Roland Eschenburg

Roland Eschenburg

@reschex.bsky.social

Principal Engineer - AI Engineering & Developer Experience. Trying to harness AI in a sustainable way to improve software delivery outcomes.

74 followers68 following263 posts

What is happening to me?? At the weekend - driving to the shops without satnav, confusing the children ("what do you mean, you don't know how many minutes??"). I blurt out "I'm vibe driving" 😭 Load of rubbish.. I was iteratively designing and crafting the route 🤦‍♂️

Roland Eschenburg reposted

Whoever started that thing of holding your phone like a tray out in front of your face in speaker mode needs to be punished without mercy. And with them every tv person who ever was part of showing it on tv. I have spoken.

Fascinating to watch the world burn from a chair with not much Php around. I got the full experience with log4j back in the day, which was great 'fun'. Not being impacted by this makes a nice change

Btw.. "software/dark factory". Stop saying it, it's not going to become a thing, it's stupid and whoever twisted this term into something that's supposedly "desirable" rather than to be avoided at all cost, deserves their pillow to be warm on both sides for the rest of their life.

I enjoyed #FastFlowConf yesterday, overall. But I can't help to also feel frustrated that most/all of what I saw was dedicated to AI in greenfield projects. I'd say 'slow down to allow for human judgement' was the overall theme - very on-brand for the current AI discourse.

View full thread

Most of us sit in some large enterprise maintaining a 5, 10, 15yo 'legacy' product that still makes tons of money. On most of those projects, coding was never the bottleneck. human review .. probably also no.

AI has made individual engineers able to produce more code and PRs faster - yet none of those companies have been reporting a noticeable jump in revenue - which is, after all, the only real measure that counts when it comes to 'have we delivered value?' - I don't know .. have more people bought it?

I enjoyed #FastFlowConf yesterday, overall. But I can't help to also feel frustrated that most/all of what I saw was dedicated to AI in greenfield projects. I'd say 'slow down to allow for human judgement' was the overall theme - very on-brand for the current AI discourse.

"Coding is no longer the bottleneck, human review is". I'm sure there are some companies out there where this is true. I'm also sure that most of them will be startups/greenfield projects. Which is not the reality for most software engineers.

LinearB considers %tage of refactored code a quality metric - nice. LinearB only considers code refactored if it was changed when it was 21 days or older - huh?

PR maturity is another great metric - measuring how much a PR changes after submission. less change being 'good'. I guess at least it really codifies that PRs are quality theater?

LinearB considers %tage of refactored code a quality metric - nice. LinearB only considers code refactored if it was changed when it was 21 days or older - huh?

Having to code on a personal Claude Pro subscription is cramping my style. Only a couple of weeks until the new job starts, which will hopefully have me back on an enterprise plan!

Listening to @lennysan.bsky.social’s podcast episode about Grok Bot - sounds amazing - shame it’s associated with Elmo - one of the worlds biggest evil idiots. Can’t believe someone with that much influence still talks about the platform formerly known as twitter and gives these people air time.

Roland Eschenburg reposted

"Do the thing you can do instead of beating yourself up over not doing the thing you're not doing" is honestly life-changing, even life-saving, advice.

Leiracal@leiracal.bsky.social

This concept was introduced to me as "run the dishwasher twice" and was an important part of my executive functioning recovery as well as my path to self-acceptance.

Another one.. new outlook.. why is the 'accept/decline' meeting button the smallest, most difficult button to find on the entire meeting invite? Seriously, have you never talked to a user?? I know what's missing from the UX.. a little Where's Waldo.

Notice how he doesn't say 'don't plan or analyse'. You are allowed to look at a map before setting off somewhere. Just don't assume one look at the beginning of the journey is enough to help you navigate around traffic jams, road closures or bio breaks.

This post is unavailable.

Roland Eschenburg reposted

WELCOME TO THE MODERN SOFTWARE ENGINEERING CHANNEL | Gojko Adzic We are extremely excited to welcome Gojko to the channel, and we are looking forward to presenting his first video in MSE colours tonight. At 7pm (UK). Join us in welcoming and celebrating Gojko 🎉💬

Roland Eschenburg reposted

Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in useful features. This is the first time I’ve seen a company say this directly. www.businessinsider.com/uber-coo-and...

Murphy’s Law is missing a crucial bit: “At some point”. What can go wrong, won’t go wrong most of the time, until one of the many things does when you least expect it. Otherwise we would always have a nicely observable set of unhappy paths.