Basil Berntsen

The complexity of using k8s is less than running a cluster. I believe a lot people accusing k8s of complexity are often mixing these up.

I like Apple but every time I think I want to do more with them they remind me that they are very firmly oriented on another kind of client.

It’s more common to see new things built with this paradigm than it is to see old things transformed. It’s also far more common to see new things using other frameworks than .NET, at least in my experience

That’s actually what I meant to imply earlier, what I’m describing as the paradigm is not that common.

Cloud native applications with continuous delivery.

Big feet take a big gun.

There’s a bit of a bias: .NET was the de facto tool for so long used to build so many things before the current paradigm became dominant that it’s easier to find examples well they have not yet embraced CD.

Another beautiful option but my vote remains Montreal :)

Out of curiosity, what would you do with macOS containers?

I vote Montreal!

When I wrote this I in fact didn’t know. @meredithmeredith.bsky.social clarified that in another thread :)

If it’s a question or one or the other, then yeah of course the better one.

I wouldn’t investigate switching to a smaller provider because that would be a loss, but using one to shore up availability in case of the failure of a larger one.

I currently find myself discussing the merits of multi cloud on a regular basis, what’s your perspective?

What about the second tier of cloud companies? Companies that have less scale but still decent size for example Digital Ocean or OVH?

This is a very insightful write up, and it makes me think. Is it possible for you to have multiple hyperscalers? If not in full redundancy, such that the failure of one would affect only a part of your user base?

🫡

It came up in my research, I haven’t tried it yet though.

Indeed, the Windows 10 end of life is why I switched to Linux :)

I would’ve gone to 11 but it doesn’t support the hardware.

I thought Mac dropped security updates after seven years?

Simple, in my context, includes simple hardware. Machines that have been running every version of windows they support for years.

Did you look into Garage?

My use case is pretty typical. Apple have beautiful designer machines and we have some in the house, there’s also a need for something simple. Historically that has been windows on low end hardware. Now since windows 11 won’t run on that hardware, it’s Linux.

Would this be similar to for example gitlab? They have a community edition with an MIT open license and then an enterprise edition with a commercial license. If you want compliance functionality for example, you need the enterprise edition.

I know how to code it I’m impressed by it.

Everybody’s a little nervous about AI, and it’s easy to channel that publicly. It’s also easy to jump on the hype train. Between these two extremes, the rest of us are just sitting here bewildered to how good this new tool is as we discover how to use it.

I’ve watched people I care about age out of this work because they stopped enjoying the process of learning. Every new advancement became something they resisted and resented.

I feel the same thing brewing in me as I age and my response is to double down. AI came at a pretty good time for me :)

There’s also how these tools enable people to learn and do novel tasks more quickly. I’ve found that becoming proficient at something I’ve never done before with an LLM is much faster than unassisted.

What are you planning on doing with it?

I’m still waiting for them to mail me the installation CD