its actually using juyipter to run the code ... this is a good method as it means mini kernals can be created for different dev kitss.... obviously using jupyter_Client without the note book .... but perhaps it should use notebooks for the sessions ... hence having the sessions stored in notebook form... as well as the environmenets can be created and reloaded for each project type, ie research toolts would be in a single notebook .... i think these guys have missed that the folder the interpreter uses to store code can be augmented by dropping modules inside the folder ....giving the model automatic skills by dropping them in the folder... there should also be a document cache for the model as it finds information it should convert them to text files to load as a rag when requred as the rag should be disposable ie session based but the text files static... such as txt,md,html,json etc... but its ok !! ipython/ikernal
![[1hr Talk] Intro to Large Language Models](http://i.ytimg.com/vi/zjkBMFhNj_g/mqdefault.jpg)
![[1hr Talk] Intro to Large Language Models](/img/tr.png)
This Open Interpreter thing sounds really nifty
its actually using juyipter to run the code ... this is a good method as it means mini kernals can be created for different dev kitss....
obviously using jupyter_Client without the note book .... but perhaps it should use notebooks for the sessions ... hence having the sessions stored in notebook form... as well as the environmenets can be created and reloaded for each project type, ie research toolts would be in a single notebook ....
i think these guys have missed that the folder the interpreter uses to store code can be augmented by dropping modules inside the folder ....giving the model automatic skills by dropping them in the folder... there should also be a document cache for the model as it finds information it should convert them to text files to load as a rag when requred as the rag should be disposable ie session based but the text files static... such as txt,md,html,json etc...
but its ok !!
ipython/ikernal
Amazing talk! To quote Willam Gibson - "The future has arrived - it’s just not evenly distributed yet."
Can't wait for the desktop app (on windows~)
Thanks so much for recording this. I was not able to come as I anticipated