HUSTEF was on my list of conferences to go to for quite a long time. Lots of people raved about it and I was curious to experience it for myself. This year, I finally made it there!
It all started with getting invited to the review board, helping out with evaluating proposals so that the program committee could make an informed decision. So I thought, why not submitting sessions myself and see if I'd also get a chance to speak? And it worked! I consider myself fortunate to have been selected among the many papers submitted.
Time flew by this year, and now it was time to travel to the beautiful city of Budapest. A long but comfortable train ride, some time to settle down and rest at the hotel, then I met Micha Kutz for dinner. Always nice to catch up with long-time friends from the community!
Tutorial Day
The first day was reserved for full-day and half-day tutorials. I gave my workshop "Secure Development Lifecycle Applied – How to Make Things a Bit More Secure than Yesterday Every Day" as a half-day tutorial at one of the sponsoring companies who kindly provided their office space for it.
I've given this workshop plenty of times already and it's always super interesting to see different kinds of groups engage with the material in their own way, based on their own experience. I'm also very happy to see that the concept does work out with most folks, coming from very different backgrounds! This time I had mostly people working in testing and quality roles as they are the main audience for HUSTEF. I really enjoyed seeing them work through the exercises, ask questions, share their own related stories, and generally having fun doing so.
And then a truly special thing happened. The tutorial was over - yet participants didn't leave. They stuck around. They waited for each other having a bio break, they kept talking. They all left the location together, heading back to the city center, planning to have lunch together - and invited me in! That was too good an opportunity to decline, no matter how tired I was or what else I needed to tackle in the afternoon. We had a lovely lunch together, sharing even more experiences - what could be a better start for the conference?
At some point, I needed to head back to the hotel and do final preparations for the next day. Then it was time for speakers dinner. I met Sonja Nešić on the way to the restaurant, a lovely chance to catch up with each other. The dinner place was literally on a boat, yet I got lucky that it was firmly anchored and did not move in any perceivable way. I already had the experience of having speakers dinner on a boat in the past and I can only tell you it's not a great idea if you have motion sickness. The place's acoustics unfortunately couldn't handle the noise well and I struggled a lot with processing my environment, barely hearing what people said and straining my voice to make myself heard. On the bright side, the food was excellent and I had really great company. For example, I just loved catching up with Clare Norman and Jitesh Gosai, talking about light and deep topics, and finding out that all three of us are scripting our talks. I'm in good company!
Conference Day 1
The first conferencing day arrived and it was packed with amazing talks - it was a true challenge to pick which sessions to witness live. Also, I had to skip two slots in order to be ready for my own talk and then catch my breath right after. It was a pity to miss Nicole van der Hoeven's keynote "Asimov’s Zeroth Law of Robotics: Observability for AI" this way, I was really looking forward to it. Also, I would have loved to see Micha Kutz's talk "WTFM: Where’s the F****** Manual⁈ – Why we avoid writing documentation, why we shouldn’t, and how to keep it useful with minimal pain". Good thing the talks were recorded and I can catch those I missed later on - even though it's never the same as experiencing them live on-site.
-
Keynote "Beyond Vibe Coding: What It Actually Takes to Build with AI and Stay
Accountable" by
Kristoffer Nordström. Well, AI tooling usage can found around many corners these days. Yet what does it really take to get good value from it, how to build software with these tools and still know exactly what you're doing and especially why? Kristoffer shared his story and what he found made the difference. I really liked the following of his takes: When building becomes cheap, we need judgment in the room.
-
"The art of getting out of the way: Facilitation for quality
professionals" by
Clare Norman. Loved that Clare shared her insights on such an underrepresented and yet
so crucial topic. Lots of invaluable advice, brilliant presentation! I especially loved her emphasis that facilitation is all about listening and supporting everyone in the group in doing their best thinking together - and that includes generosity and fairness.
-
"The Rubber Duck that talks back" by
Benjamin Bischoff. Benjamin pointed out that we had hype cycles before AI tooling, like search. And all the shortcomings that come with new technology, hyped or not. He also shone a light on how we can use our skepticism for good. Benjamin encouraged us to make use of our brains, apply discipline and differentiate unnecessary from necessary friction.
-
"AI applications security puzzle" by
Maryia Tuleika. Maryia told a story on how security concerns evolved over the years, while some are still remaining alike. Especially for current AI tooling, how do we secure them? She points out that we should rather focus on the system and constrain the tools we have at that level. I really liked her take: "Don't ask only whether you can trust AI - build and test for the moment you can't." Maryia based her talk on detailed research in the space, gathering insights from the community - worth a read, and I'm honored to be among the people she consulted.
-
"Out of Security Exception – What to Do Without an Expert to Secure Your
Software" by me. I presented what engineering teams can do themselves to improve their product's security posture without having a security expert at hand. I've given this talk already at BSides Luxembourg this year, yet in a longer and more detailed form. This time I needed to cut it down to the essentials to fit the time slot. It seems the message came across nonetheless!
-
"How quality is created, maintained and lost in complex software
systems" by
Jitesh Gosai. I've seen Jit present many years ago and was already back then impressed by both his presentation skills and by how much his point of view overlaps with mine. His talks and thoughts just resonate a lot with me, so I had to check out this one as well. He walked us through the famous CrowdStrike outage that made the world stop turning for a bit, and what we can learn from it. It was a great reminder that complex failures never have a single root cause and they require us to acknowledge and improve the socio-technical system.
-
Keynote "Could Personas Be Your Insight Superpower?" by
Steve Watson. Are personas still relevant or needed in the age of fancy AI tooling? Steve says yes, and I agree. He presented how we can use them in a meaningful way and demonstrated the pitfalls when creating actually insightful personas without missing out on what's relevant for our target user group. As he shared, the risks are in the gap between our assumptions and our users' reality.
I truly appreciate the lunch breaks were decently long to get some rest and also socialize. I had the chance to briefly speak with Vitaly Sharovatov - always a pleasure, ever since we met at NewCrafts Paris 2024. Same with Sérgio Freire and Mazin Inaad! Would have loved to catch up more with each of them.
What I thoroughly enjoyed was the opportunity to have a quiet and relaxed dinner with Heather Reid, Léna Nyström and Rachel Kibler. We had discovered the perfect place for it as well, with delicious food, calm atmosphere and awesome service. For me it was an utterly needed chance to take a break for a little while before the pressure will be on again the next weeks.
Conference Day 2
The final conference day came with yet another bunch of awesome talks.
- Lean Coffee with Gáspár Nagy. I thoroughly enjoy lean coffee sessions, and yet I did not plan to join this one. That being said, I had arrived early enough at the venue, so I decided to just join in. We had insightful conversations on the usage of AI for slides, how spending money on tokens that goes to foreign companies destroys our local economies, how everyone turning into a vibe coder and wanting to bring their projects into production causes a huge problem and how to navigate it.
-
Keynote "Now You See Me" by
Heather Reid. What can I say. Awesome content, awesome delivery. More people need to hear
this. Heather talked about what makes upper leadership not see what people achieved, while they themselves feel it's clearly there to see. Perspectives matter and we all only have our own viewpoint available. Making things visible in ways that allow others to see things is crucial for any career - no matter if you have a leadership title or not.
-
"'As is' is not a requirement – it’s failure in a fancy suit" by Léna Nyström. I loved how Léna walked us through three cases in her career where it seemed people wanted a copy of what's been built before. Yet is it ever a true copy? Of course not. It needs to be modern. Some things aren't relevant anymore. No one has the complete picture. Did anyone even ask what's actually needed or what wasn't working well with the old solution? We should listen to this signal of "It should work as is" and find out what questions are unasked.
-
"Looks Perfect. Runs Fine. Completely Wrong. A Live Hunt." by
Christine Pinto. Christine took us on the investigation of an incident where all things came together so that a bunch of AI tools enabled attackers to exfiltrate data and demand ransom from a company. Based on the audience response it seems the message landed - supply chain attacks using AI tooling are way too common these days to stay inactive.
-
"Quality Signals on Pull Requests: How We Stopped Shipping Bugs Without
Slowing Delivery" by
Marco Andrade. Marco shared an all too well-known story of flaky end-to-end tests. He pointed out what the underlying issues were and what steps they took to make sure they got actual feedback and hence value from these tests. Marco warned people not to become the gatekeeper and scapegoat but have teams trust their pipelines.
-
Keynote "You’re not really ready for AI (but testing can help)" by
Callum Akehurst-Ryan. I really appreciate that Callum shared his personal story and showed vulnerability on stage - we need to see more of such good examples. These are the cases we all can learn the most from. Callum shared how much AI amplified all the dysfunctions that already existed before in the teams and how that ended up badly for his organization and him personally. He emphasized that we can't just demand attention from people who don't have capacity to listen. We have to roll up our sleeves, make things better ourselves, and hence become part of the solution.
-
"The Tale of Two Pipelines: Why Your CI/CD Can’t Fix Cultural Debt" by
Chris Armstrong. Technical debt oftentimes isn't technical at all. It's rooted in a lot of other types of debt, deeply human ones. Chris presented how it doesn't matter if your engineered delivery pipeline is fast if your team can't actually get things through due to other types of debt. He also showed how such debt can look like and what to do about it. The core: optimize for flow.
-
"When Your AI Testing Partner Lies to Your Face: Detecting and Defeating
Completion Theater" by
Dragan Spiridonov. Dragan addressed how AI tooling can claim to offer confidence, yet there's either no evidence for it or the evidence indicates a different situation. He shared how you can spot this and what to do instead not to let ourselves be swayed by AI. Now, Dragan didn't only give this talk, but also a full-day tutorial and a masterclass. People clearly appreciated what he shared, in the end he won the conference's best presentation award! Check out his experience report on the conference as well.
-
"The Quality Nervous System" by
Kevin Pyles. Kevin presented a different approach to solve the overwhelm of too much data, getting flooded while still not knowing whether you can release a new software increment. He compared it to the human nervous system which also has to deal with huge amounts of data points and signals and transform them into meaningful output at scale. He also demonstrated how quality as a living capability could look like.
-
Keynote "Rethinking the Automation in Testing Principles" by
Richard Bradshaw. Richard looked back 10 years ago when he and Mark Winteringham came up with a set of principles to achieve more effective testing through the help of automation. He contrasted them with today's world, presented revisions made and why (check out the reasoning in Richard's own words). The principles in their new form are staying relevant and keep serving as an invaluable tool to use our thinking skills and help make better software - focusing on intent and judgment. "If you can't answer why something exists or should exist, change it." Hear, hear!
Conferences have a way to start slowly, make you forget how time flies by, and suddenly they're over. That's what happened here as well. Good thing we could prolong the community spirit for a while on the last evening. I joined a great dinner initiated by Kristoffer Nordström, with Cassandra H. Leung, Chris Armstrong, Marta Firlej, Mateusz Holewski and Richard Bradshaw. Later on, we were joined by Alex Cusmaru, Jamina Pöllänen and Gek Yeo. We moved on to a bar and stayed until it closed - it was finally time to end the evening. What can I say. Amazing food, amazing company. The best part of every conference. Thanks everyone for making this such a perfect end for my first HUSTEF!
Well, it wasn't quite over for me yet. I still had one day in Budapest. My chance to sleep in, go for a walk, do a bit of sightseeing. Truly ending the conference with another great dinner and conversation together with Jamina Pöllänen. Time to go home, time to share my experiences. Thank you everyone who made this conference a pleasure to be at!