Writing this from London, England

I’m grateful for the opportunity to visit London accompanying my wife on a work trip. After 35+ years of being an Arsenal fan, I was able to attend a match at Emirates. I was like a kid in a candy store. I’ll never forget it. After only a few days, I love this city. I could see myself living here. We’re having a great time. Customs was so fast. Public transit is so easy. Everything uses Apple Pay or contactless payments. And of course the architecture. The narrow streets. The ornate buildings. The scale of Regent street blew me away. All the pubs and shops in Soho. It really has been amazing.

I’ve not noticed a ton of tech, especially AI, advertising since being here. I don’t see as many iPhone advertisements either. The States is way too advertising heavy. Things in general feel more laid back here.

I’ve been spending a lot of time with Claude, Gemini, Codex and now Antigravity. The experiences are mixed. I’m still leaning towards Claude for my day to day work. It fits into my toolbox the best. I’m a long time terminal user. The output is better as well. Gemini has had timeouts. It has led me down rabbit holes that are incorrect. One in particular was looking for an ActiveStorage issue that in the end was simply changing a factory from build to create. Antigravity doesn’t fit my workflow and again Gemini timeouts. Codex has been impressing some though. The winners are definitely Claude and Codex for me.

At my day job, I’m leading a team migrating customers from two older legacy platforms to a new modern offering. One of the platforms first line of Perl was written in 1995. Amazing and terrifying. This coincides with the company transitioning from a white glove services offering to a Product/SaaS model. There is a lot to unwrap and untangle. I’m able to provide a lot of value for the product I am leading. The team is now shipping at an accelerated pace. We’re changing peoples’ views of not being confident to observing how quickly we are shipping features. That feels good but the work is not done. We have a very full roadmap.

A company transitioning from services to a SaaS is something I’ve been in a lot of conversations about over the last 6 months. Including with a local Cincinnati company that is struggling to commit 100%. It’s a hard transition. People worry about their jobs when everything becomes self-serve. But I’ve reminded people that domain knowledge is still key. Just because they won’t be clicking the buttons on behalf of the customers, their domain knowledge will help the Product offering evolve.

Engineers can be worried about transitioning from fire drill/one off bespoke solutions to multi-tenant Product offerings. Bespoke solutions by nature can feel more focused. Less scope. More siloed. The definition of done can be more absolute. Or at least, feel more absolute. But, for anyone that has worked for an agency, or one these bespoke solutions, they can be absolute fire drills with little requirements gathering. Product development is more long term or far sighted. It’s a bit of a mindshift. You have to think through multitenancy. Scale. Performance. User Experience. Archcitecture.

And then there are migration details. Do credentials and other data need to be migrated from the old services/solutions? Do the interfaces exist to do that? Is it a complete cutover, or do both systems need to be operated/managed at the same time? Is it possible to have both the systems interacting with downstream systems simultaneously.

There is no prescribed way to transition from a services company to a “Product Led Growth” company. Each service offering was different and bespoke. We’re making progress but there have been unexpected surprises along the way. Realizing momentum is starting to sway people’s opinions though. That feels good.

Getting away from the keyboard to explore more of London!