workarounds, and AI companies aren't even charging a fraction of what it costs to run the models. What happens when they run out of inve...

The AI cost issue is already rearing its head. People are using workarounds, and AI companies aren't even charging a fraction of what it costs to run the models. What happens when they run out of investor money and need to make a profit?

It’s English Jim, but not as we know it - Jon Ayre If you think talking in plain English to an LLM (large language model) will produce a coded solution that does what you want, does it well and will stand the test of time, you're wrong. This post atte...

And as a blog post:

If you think talking in plain English to an LLM will produce a coded solution that does what you want, does it well and will stand the test of time, you're wrong. When anyone first tries to learn... If you think talking in plain English to an LLM will produce a coded solution that does what you want, does it well and will stand the test of time, you're wrong. When anyone first tries to learn to ...

If you think talking in plain English to an LLM will produce a coded solution that does what you want, does it well and will stand the test of time, you're wrong. Here's why:

And yes, that's on a repo for a live system. All changes pushed to prod. 365 days and counting, 100% uptime.

Who says bugs and outages are inevitable?

I don't have any particularly wild claims about creating 50 apps before breakfast with my hoard of AI agents, but I do seem to be doing okay by coordinating the fingers on my left and right hands.

Coding is not my day job anymore, so this GitHub chart is just my spare time.

Endoplanets, surely?

You should never open... *single shot rings out*

There was a bit of toilet trouble on NASA's Artemis 2 mission to the moon It wasn't their No. 1 problem, but it was a problem for going number one.

I have a phrase about good engineers and bad engineers that goes like this:

"There are engineers who can put a man on the moon and there are engineers who can't design a toilet that flushes first time."

I've been using it for decades.

Today, Artemis II has an issue with its loo.

The software engineer is dead, long live the software engineer! - Jon Ayre The IT industry is only ever as good as the software it produces. There has always been great software, and there has also always been a lot of bad software.

Also available as a blog post for those who don't want to visit LinkedIn:

The IT industry is only ever as good as the software it produces. There has always been great software, and there has also always been a lot of bad software. The people writing the great software… | ... The IT industry is only ever as good as the software it produces. There has always been great software, and there has also always been a lot of bad software. The people writing the great software did...

The software engineer is dead, long live the software engineer!

The IT industry is only ever as good as the software it produces. There has always been great software, and there has also always been a lot of bad software.