Hello everyone!
Last week some of you started testing Appfarm AI. We’ve received a lot of valuable feedback - keep them coming and thanks to all beta testers! You may DM me if you run out of credits during this phase.
There are some very important and known limitations we are working on, and some key extensions that are also highly prioritized.
Current weakness, workaround and what we are doing to solve it:
The flow for creating new apps (with a certain scope) is unstable. Not because of Claude, but once you get passed e.g. task number 5, it starts doing weird stuff. This is because of the length of the thread, and that it gets confused trying to solve its tasks grounded in both previous thinking and messages, plus “our” instructions.
We are currently working on automatic summarization of previous messages + improvements on handling long threads, as a number one priority. Until then, try starting small, and work incrementally, and clear the thread / start new thread often.
Moving ahead:
We are working on general improvements (make it even better on the “small cases”), and new functionality based on your feedback. Hoping to release some new cool features very soon! Also, enabling AI for existing solutions is of high priority once the beta has been running for a few weeks.
We will keep you posted here in the community!
