Maaret Pyhäjärvi

Having a position on a higher career level than your capabilities are means you are more likely to not exceed expectations. Now that I have three organizations to think about around this, I realize that half of people promoted were promoted on potential, and the potential did not always realize. Others needed merit of the higher level before getting into it.

There would appear to be a gender difference as articles around world suggest where men are more likely to be promoted on potential.

I've been setting up collections of my notetaking streams (12 years of twitter, 4 years of mastodon, blog, posted articles, Linkedin) into a second brain kind of structure on Obsidian, and I am really liking what I get with that.

18 years of sharing all public notes in public is an impressive data source on my growth. And accessing it with a harness + LLM is very useful to me for personal productivity.

Reminded a guy today that when they look at their conference submissions next week, ooing and aaing about why there are no women, that the amount of extra work to exist in this tech space and world does not leave equal amounts of free time to submit proposals just in case, and inviting works a lot better.

Blogging since 2010. What about? to 2.04M views.

Explained this a week ago to someone who discussed public speaking with me, that some of the attention you get as speaker is not about you but the company you represent. You will learn how much when you change to an organization classified disliked-at-large, a consultancy.

Nothing of you changed. But everything can change. It's not bad, it's just how things are.

[Blog] Public notetaking, a 12 year sample
https://visible-quality.blogspot.com/2026/08/public-notetaking-12-year-sample.html

I interviewed a knowledge worker (sales) about their use of AI at work. One of the experiences I was tracking was if they have pitched a solution with AI.

I did not expect the response. They have, but every single one of them has turned into the client taking the idea and running with it.

The ideas have a different position now.

Refused to submit a proposal to a conference, insisting that I let them know that if they would want something I can contribute, I have now told them I am available. Asking for free work beyond that is not right, and is not for me at this time.

They submitted my two ideas from the conversation we had themselves.

Conferences should do their own work of sourcing great content and not rely on people who can afford to copy paste proposals in their submission systems.

[Blog] Breaking what you thought you knew about testing in 30 minutes
https://visible-quality.blogspot.com/2026/08/breaking-what-you-thought-you-knew.html

Ok I’m so proud of my team’s people. For 3 testers elsewhere rejected, mine was approved. One more data point that what I coach them to is what is wanted and hard to find. Teachable skills and all that and sad that so many testers were never taught.

Turned on 'quietly looking' as I can't really be loudly looking on LinkedIn as I can be here. I am thinking of trying out Central Europe and relocation now that my young adults allow for it. It will take me some work to only work 7.5 hours a day to make space for serious looking.

My vibe code ISTQB coach app helped 1st colleague pass through the Advanced Test Manager certificate. Useless certificate that does not cost the course fee is real savings.

I may hate that it’s required but I got all of them. Not because it’s good but because someone decided this brand is blocking access to work.

Someone made the usual ‘more work for same salary’ remark when I spoke of task expansion.

Every month gives us more salary. And the monthly sum does go up with perceived value.

I was asked to take 15 more direct reports, interim. Not for more salary but to optimize the flow of AI + testing upskillling I’m with them anyway.

Appreciating 'careful analysis' that good system specialist testers without ideas about improved tooling / automation bring to the table.

When I explore, I do 'careful analysis'. Since I move around different systems, I specialize in learning fast the system specialization, surprising people on how fast I can know things they don't through connecting existing information.

Careful analysis that takes years of system learning to build is valuable. Product historians are valuable.

Processing the falling kind of failing. No major troubles, but thinking about what tripped me in teaching AI for testing in a small group today.

I had 3 people. Everyone was at an entirely different place.

One was a developer turned tester, and could easily see how to apply GitHub Copilot + MCPs + CLIs + agents/skills, and where the client rules on test data being confidential would block it.

One was tester dipping toes to something technical.

One was tester who was clicking UI.

Turn anger to power for good. So today I made agreements necessary to bring in Agentic accessibility assessments and agentic fixing of those problems to pull requests while doing the assessment to a project, and got to doing it.

I’m starting to come to terms with the idea that I’m not cut out for consulting. I hate the continuous negotiation of someone paying for my time, and resulting guilt of never being enough. And people aren’t meant to work for this many organizations in a year.

The new problems "I used 3k€ on tokens for practicing for work-related hobby projects" and now someone wants to know who pays all the Facebook clones we generate with AI :D

The idea that subcontractor would want to raise prices because their work requires now more tokens, working on a no-AI-allowed case makes up a funny story of how the world works.

It may have taken me a few hours but I can blog (with images!) again. 🎉

‘Do you do anything else but work?’ Exhibit 1. Only one of the books I read this year are work-related and even that I read because a friend read it too and we met to discuss it.

I have felt guilty on how much I do with a lot of these well-intentioned questions, but I wonder if all of my high-performing colleagues get this on a weekly basis.

Scheduled a 1.5-hours session for 3 new testers with GitHub Copilot access, with the idea of leading them through ensemble working to realize some things that would take them weeks solo.

I did not foresee how many obstacles one can have to get to a working environment, from functional license to admin rights to vpn exceptions, or how much we could be thrown off by one form having become 3 since last looking at them.

If ‘3000 person change journey’ in a company with 37% women manages to fit one woman and that is ‘change management specialist’, I’m calling it badly done. I just wish it’s not as bad as pictures shared on LinkedIn show.

Three testers go to an interview with the same feedback: missing knowledge of TDD, test automation and AI. These aren’t negotiable any more.

12 years on twitter. What did I write about on #ExploratoryTesting.

1. What Exploratory Testing Is (core concepts — mindset, learning, agency)
2. ET vs. Scripted Testing (test cases as output, not input)
3. ET and Automation ("contemporary exploratory testing," the automationist's gambit)
4. History and Attribution (your research into Cem Kaner, Elfriede Dustin, the 1984/1988/1999 dates)
5. Defending the Term (pushback on deprecation and "ad hoc" framing)
6. Teaching, Courses, Mob/Ensemble Testing, and the Book
7. Community and Peer Conferences (ET Dinners, DEWT, ET19)
8. Managing Exploratory Testing (charters vs. sessions, agency)
9. Developers and Exploratory Unit Testing
10. Personal Journey and Career (2010 → 2022 arc)
11. Craft Notes (bugs, serendipity, real sessions)

It was 1304 tweets out of 33551 on this.

Today in 'learn testing by testing', I offered people three options for their target and frame. Interestingly, 2/3 of crowd chose second option, 1/3 the first option and none the third option where they could have placed themselves on a benchmark.

Explained this again today: there is such a thing as manual unit testing. It makes little sense and is not talked about much since modern unit testing frameworks are better promise, but it exists. It is often that first functional testing in small scale that developers do, and may not separate.

If you've been promised "unit testing" contractually, it does not automatically mean you get the repeated 80% code coverage after each change. Or have invested in having that.

‘Your guidance has been invaluable in driving improvements’ as client feedback of the day.

If invaluable would show up in business numbers as positive I could stop feeling like a failure. And sometimes it does but I have so much to learn.

I have a coaching session (getting coached) booked up for tomorrow morning, and trying to make sense of what do I want to be coached on.

The two previous sessions were on:
- choosing between testing and AI workforce transformation at large
- sense of invisibility, and what really matters to me and why

We decided (well, I, while coached) to take on something small next time. And now I feel like everything I do is too big to do alone, and almost also too big to do with the people I do things with.

Exploratory, in minds of a lot of people, is to testing what Vibe is to coding. Suggestion of low skill on whoever is doing it.

I know the first one is high skill, and I am still undecided on whether Vibe coding at its core is low-skill, and they more to AI native software engineering when it is high-skill.

Because the ask-and-receive vs. spec-world-first is not the real differentiator.

Even if they call these things "annual goals", getting stuck on the fact that understanding has grown and something specific that wasn't specific back then is now asked, well, feels unnecessary. Especially when that new thing is 5 minute thing, and the thing you skipped is a full certification.

I have used my 3k€ on learning how usage-based invoicing of cloud services work on datadog. It's been years since, but I get the credit of learning that the synthetic monitoring over a lot of locations could ramp up quite a bill.

So today I am wondering if everyone needs that 3k€ for their learning. I did not use mine on tokens, but consumables are all the same.

A boss at worked bought coaching and the coach got them to seek a weekly pair on AI learning that drives the org. The boss asked me to pair.

5 months later we have actions that were born in that pairing. The conversations balance outside and inside signals.

Serendipity of connections. Strengths of two.

I find it increasingly difficult to explain how I’m hiring testing specialist while having testing specialists without work. The difference is that people I hire get selected for projects. The learning gap is easily a year with people with similar years in industry.

And there is no school or course I can send them. It’s all stuff you pick up (or not) from the work.

[Blog] More work for the same salary
https://visible-quality.blogspot.com/2026/05/more-work-for-same-salary.html

How to learn the difference in writing in personal voice vs company voice? There must be more to it than an observation of company voices including levels of promise I consider almost dishonest.

My CV already mentioned I’ve done things on Azure and AWS, but I really did not foresee the work that added Google Cloud to it. Doing one thing and the choice wasn’t mine to make anymore.

If you have file (agent/skill) and copy it to a client project (copy-paste), it tends to become client owned. If you want it to remain owned by where you found it, you would bring it through package managers.

This world of sharing changed faster than people understood the implications.

Drawing random sketches with icons again, thinking through how patterns of potential/merit unfold over time.

If you end up hired on potential, you get a year where you perform as expected even when you didn't, because people keep telling that you're still learning. The standards shift though, and you are yourself supposed to notice that same next year is usually not the expected.

The higher your level, the higher the delta on your self-driven learning.

Feeling like swearing but these practical experiences of how things of past - 3 days attended - are so different too do today - 15 minutes unattended.

Not generating walls of text but seeing structure. Works so well in an exploratory intentional frame.

Reading these posts about insufficient onboarding and unclear goals as reasons to fail at your job that aren’t your responsibility and thinking about knowledge work. You ask when you don’t know. You research. You work with things you didn’t do before. It feels a little off to externalize to onboarding and goals.

What am I missing?

I have been really puzzled with my latest mentorship engagement. I think of mentorship to be about guidance, perspective, and navigating real situations. This time I had one that wanted me to teach them UML and use of tools. There are courses for those skills.

I did not do what was asked, but discussed navigating decisions on when UML would be useful, and why is it required in job ads. And showed how to model mentor-mentee relationship in AI generated mermaid UML.

Tried codeburn. I like this one. Explorers always want to analyze their pattern in hindsight to learn going forward.

[Blog] Talks turned articles: A Practitioner's Guide to Contemporary Exploratory Testing
https://dev.to/maaretp/a-practitioners-guide-to-contemporary-exploratory-testing-3d43

Arguing about use of a test tool labeled "AI tool", where the AI in it is GitHub Copilot / Claude Code. Testing area seems like it is not aware enough on architectures to assess risks in world as complex in how it is connected.

Mentor, manage, coach. I do similar things for all. I have four I mentor. 2 devs, one in same company, one in another. 1 business analyst. 1 manual tester expanding to automation. I have 8 I manage. And 5 I coach - 4 as a group, 1 solo.

Been splitting the IT world into knowledge workers and developers. I guess the difference is where problems gets handed off to. Other people or machines.

While constrained as a knowledge workers, I would not have had the drive and tools to get back to blogging against unintended corporate power. As developer there are other routes if it’s worth the effort.

[Blog] Lessons learned from AI for testers coaching
https://visible-quality.blogspot.com/2026/05/lessons-learned-from-ai-for-testers.html