-
Notifications
You must be signed in to change notification settings - Fork 15
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove while and storages #16
Conversation
|Consider the ultimate objective of your team: ${objective.value}. | ||
|Do not remove any tasks. Return the result as a numbered list, like: | ||
|#. First task | ||
|#. Second task | ||
|Start the task list with number $nextTaskId.""".trimMargin() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not sure why this nextTaskId was used 🤔
I currently start with TaskId
1 for the original objective, and then increment TaskId
s when new generating new tasks through the getNewTasksOrComplete
function. The prioritizationAgent
should re-use the TaskId
s generated by getNewTasksOrComplete
but not create new TaskId
s, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yes, we need to revisit the prompts to make them more precise and ensure we are not just generating new tasks. currently seems to fail to complete an objective if it goes into task generation. Seems we are not adding the context properly or in the right format
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I’ll setup logging properly, so we can more easily figure out what’s going on.
0507f18
to
a886b71
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thanks @nomisRev !
* AI as a language primitive via an auto-agent * Small clean-up, and fix ktor-client-js issue * Move around HttpClient dependencies * Add Simon's PR suggestion to fix bug listing tasks + additional example * Remove while and storages (#16) * Flatten structures * Improve Task and TaskResult model * Improve logging, and setup example module (#17) * Improve logging, and setup example module * Fix package, split logback into TOML, and remove old reference to storage * Simplify Auto agent by depending on just `ai` function + basic agents draft (#18) * autonomous agent in terms of `ai` function with self reasoning and agent support. Extended `ai` function to accept agents * wikipedia agent, currently failing JS node on import child_process.ExecOptions, io multiplatform solution based on https://github.com/jmfayard/kotlin-cli-starter * wikipedia agent, currently failing JS node on import child_process.ExecOptions, io multiplatform solution based on https://github.com/jmfayard/kotlin-cli-starter * Fix build --------- Co-authored-by: Simon Vergauwen <nomisRev@users.noreply.github.com> --------- Co-authored-by: Simon Vergauwen <nomisRev@users.noreply.github.com>
Having context receivers would drastically improve this design, as everything could be top-level functions and we could use tailrac in
ai
DSL. This is currently non-problematic sinceai
will never be such deeply nested stack. It is only as deep as Json retries.