WEBVTT

00:00:00.400 --> 00:00:01.920
One of the hardest

00:00:01.920 --> 00:00:05.040
things to do with AI assisted coding is keeping

00:00:05.520 --> 00:00:08.000
the AI on track in large

00:00:08.000 --> 00:00:09.280
coding project.

00:00:10.240 --> 00:00:12.560
And I will say probably something kinda bold,

00:00:12.975 --> 00:00:16.095
but I think AugmentCode might have just released

00:00:16.495 --> 00:00:17.775
one of the best

00:00:17.935 --> 00:00:19.375
task orchestrators,

00:00:19.375 --> 00:00:24.095
task list implementations that I've seen so far in any of the AI coding tools.

00:00:25.070 --> 00:00:26.750
And to back up a little bit,

00:00:27.310 --> 00:00:30.750
we can kind of see this sort of split direction

00:00:30.750 --> 00:00:33.390
that people are going that these companies are going.

00:00:33.950 --> 00:00:35.230
On one hand,

00:00:35.710 --> 00:00:38.270
we've got the orchestrator mode in RUCODE.

00:00:38.825 --> 00:00:47.225
What that does is it is it's kind of like a task list if you think about it. You give it a large thing. You give it a large amount of of work to do.

00:00:47.625 --> 00:00:51.465
It actually breaks it up and then orchestrates out the smaller jobs to,

00:00:52.160 --> 00:00:57.680
like, other modes. For example, you might need to do an architecture mode or you might need to go do

00:00:58.000 --> 00:00:59.120
some coding.

00:00:59.920 --> 00:01:04.720
But that actually is a really interesting thing because you're no longer, like, confined to a single context,

00:01:05.205 --> 00:01:11.605
and you're able to really keep the AI on track by that orchestrator mode, kinda dishing out the work.

00:01:12.165 --> 00:01:18.165
Where on the other side, there's we're gonna keep it all in the same context. One model chat log

00:01:18.530 --> 00:01:22.210
and so on. And one of the things that I've actually done

00:01:23.170 --> 00:01:29.490
probably for the last, I don't know, year now at this point, is I found that just small surgical changes,

00:01:29.890 --> 00:01:33.330
basically, not letting the AI do these big sweeping things, actually,

00:01:34.575 --> 00:01:38.015
has honestly helped me a lot because I can review the code easier.

00:01:38.495 --> 00:01:44.255
I can go in, and I can make sure that the AI doesn't get off track to really being surgical about it.

00:01:44.655 --> 00:01:47.615
But some other things have evolved over time. So

00:01:48.040 --> 00:01:52.280
one thing I did for a while was I was using kind of text based

00:01:52.920 --> 00:01:55.640
task list. So I would do I would have,

00:01:55.960 --> 00:01:57.320
you know, maybe ChatGPT

00:01:57.320 --> 00:01:58.920
or Claw generate

00:01:59.160 --> 00:02:05.935
a list of things that needed to be done, And then I would use the AI assisted coding tool to actually go through, read that file,

00:02:06.335 --> 00:02:08.735
and use that sort of thing as

00:02:08.975 --> 00:02:10.335
what it needed to do.

00:02:10.735 --> 00:02:13.855
But then we've had things kind of add on like Taskmaster.

00:02:14.880 --> 00:02:18.240
So Cloud Taskmaster might be one of the more popular ones.

00:02:18.880 --> 00:02:23.200
So here's an image of kind of what that looks like. This thing's actually pretty sweet, honestly.

00:02:23.760 --> 00:02:25.200
But at the same time,

00:02:25.440 --> 00:02:29.915
I would say, like, it does take a little bit extra work to get it set up. So we've got

00:02:30.395 --> 00:02:33.595
these really cool, really great implementations

00:02:33.915 --> 00:02:40.875
that basically allow you to orchestrate and keep the AI on track. Because if if any of you know, like, if we can

00:02:41.700 --> 00:02:43.860
constrain the AI to do our bidding,

00:02:44.100 --> 00:02:51.220
we can actually accomplish some pretty amazing things with it. And a lot of people talk about the big complications with AI,

00:02:51.540 --> 00:02:58.655
and I think one of the biggest complications with AI is just making it stay on track. And as the context fills up,

00:02:59.215 --> 00:03:11.150
its likelihood of staying on the path that you want it to be on kind of goes off the rails a little bit. And AI has gotten a lot better at this from where it was even, let's say, eight months ago to today.

00:03:11.310 --> 00:03:12.750
It's gotten a lot better.

00:03:13.470 --> 00:03:14.830
So that that's where

00:03:15.150 --> 00:03:20.350
Cloud Code and Augment Code have kind of come in. And there may be others too, but these are the two I know the most.

00:03:20.670 --> 00:03:24.110
Cloud Cloud Code has this built in to do list function

00:03:24.345 --> 00:03:25.465
where you

00:03:25.785 --> 00:03:31.225
can actually get it to generate its own task list so you don't have to orchestrate anything.

00:03:31.465 --> 00:03:34.825
It just manages it, and it keeps itself on track.

00:03:35.225 --> 00:03:36.505
And what AugmentCode

00:03:36.505 --> 00:03:42.600
has done is it has also just released its own version of the task list.

00:03:43.240 --> 00:03:44.680
And this thing is

00:03:44.920 --> 00:03:46.200
freaking awesome.

00:03:46.600 --> 00:03:50.760
And I'm gonna go through some of the features on it, but the big thing I wanna cover is

00:03:51.645 --> 00:03:54.045
it went from 63.2%

00:03:54.045 --> 00:03:56.845
in my evals to 67.5.

00:03:57.085 --> 00:04:01.245
And I've actually got an early version of my site up if you go to the best AI agents.

00:04:01.885 --> 00:04:04.205
You can see here if we search by

00:04:04.845 --> 00:04:05.965
CloudForms on it,

00:04:06.820 --> 00:04:11.780
the last time that I had ran AugmentCode was sixty three point two o, and that was on May 30.

00:04:12.260 --> 00:04:14.260
The task manager alone

00:04:14.580 --> 00:04:19.460
has improved its scoring quite substantially right in line with Klein and RootCode,

00:04:19.975 --> 00:04:22.775
know, maybe slightly behind that. Still number seven spot,

00:04:23.015 --> 00:04:29.815
but it is all relatively within margin of error when you see, like, a fraction of a point difference between them.

00:04:30.455 --> 00:04:33.175
So pretty huge boost very consistently

00:04:33.960 --> 00:04:38.680
from claw or from AugmentCode because of the task list management.

00:04:39.000 --> 00:04:39.560
Now

00:04:39.800 --> 00:04:42.600
they just make a huge difference.

00:04:43.800 --> 00:04:45.800
Ever since ClogCode,

00:04:46.360 --> 00:04:51.165
and I've been working with that, I very rarely have Cloud Code go off the rails.

00:04:51.645 --> 00:04:52.205
Where

00:04:52.445 --> 00:05:03.260
if I go back a few months ago, some of you might have known that I was venting a little bit about Cloud three dot seven because it would just go do things I didn't ask it to do. I don't have that problem anymore.

00:05:03.580 --> 00:05:11.660
And in fact, I would say clog ClogCode, in particular, has kept my my, um, what I'm actually executing, like, really, really well constrained.

00:05:11.820 --> 00:05:27.315
And I've just totally enjoyed working with it because I don't have to worry about it doing things I don't want it to do. But let's talk about AugmentCode a little bit because what what I said before is I think they have created one of the best, if not the best implementations

00:05:27.635 --> 00:05:28.675
for task list.

00:05:29.640 --> 00:05:34.440
So one thing to note is not everything is going to actually generate a task list.

00:05:35.000 --> 00:05:45.400
So in this particular one, I was actually had it debugging an error message for me. It did not generate a task list here. But if I go into some of my past ones let's take a look at this one. I think this one

00:05:45.875 --> 00:05:46.915
may have.

00:05:47.235 --> 00:05:51.155
It did not. Okay. So let me go back one more. K. Here's one that actually did.

00:05:51.475 --> 00:05:55.395
So this one actually generated a task list of things that it wanted to do.

00:05:56.195 --> 00:06:04.810
And I'm just gonna stay here for a second and kinda show some of the functionality that I've really enjoyed. So the first thing is you can just manually add a new task.

00:06:05.290 --> 00:06:06.490
And this is

00:06:06.970 --> 00:06:08.570
incredible, in my opinion,

00:06:08.890 --> 00:06:12.570
because I can come in and be like, I want this to do x y z.

00:06:13.335 --> 00:06:16.455
And I've used this quite a lot over the last day and a half.

00:06:16.935 --> 00:06:23.015
Because sometimes I don't even I like what the AI did or I wanna actually come in and I wanna change

00:06:23.095 --> 00:06:24.215
this to something else.

00:06:25.250 --> 00:06:30.770
I can do that without having to actually have the AI generate and use tokens to make that happen.

00:06:31.010 --> 00:06:33.170
I can go in and actually edit it myself.

00:06:33.410 --> 00:06:38.770
Now once the task list is in place, you literally can hit run all task, and it will just run through them.

00:06:39.825 --> 00:06:43.585
The other thing you can actually control is the status yourself,

00:06:43.825 --> 00:06:49.505
which I thought was really interesting. It will auto get updated as well as it actually gets completed,

00:06:49.905 --> 00:06:53.425
but you can also come in and you can say, oh, I don't wanna do that one.

00:06:54.210 --> 00:06:57.010
That one's complete, or you can just go ahead and delete them.

00:06:57.330 --> 00:06:59.170
So really, really very,

00:06:59.170 --> 00:07:02.130
very cool because when you think about clogged code, for example,

00:07:02.930 --> 00:07:05.250
you know, you get, like, this plan, for example,

00:07:05.890 --> 00:07:08.050
that I could go ahead and approve or not.

00:07:08.915 --> 00:07:10.915
I can't really change

00:07:11.555 --> 00:07:12.435
individual

00:07:12.435 --> 00:07:13.715
things very easily.

00:07:14.355 --> 00:07:19.235
So for example, if I if I like phase one, but I wanna change something in phase two,

00:07:19.635 --> 00:07:30.400
the way I need to do that is communicate back to it to have it make that plan. And that has been fine for me. Like, honestly, I I haven't had a lot of issues or concerns with doing that.

00:07:30.720 --> 00:07:38.320
But as I've started using all of my codes new task list, I've realized how much nicer it is to be able to come in and actually tune

00:07:39.335 --> 00:07:40.695
the task itself,

00:07:41.175 --> 00:07:43.175
remove a task, add a task.

00:07:43.415 --> 00:07:45.895
So this thing is incredibly powerful.

00:07:46.535 --> 00:07:51.255
And what I would say is, um, if you want it to generate a task list,

00:07:51.780 --> 00:07:59.060
it seems to do, I would say, a really good job of finding the right times rather to actually use a task list,

00:07:59.460 --> 00:08:04.660
and it finds the right times when it shouldn't use a task list. Because I've I've had somewhere

00:08:04.660 --> 00:08:06.020
when I put in a query,

00:08:06.905 --> 00:08:14.825
it knows that it's complex enough and it's like, boom. Here's my eight things. I'm gonna go ahead and do it. And it did that with every one of my evals that I put in.

00:08:15.225 --> 00:08:16.345
Where before,

00:08:16.425 --> 00:08:18.185
I think it scored poorly

00:08:18.425 --> 00:08:34.340
because a lot of times the agent would just kinda give up. Give up is the wrong word, but it wouldn't complete it to the fullest extent that it needed to. And, the task list is harnessing the AI to accomplish the particular goal that we actually have here. I would highly recommend

00:08:34.965 --> 00:08:37.845
giving this a try. It's this little icon here

00:08:38.725 --> 00:08:43.045
right beside the file change. Now the one thing that I would say is I I sorta wish

00:08:43.605 --> 00:08:46.885
I had a little bit more control on the size of this thing.

00:08:47.365 --> 00:08:48.005
I

00:08:48.300 --> 00:08:55.260
it's just either open or close, and I would love to be able to, like, bring it down to, like, two or three show. I know this is a minor nitpick,

00:08:55.820 --> 00:09:03.020
but I do keep trying to, like, drag it up and down this. And I want this a certain size, and I want this a certain size,

00:09:04.105 --> 00:09:08.265
kind of a minor a minor issue. But it is just worth kinda calling out.

00:09:08.825 --> 00:09:13.785
The other thing that I would say is you can actually take your task list and put it into a new chat.

00:09:14.425 --> 00:09:17.865
So I've done this a little bit where I was actually planning through something.

00:09:18.810 --> 00:09:25.210
And then then now that I have my task list, I just started a brand new context window with it in a brand new chat.

00:09:25.610 --> 00:09:32.170
I have not tested the import from markdown yet. I am very interested in checking that out because you could go in,

00:09:32.775 --> 00:09:43.095
you know, o three, for example, have it help you generate the PRD or the technical task list that you want and bring it in, and that could be kinda sweet as well.

00:09:43.735 --> 00:09:52.390
And I haven't really played around with exporting yet because I haven't really had much of a need to. But continuing new chat is the one that I've I've used a couple times at this point now.

00:09:53.030 --> 00:09:58.710
So, anyway, I just wanted to get kinda touch on this real quick and just share that massive improvements

00:09:58.710 --> 00:09:59.430
on

00:09:59.990 --> 00:10:03.605
evals with this. It really does rain in the AI.

00:10:04.245 --> 00:10:10.165
I felt this in Clog code, and that's probably one of the reasons why I like Clog code so much. I hadn't quite put that,

00:10:10.485 --> 00:10:12.085
you know, that into

00:10:12.485 --> 00:10:26.360
into words yet because Cloud Code is so controllable. I've said that. It's so controllable. And I think a lot of it has to do with the way it kind of breaks out its own tasks that it's doing. And I think Augment code, now that it has this ability,

00:10:27.645 --> 00:10:30.525
it's just gonna become a lot more powerful for people.

00:10:30.845 --> 00:10:37.645
Combined with its amazing context engine, this thing is a beast, and I'm excited to kinda see what they go next because I was not expecting

00:10:37.885 --> 00:10:39.805
something like this. And in fact, I would actually

00:10:40.300 --> 00:10:48.860
be surprised if we did not see things like this in some of the other AI coding tools because it just makes way too much sense for the way they have this implemented.

00:10:49.020 --> 00:10:50.460
Really great implementation.

00:10:50.780 --> 00:11:04.265
Really excited to kinda see what they do next. Anyway, I'm gonna wrap it up there. Let me know what your thoughts are below. Have you had a chance to try this out? If not, you should just definitely go check it out because this thing is freaking awesome. Till next time, everyone. Have a wonderful day. Peace out.