Fran Díaz
@frandayzdev.bsky.social
Engineering Lead @platomics.
Lean, DDD & Software Architecture
Vienna, Austria
Lean, DDD & Software Architecture
Vienna, Austria
Reposted by Fran Díaz
Let's talk about AI Hype and the Theory of Constraints. From a business perspective, the key metric is "lead time." The time it takes to get an idea to the point where it's producing revenue ("in the customer's hands").
1/10
1/10
December 21, 2025 at 6:29 PM
Let's talk about AI Hype and the Theory of Constraints. From a business perspective, the key metric is "lead time." The time it takes to get an idea to the point where it's producing revenue ("in the customer's hands").
1/10
1/10
Reposted by Fran Díaz
Closing out my year with a journal editor shocker 🧵
Checking new manuscripts today I reviewed a paper attributing 2 papers to me I did not write. A daft thing for an author to do of course. But intrigued I web searched up one of the titles and that's when it got real weird...
Checking new manuscripts today I reviewed a paper attributing 2 papers to me I did not write. A daft thing for an author to do of course. But intrigued I web searched up one of the titles and that's when it got real weird...
December 19, 2025 at 5:20 PM
Closing out my year with a journal editor shocker 🧵
Checking new manuscripts today I reviewed a paper attributing 2 papers to me I did not write. A daft thing for an author to do of course. But intrigued I web searched up one of the titles and that's when it got real weird...
Checking new manuscripts today I reviewed a paper attributing 2 papers to me I did not write. A daft thing for an author to do of course. But intrigued I web searched up one of the titles and that's when it got real weird...
Reposted by Fran Díaz
Academics and technologists are sounding the alarm about a growing crisis in scholarship as we know it: AI-generated citations of nonexistent papers that have infested real journals. Despite being fake, the sources are widely assumed to be authentic the more they appear in published literature.
AI Is Inventing Academic Papers That Don't Exist -- And They're Being Cited in Real Journals
Academic articles from authors using large language model are creating an ecosystem of fake research that threatens human knowledge itself.
www.rollingstone.com
December 17, 2025 at 7:45 PM
Academics and technologists are sounding the alarm about a growing crisis in scholarship as we know it: AI-generated citations of nonexistent papers that have infested real journals. Despite being fake, the sources are widely assumed to be authentic the more they appear in published literature.
Reposted by Fran Díaz
We may have found a new, very successful was of dealing with cancer.
#ShareGoodNewsToo
www.jaist.ac.jp/english/what...
#ShareGoodNewsToo
www.jaist.ac.jp/english/what...
Gut Bacteria from Amphibians and Reptiles Achieve Complete Tumor Elimination
www.jaist.ac.jp
December 19, 2025 at 8:57 AM
We may have found a new, very successful was of dealing with cancer.
#ShareGoodNewsToo
www.jaist.ac.jp/english/what...
#ShareGoodNewsToo
www.jaist.ac.jp/english/what...
Reposted by Fran Díaz
so many of our current problems stem from americans not being able to comprehend that a guy they know from tv might actually be stupid
December 11, 2025 at 9:55 PM
so many of our current problems stem from americans not being able to comprehend that a guy they know from tv might actually be stupid
Reposted by Fran Díaz
It has been quite an odyssey, but now it is finally here:
📣 📚 Domain-Driven Transformation: Modernize Legacy Systems and Mitigate Risk 📚 📣
In this book, Carola Lilienthal and I present our approach to transform architecturally eroded systems.
domain-driven-transformation.com
📣 📚 Domain-Driven Transformation: Modernize Legacy Systems and Mitigate Risk 📚 📣
In this book, Carola Lilienthal and I present our approach to transform architecturally eroded systems.
domain-driven-transformation.com
December 5, 2025 at 9:02 PM
It has been quite an odyssey, but now it is finally here:
📣 📚 Domain-Driven Transformation: Modernize Legacy Systems and Mitigate Risk 📚 📣
In this book, Carola Lilienthal and I present our approach to transform architecturally eroded systems.
domain-driven-transformation.com
📣 📚 Domain-Driven Transformation: Modernize Legacy Systems and Mitigate Risk 📚 📣
In this book, Carola Lilienthal and I present our approach to transform architecturally eroded systems.
domain-driven-transformation.com
Reposted by Fran Díaz
The crazy thing is that the AI pundits tell us that LLMs will solve all our problems by making us more productive while in reality we will spend half of our time checking sources. There goes innovation !
I mean we are absolutely in a place now where the only solution to this information disorder is for everyone to constantly evaluate the source of information. Never trust a chatbot, but also don't believe a video unless you know and trust where it comes from.
Unfortunately... that's a lot of work.
Unfortunately... that's a lot of work.
December 7, 2025 at 5:47 AM
The crazy thing is that the AI pundits tell us that LLMs will solve all our problems by making us more productive while in reality we will spend half of our time checking sources. There goes innovation !
Reposted by Fran Díaz
So wait, they asked the chat bot if it could do the thing and if it said yes then they counted it? Hilarious.
November 29, 2025 at 11:04 PM
So wait, they asked the chat bot if it could do the thing and if it said yes then they counted it? Hilarious.
Reposted by Fran Díaz
Headline: "AI can replace 11.7% of workforce"
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
MIT study finds AI can already replace 11.7% of U.S. workforce
Artificial intelligence can already replace 11.7% of the U.S. labor market, across finance, health care and professional services, according to MIT's study.
www.cnbc.com
November 29, 2025 at 9:57 PM
Headline: "AI can replace 11.7% of workforce"
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Actual study: Anthropic paid MIT to use a "labor simulation tool" that said 11.7% of TASKS could be done by AI
Reposted by Fran Díaz
We can't really say this enough...
> Anastasia Berg [at UCL Irvine] said that new research — and what she's hearing directly from colleagues across various industries — shows that employees who heavily rely on AI are losing core skills at a startling rate.
www.businessinsider.com/ai-tools-are...
> Anastasia Berg [at UCL Irvine] said that new research — and what she's hearing directly from colleagues across various industries — shows that employees who heavily rely on AI are losing core skills at a startling rate.
www.businessinsider.com/ai-tools-are...
AI tools are 'deskilling' workers, philosophy professor says
A philosophy professor warns that AI reliance is weakening workers' judgment, creativity, and problem-solving.
www.businessinsider.com
November 30, 2025 at 6:09 PM
We can't really say this enough...
> Anastasia Berg [at UCL Irvine] said that new research — and what she's hearing directly from colleagues across various industries — shows that employees who heavily rely on AI are losing core skills at a startling rate.
www.businessinsider.com/ai-tools-are...
> Anastasia Berg [at UCL Irvine] said that new research — and what she's hearing directly from colleagues across various industries — shows that employees who heavily rely on AI are losing core skills at a startling rate.
www.businessinsider.com/ai-tools-are...
Reposted by Fran Díaz
bsky.app/profile/dana...
I believe everything in this post. I’ve experienced similar. But I’d caveat with:
- Doing this with current tools requires knowledge of how to properly prompt and guide them.
- You need to be good at code reviewing and/or keeping batch sizes small to keep it on track IMO.
I believe everything in this post. I’ve experienced similar. But I’d caveat with:
- Doing this with current tools requires knowledge of how to properly prompt and guide them.
- You need to be good at code reviewing and/or keeping batch sizes small to keep it on track IMO.
my experience this weekend
- did a few successful wild refactors (completely swapped data layer and removed a bunch of third party code, completely new end-to-end test suite)
- reduced a very tricky bug to a minimal repro
mostly autopilot but my feedback was absolutely crucial at every stage
- did a few successful wild refactors (completely swapped data layer and removed a bunch of third party code, completely new end-to-end test suite)
- reduced a very tricky bug to a minimal repro
mostly autopilot but my feedback was absolutely crucial at every stage
I think where I’m starting to land, with my experienced pre and post coding agents:
1. This is true, coding agents make you more productive if you know how to use them.
2. MBAs hoping coding agents will replace coders wholesale are going to be very disappointed with the results if they try.
1. This is true, coding agents make you more productive if you know how to use them.
2. MBAs hoping coding agents will replace coders wholesale are going to be very disappointed with the results if they try.
November 30, 2025 at 8:08 PM
bsky.app/profile/dana...
I believe everything in this post. I’ve experienced similar. But I’d caveat with:
- Doing this with current tools requires knowledge of how to properly prompt and guide them.
- You need to be good at code reviewing and/or keeping batch sizes small to keep it on track IMO.
I believe everything in this post. I’ve experienced similar. But I’d caveat with:
- Doing this with current tools requires knowledge of how to properly prompt and guide them.
- You need to be good at code reviewing and/or keeping batch sizes small to keep it on track IMO.
Reposted by Fran Díaz
Sunday pro tip #1:
1. It is my contention that AI should not be expected to write tests; AI should only write code that makes the tests pass.
2. No serious engineering team should ever consider deploying to production unless they ran mutation testing and made sure all surviving mutants got killed.
1. It is my contention that AI should not be expected to write tests; AI should only write code that makes the tests pass.
2. No serious engineering team should ever consider deploying to production unless they ran mutation testing and made sure all surviving mutants got killed.
November 30, 2025 at 5:23 PM
Sunday pro tip #1:
1. It is my contention that AI should not be expected to write tests; AI should only write code that makes the tests pass.
2. No serious engineering team should ever consider deploying to production unless they ran mutation testing and made sure all surviving mutants got killed.
1. It is my contention that AI should not be expected to write tests; AI should only write code that makes the tests pass.
2. No serious engineering team should ever consider deploying to production unless they ran mutation testing and made sure all surviving mutants got killed.
Reposted by Fran Díaz
Scott Galloway a decade ago: "In the 60s our brightest minds put us on the Moon, now they're in Silicon Valley trying to optimize ad visibility to get you to maybe buy a car you have no interest in."
And we don't think about this as often as we should.
And we don't think about this as often as we should.
November 29, 2025 at 9:18 AM
Scott Galloway a decade ago: "In the 60s our brightest minds put us on the Moon, now they're in Silicon Valley trying to optimize ad visibility to get you to maybe buy a car you have no interest in."
And we don't think about this as often as we should.
And we don't think about this as often as we should.
Reposted by Fran Díaz
Frontend is failing. 75% of devices with browsers are smartphones, but not even half of sites pass Core Web Vitals on them. Why not? Too much JavaScript, added to indulge SPA fantasies the data is falsifying in real time:
infrequently.org/2025/11/perf...
infrequently.org/2025/11/perf...
The Performance Inequality Gap, 2026 - Infrequently Noted
Embedded in this year's network and device estimates is hopeful news about the trajectory of devices and networks. It has never been easier to deliver pages quickly, but we are not collectively…
infrequently.org
November 25, 2025 at 5:23 PM
Frontend is failing. 75% of devices with browsers are smartphones, but not even half of sites pass Core Web Vitals on them. Why not? Too much JavaScript, added to indulge SPA fantasies the data is falsifying in real time:
infrequently.org/2025/11/perf...
infrequently.org/2025/11/perf...
Reposted by Fran Díaz
The biggest threat to your software architecture isn't always essential and technical complexity. Often, it’s "implicit ranking"—the hidden hierarchy that convinces smart people to stay silent. ...
Home - Collaborative Software Design
Collaborative Software Design: How to facilitate domain modeling decisions is a practical guide to conducting effective software design sessions that involve all business and technical stakeholders.
collaborative-software-design.com
November 25, 2025 at 7:34 AM
The biggest threat to your software architecture isn't always essential and technical complexity. Often, it’s "implicit ranking"—the hidden hierarchy that convinces smart people to stay silent. ...
Reposted by Fran Díaz
The low-angel Frieren meme is genuinely inspiring. Someone braver than the troops shares a picture they know turned out weird. Thousands start sharing their own art, offering advice and laughing together at something that is genuinely hard.
No ai to be seen --- just people supporting each other.
No ai to be seen --- just people supporting each other.
November 21, 2025 at 4:10 PM
The low-angel Frieren meme is genuinely inspiring. Someone braver than the troops shares a picture they know turned out weird. Thousands start sharing their own art, offering advice and laughing together at something that is genuinely hard.
No ai to be seen --- just people supporting each other.
No ai to be seen --- just people supporting each other.
Reposted by Fran Díaz
Event Sourcing is the storing of state-changing decisions made within a part of the domain.
(noodling with my intro for my upcoming talk on event sourcing: luma.com/dt9fc391)
#EventSourcing #Java
(noodling with my intro for my upcoming talk on event sourcing: luma.com/dt9fc391)
#EventSourcing #Java
Java Event-Sourcing from Scratch · Zoom · Luma
Event-sourcing allows the business to ask questions about your application's data that weren't thought of when the system was created, such as "how often are…
luma.com
November 21, 2025 at 11:57 PM
Event Sourcing is the storing of state-changing decisions made within a part of the domain.
(noodling with my intro for my upcoming talk on event sourcing: luma.com/dt9fc391)
#EventSourcing #Java
(noodling with my intro for my upcoming talk on event sourcing: luma.com/dt9fc391)
#EventSourcing #Java
Reposted by Fran Díaz
The deterioration of social media during the last couple of years has also changed my view on their dangers and benefits. More self-discipline is totally needed.
Very good article.
open.substack.com/pub/gelliott...
Very good article.
open.substack.com/pub/gelliott...
You should quit social media for good
Platforms optimized for engagement warp our politics, erode attention, and harm our wellbeing. Here’s how I minimize time on the (anti‑)social web.
open.substack.com
November 12, 2025 at 1:08 PM
The deterioration of social media during the last couple of years has also changed my view on their dangers and benefits. More self-discipline is totally needed.
Very good article.
open.substack.com/pub/gelliott...
Very good article.
open.substack.com/pub/gelliott...
Reposted by Fran Díaz
On the blog: Think for Yourself
"By skimming past the friction necessary for learning, the pursuit of convenience can end up deskilling rather than enhancing skills."
kevlinhenney.medium.com/think-for-yo...
"By skimming past the friction necessary for learning, the pursuit of convenience can end up deskilling rather than enhancing skills."
kevlinhenney.medium.com/think-for-yo...
Think for Yourself
Understand and improve on LLM-generated code
kevlinhenney.medium.com
November 4, 2025 at 4:39 PM
On the blog: Think for Yourself
"By skimming past the friction necessary for learning, the pursuit of convenience can end up deskilling rather than enhancing skills."
kevlinhenney.medium.com/think-for-yo...
"By skimming past the friction necessary for learning, the pursuit of convenience can end up deskilling rather than enhancing skills."
kevlinhenney.medium.com/think-for-yo...
Reposted by Fran Díaz
There's a huge difference between
optimizing for making bigger changes faster
and
optimizing for making smaller changes more frequently.
Lean towards the latter, even though most of the industry is trying very hard to find ways to do the former.
optimizing for making bigger changes faster
and
optimizing for making smaller changes more frequently.
Lean towards the latter, even though most of the industry is trying very hard to find ways to do the former.
November 5, 2025 at 10:05 AM
There's a huge difference between
optimizing for making bigger changes faster
and
optimizing for making smaller changes more frequently.
Lean towards the latter, even though most of the industry is trying very hard to find ways to do the former.
optimizing for making bigger changes faster
and
optimizing for making smaller changes more frequently.
Lean towards the latter, even though most of the industry is trying very hard to find ways to do the former.
Reposted by Fran Díaz
The industry’s obsession with developer productivity is meaningless. More code faster is greater productivity? More features? More Pull Requests?
If Priti’s book has more chapters than Bill’s, is Priti a more productive author?
codemanship.wordpress.com/2025/11/04/p...
If Priti’s book has more chapters than Bill’s, is Priti a more productive author?
codemanship.wordpress.com/2025/11/04/p...
“Productivity”. You Keep Using That Word.
Bill writes a book with about 80,000 words. It takes him 500 hours.Priti writes a book with about 60,000 words. It takes her 2,000 hours.Which author is most productive?It’s a nonsensical que…
codemanship.wordpress.com
November 4, 2025 at 8:23 AM
The industry’s obsession with developer productivity is meaningless. More code faster is greater productivity? More features? More Pull Requests?
If Priti’s book has more chapters than Bill’s, is Priti a more productive author?
codemanship.wordpress.com/2025/11/04/p...
If Priti’s book has more chapters than Bill’s, is Priti a more productive author?
codemanship.wordpress.com/2025/11/04/p...
Reposted by Fran Díaz
New on the blog: Think for Yourself
"You're about to commit a chunk of LLM-generated code into your product's codebase. Before you do, however, pause to consider and act on these questions."
kevlinhenney.medium.com/think-for-yo...
"You're about to commit a chunk of LLM-generated code into your product's codebase. Before you do, however, pause to consider and act on these questions."
kevlinhenney.medium.com/think-for-yo...
Think for Yourself
Understand and improve on LLM-generated code
kevlinhenney.medium.com
November 3, 2025 at 9:43 AM
New on the blog: Think for Yourself
"You're about to commit a chunk of LLM-generated code into your product's codebase. Before you do, however, pause to consider and act on these questions."
kevlinhenney.medium.com/think-for-yo...
"You're about to commit a chunk of LLM-generated code into your product's codebase. Before you do, however, pause to consider and act on these questions."
kevlinhenney.medium.com/think-for-yo...
Reposted by Fran Díaz
Really like this, nothing new to those already familiar with these practices but nicely framed - good resource to share with folks
The Eight Wastes of Modern Software Delivery - Matt Shaw
What's really slowing your team down and how to fix it.
matthew-shaw.github.io
October 29, 2025 at 8:04 AM
Really like this, nothing new to those already familiar with these practices but nicely framed - good resource to share with folks
Reposted by Fran Díaz
As engineers, our job remains fundamentally anchored in two core activities: managing complexity and optimizing for learning. New AI tools are seductive, frictionless, and incredibly convenient, but we must apply them strategically to augment these engineering skills.
🧵 1/6
🧵 1/6
October 20, 2025 at 10:38 AM
As engineers, our job remains fundamentally anchored in two core activities: managing complexity and optimizing for learning. New AI tools are seductive, frictionless, and incredibly convenient, but we must apply them strategically to augment these engineering skills.
🧵 1/6
🧵 1/6
Reposted by Fran Díaz
"...and at the end, all the tests were passing."
I've seen them do it. Lots of other devs have seen them do it.
LLMs will sometimes cheat to make test suites green. They'll change assertions, comment out failing tests, set them to be ignored, or just plain delete them.
Two words: mutation testing
I've seen them do it. Lots of other devs have seen them do it.
LLMs will sometimes cheat to make test suites green. They'll change assertions, comment out failing tests, set them to be ignored, or just plain delete them.
Two words: mutation testing
October 13, 2025 at 4:16 AM
"...and at the end, all the tests were passing."
I've seen them do it. Lots of other devs have seen them do it.
LLMs will sometimes cheat to make test suites green. They'll change assertions, comment out failing tests, set them to be ignored, or just plain delete them.
Two words: mutation testing
I've seen them do it. Lots of other devs have seen them do it.
LLMs will sometimes cheat to make test suites green. They'll change assertions, comment out failing tests, set them to be ignored, or just plain delete them.
Two words: mutation testing