-Andreas (part of the Neo4j community team)
Also ThoughtWorks is using it and writing about it: http://www.markhneedham.com/blog/2012/04/20/neo4j-searching-...
- James, author of Bulbs (http://bulbflow.com), one of the Python bindings
I am aware of the Heroku addon, but what happens if you need more than that?
For write scaling, you would have to stay tuned. There's a team sequestered in England feverishly working on a sharding solution to spread write load and data set size across multiple machines without compromising query performance.
Heroku is just the first cloud service to offer Neo4j. We've been chatting with other PaaS providers, and some are moving forward with their own plans. A few enterprise customers simply orchestrate their own clusters (up to global, multi-continent spanning deployments) using EC2.
You're right that all of the NOSQL solutions don't have as many experts and professional services as the relational db options. Seems like a great opportunity. ;)