Ask HN: How is Quilt different from ClusterHQ?

Last day at my job the co-founders were ousted by the board and CEO (the CEO who eventually wrote that post about Clusterfucked), and all the employees in the UK were laid off. Much of the US engineering team left over next few months, and new members hired, so by the time you interviewed it was almost a completely new team.

One of the those co-founders, Luke Marsden, and some other people from original ClusterHQ have started https://dotmesh.com/, which perhaps expresses the original vision better.

I can't say whether the product idea is a good one, though I've definitely talked to people who would find it useful.

As far as Quilt goes:

* The ClusterHQ vision was based on filesystem snapshots (probably Dotmesh too). Thus use case might be "take cheap snapshot of database whenever a 500 error occurs, for later debugging".

* The Quilt vision is data science-oriented, it understands the underlying data, and e.g. exposes it to data science tools like Jupyter.

So they're going after very different markets.

[0] https://quilt-data.breezy.hr/p/15c4a8807cf6-technical-lead

[1] http://archive.is/vjiLw

[2] https://news.ycombinator.com/item?id=13239032