Tonight's recovery meeting featured another very familiar round of men being honest about their own frailties and fears, one of them breaking into tears in love and gratitude and astonishment, and...well, it's kind of amazing how commonplace this is in my life.

Men, if you learn how to get honest and open and vulnerable, really share with someone who isn't your wife, girlfriend or mother, *you will be unlocking the door to your cage*. Also, you'll realize you were in a cage. It's freedom.

Once again a copy of War And Peace has defeated input validation during a blitz test 😈

It's not that people will do this on purpose, it's just a really easy to way to check if your system actually has limits and whether it tells your users how to recover.

https://www.gutenberg.org/files/2600/

I've started as the ABC’s national AI tech reporter, after 5 unreal years at Crikey. AI is one of the biggest stories of our time. The sheer amount of money, influence and hype makes it ripe for critical, nuanced and accessible reporting.

Got a story about Aussies and AI? [email protected]

Hello #parents — I imagine most of you have received some kind of communication from your kids' schools about the #Instructure / #Canvas data breach. I'm collecting examples of how schools have communicated the issue to parents and students (and *if* they've communicated directly to students).

Shoot me an email if you have anything you could share: [email protected]

I spent years manually taking product screenshots. Cropping them. Replacing them. Seeing many outdated images. Losing the will to live.

So I automated it with Playwright. 585 screenshots, every theme, every viewport, generated in minutes.

But the real surprise? Opening the output folder turned into the best design QA tool I've ever had. Inconsistencies you’d never notice while navigating suddenly become obvious when you see everything at once.

https://calibreapp.com/blog/screenshot-pipeline-qa-review

Been thinking about this old post where I failed to convince people about the unreasonable effectiveness of randomized testing, fuzzing, etc.: https://danluu.com/testing/

Since switching from hardware to software, I've constantly had people tell me, that won't work for their project because of X.

Now that I've been doing software for a while, I've fuzzed a thing that has X for every X people told me makes it infeasible and they were all easy to do, so I no longer believe comments of that form.