IIUC this is addressed with the ClickHouse JSON type which can promote individual fields in unstructured data into its own column: https://clickhouse.com/blog/a-new-powerful-json-data-type-fo...

Parquet is getting a VARIANT data type which can do the same thing (called "shredding") but in a standards-based way: https://parquet.apache.org/blog/2026/02/27/variant-type-in-a...

If it’s something where they have fewer unknowns and more control and lots of experience building the same thing, then I would expect an estimate: building a deck, re-roofing a house, etc

There is a bridge in my town that is finally nearing completion, hopefully, this year. It was estimated to be completed 2 years ago.

This changes when it’s a project that has fewer unknowns, where they’ve built the same thing several times before. The same is true in software.

This seems like a good idea to have plentiful dev environments and avoid a bad pricing model. If your production instance is still on Heroku, you might still want a staging environment on Heroku since a Hetzner server and your production instance might have subtle differences.

> minimum resource requirements for good performance to be around 2x CPUs and 4 GiB RAM

This is less compute than I regularly carry in my pocket? And significantly less than a Raspberry Pi? Why is Fargate that expensive?

If these aren't enabled for containers / sandboxes yet, I bet they will be soon

With single instances topping out at 20+ TBs of RAM and hundreds of cores, I think this is likely very under-explored as an option

Even more if you combine this with cell-based architecture, splitting on users / tenants instead of splitting the service itself.

That's... largely what standards are?? And they are really beneficial??

That’s barely more than a raspberry pi? (4 vs 8 cores) Huge machines today have 20+ TBs of RAM and hundreds of cores. Even top-end consumer machines can have 512GB of RAM!

I do agree with the author that single machines can scale far beyond what most orgs / companies need, but I think they may be underestimating how far that goes by orders-of-magnitude

We can expect them to be better in 5 years, but your last assertion doesn't follow. We can't assert with any certainty that they will be able to specifically solve the problems laid out in the article. It might just not be a thing LLMs are good at, and we'll need new breakthroughs that may or may not appear.