NEO is now available for open beta.
We’ve had some hiccups the past week with changes done by Anthropic, but we’ve adapted to them.
Really made me think about how reliant we are on Anthropic though.
But I suppose it’s the reality of working with an LLM and it's on me to figure out systems to adapt quickly in the future.
I’ll prepare more buffers so that NEO users’ experiences are protected no matter what the corporations do.
Meanwhile, we’ve made changes to NEO’s system prompt so that it:
Gives only one suggestion
AFTER it gathers enough information about your goals and context
One question at a time
It now thinks things through from the big picture first before going down into the fundamentals.
What this means: even though we’re operating with Sonnet 4 we’re not letting the issues of the later models affect how NEO is able to ‘get’ you.
From my testing, it looks like it’s back to the OG AI Coach feeling and experience.
So far, it feels great.
But I need more people testing NEO.
If you haven’t yet already, you can start trying NEO with 100 free credits.
100 credits = about 1 long conversation.
I suggest bringing to NEO a real decision-making problem that’s quite significant.
That does not require any form of uploading knowledge files.
(I’m not quite happy with the RAG function…yet)
Get NEO here: https://www.meetneo.app
FELIX