A report published by Thought Works in February 2025 looked at the impact of coding assistants on development speed. They found out that the delivery speed improvement is between 2% and 15%. But the interesting things are qualitative: how did they measure? How can you measure by yourself? And where were the assistants most useful? […]
Do AI Assistants Speed Up Development? Read More »
A common question our clients ask is: our estimations are not precise enough, what should we do? Alex responds in this video.
Dealing With Estimation Variability Read More »
In this video, Alex asks an interesting question: what if we invented programming now, after we have LLMs and AI agents? Would we program in the same way? Would we still need programming languages? Would testing become more important than programming? And where is software development headed next?
What If We Invented Programming After LLMs? Read More »
Alex is reporting on his experience using ChatGPT, Claude, DeepSeek-R1 and Github Copilot for various programming tasks and compares it with Elisabeth Hendrickson’s experience report. The conclusion: these tools have an amplifying effect for existing skills, rooted in both the ease of access to knowledge (even imperfect) and the psychological effect of knowing that you
Programming with AI – Experience Report Read More »
Prompt injection is a vulnerability present in LLM systems and in software products that take advantage of LLM features. It’s something that every developer should be aware of and consider when implementing LLM-based systems. Links:
A Brief Introduction to Prompt Injection Read More »
Is it wrong to be wrong? Yes, if it’s about the result. No, if it’s during the process. The freedom to be wrong is fundamental for creativity, so allow yourself and your colleagues the space for it.
The Freedom To Be Wrong Read More »
Cooper that looks at AI development tools from the perspective of Fred Brooks “There is no silver bullet”. Links:
Is AI A Silver Bullet? Read More »
This page contains resources for the keynote held by Alex Bolboaca and Adrian Bolboaca in Brescia, for the “Working Software” Conference 2024. Software Development is Non-Linear We discussed briefly in the section “II. When Requirements Change” about the non-linear reality of software development. A deep dive into this topic can be found in our video
Seven Attitudes For a Better Programming Life Read More »
tl; dr: it’s not hard to learn TDD, it’s hard to unlearn the linear way of software development we’ve been used to In a recent discussion on Mastodon, Emily Bache asked the question: “In what ways is TDD easier than test after? And why is it hard to learn?”. This is the answer to the
Is TDD Hard To Learn? Read More »
Extreme Programming (XP) practitioners often talk about how they haven’t used a debugger in years. Yet this assertion often leads to potential misunderstandings. Should you really avoid debuggers? Or is there something more to unpack? tl; dr: good tests reduce the need of debugging to almost zero. A debugging session is an automated test that
Should You Avoid Debuggers? Read More »
