1
00:00:02,980 --> 00:00:10,060
How do we use OpenAI functions with the new long chain expression language?

2
00:00:10,180 --> 00:00:18,430
So in this very simple example, we are using OpenAI functions in the classic way.

3
00:00:19,240 --> 00:00:26,500
And remember that the key point was to describe the function in the proper way.

4
00:00:26,500 --> 00:00:39,220
And then with a the create OpenAI function chain, we can a create the chain that has the functions

5
00:00:39,220 --> 00:00:41,110
that we have described.

6
00:00:42,140 --> 00:00:50,870
So the new way in that Lansing is proposing is a using pipes to create the the chain.

7
00:00:50,870 --> 00:00:56,390
And right now we are using the dot bind and we are including two things.

8
00:00:56,390 --> 00:01:04,400
First, the function, the name of the function and that we have described, and then the name of the

9
00:01:04,400 --> 00:01:08,960
variable where we have included all the function descriptions.

10
00:01:08,960 --> 00:01:09,500
Okay.

11
00:01:09,500 --> 00:01:14,210
So we are using the language model dot bind.

12
00:01:14,210 --> 00:01:20,690
And then we have a function call with the name of the function we want to use.

13
00:01:20,690 --> 00:01:28,250
And then the name of the variable that A includes all the function and descriptions.

14
00:01:28,400 --> 00:01:30,470
Remember that a.

15
00:01:30,470 --> 00:01:35,120
You have more information about this new way of working with line chain.

16
00:01:35,120 --> 00:01:43,670
In the line chain documentation where you can find a more cases and more ideas to experiment.