It seems to me that building a native platform directly on top of claude to truly achieve this goal:
---
> When we released the Claude Code Plugin for EventSourcingDB at Easter, we proudly listed everything it did not need: no SDK, no MCP configuration, not even a Docker container.
---
Is building on very unstable, stochastically so, ground. SDK, scripts, etc... are reliable.
I understand the ergonomics and niceties of having the "AI OF THE DAY" work well with your database provider. But the framing initially in this article through me the wrong way.
Database providers are meant to be stable reliable infrastructure which rarely have issues. AI is generative and makes mistakes sometimes. Leaning into AI usage for Databases is nice, since there is a lot of information which needs to be synthesized. But having an AI literally do it all, is not really a pattern I see working in the future.
I find it troubling that we’re apparently reaching this level of learned helplessness where it’s considered unreasonable to assume the user is able to start a container instance.
It seems to me that building a native platform directly on top of claude to truly achieve this goal:
---
> When we released the Claude Code Plugin for EventSourcingDB at Easter, we proudly listed everything it did not need: no SDK, no MCP configuration, not even a Docker container.
---
Is building on very unstable, stochastically so, ground. SDK, scripts, etc... are reliable.
Inside their docs, https://docs.eventsourcingdb.io/getting-started/running-even... it reads more like I would expect. So I think this initial line is meant to really be a marketing term for readers.
I understand the ergonomics and niceties of having the "AI OF THE DAY" work well with your database provider. But the framing initially in this article through me the wrong way.
Database providers are meant to be stable reliable infrastructure which rarely have issues. AI is generative and makes mistakes sometimes. Leaning into AI usage for Databases is nice, since there is a lot of information which needs to be synthesized. But having an AI literally do it all, is not really a pattern I see working in the future.
I find it troubling that we’re apparently reaching this level of learned helplessness where it’s considered unreasonable to assume the user is able to start a container instance.