Heads up! To view this whole video, sign in with your Courses account or enroll in your free 7-day trial. Sign In Enroll
Preview
Start a free Courses trial
to watch this video
Explore the standout features and unique strengths that set each chatbot apart in this comprehensive comparison.
This video doesn't have any notes.
Related Discussions
Have questions about this video? Start a discussion with the community and Treehouse staff.
Sign upRelated Discussions
Have questions about this video? Start a discussion with the community and Treehouse staff.
Sign up
This last demonstration isn't a test, but
0:04
rather a showcase of
each chatbot's strength.
0:07
I've picked one category where
I feel each chatbot excels and
0:09
I wanted to show these
strengths in action.
0:13
Up first we have ChatGPT.
0:16
For ChatGPT's ultimate strength,
I chose the advanced data analysis model.
0:17
While ChatGPT has plenty of strengths,
I think this one in particular leads
0:23
the pack, especially when it comes to
tasks like coding and data visualizations.
0:28
For this example, I want to use ChatGPT's
advanced data analysis model to create
0:34
a small brick breaker game
that plays in the browser.
0:39
Now I tried to replicate this
task with other chatbots and
0:42
they didn't even come close.
0:45
Bard couldn't complete the code
without running into its token limit.
0:47
Bing refused,
saying it wasn't able to create it.
0:51
And I was never able to get
functioning code with Claude.
0:53
But let's see how far we
can get with ChatGPT.
0:56
I'll paste my prompt in, and just like
that ChatGPT begins generating our code.
1:04
It starts with the HTML,
1:09
Then generates the CSS and JavaScript as
well, because this was such a large ask
1:17
on my end, we will have to give ChatGPT
a few moments to produce the full code.
1:21
But it looks like it is done.
1:40
Let's copy the HTML and
head over to VS Code.
1:41
Once there,
I'll paste the code and re-save.
1:46
Now let's go back and get the CSS.
1:48
Then I'll paste that in my CSS file.
1:53
Now let's go back one last time for
our JavaScript.
1:58
I'll paste my JavaScript,
now let's save and
2:04
check out the browser to see what we get.
2:06
Okay, the game seems to be working but the
bricks are changing colors like crazy, and
2:09
the paddle moves kinda choppy/stiff.
2:13
Let's go back and see if we can have
ChatGPT help us debug our code.
2:16
Back in ChatGPT I'll paste my new prompt.
2:20
Can you help me debug the code?
2:23
The blocks are flashing.
2:24
I want them to be one solid colour but
each block should be unique.
2:26
Also, the paddle movement isn't smooth.
2:29
And without skipping a beat,
ChatGPT agrees to help us and
2:34
starts working on a fix.
2:38
But since this will take a while, we'll
skip the finished results like before.
2:40
All right,
it looks like ChatGPT is done, but
2:51
it didn't update the original code for me.
2:54
I'll ask it to do that, and
then we'll skip to the finished response.
2:57
Now that it's done, I'll copy the code and
I'll head over to Visual Studio Code.
3:06
Once there, I'll paste it and
re-save my JavaScript file.
3:10
Now let's refresh the browser
window in Presto.
3:18
We have a working brick breaker game.
3:20
Now you get to witness my sick brick
breaking skills, just kidding but
3:22
seriously you could even continue to
prompt ChatGPT for more advanced features,
3:26
like having each color take a different
amount of hits to fully break the brick
3:30
and more.
3:35
But let's head over to Bing AI to
check out one of its main strengths.
3:35
And we're back in Google Bard.
3:41
For this demo, I wanna show off what I
think is Bard's greatest strength, and
3:43
that is its extensions.
3:47
It's natively integrated to work with
Google workspace, Gmail, YouTube,
3:49
and more.
3:54
And for this example, I wanna find hotels
for a four-day trip to San Francisco for
3:55
New Year's Eve, and I'll tell Bard to
also include a possible itinerary.
3:59
Bard starts using
the Google Hotels extension.
4:03
You can see this is the case as the
loading icon changed from the sparkles to
4:05
a bed icon.
4:09
Bard starts off with a disclaimer that
it doesn't always get it right, and
4:11
to be sure,
check the following hotel deals.
4:14
It then offers me several hotels in
the San Francisco area with links to book.
4:16
Those same links are even
embedded in the reply.
4:20
Bard also includes
the itinerary saying Day 1,
4:23
we could walk around Union Square and
do some shopping and dining.
4:25
Day 2, it suggests visiting the Golden
Gate Bridge and Alcatraz Island.
4:29
Day 3, visit the Mission District and
explore the murals and shops.
4:34
And Day 4, take a ferry to Sausalito and
4:38
have lunch at one of the many restaurants
with views of the Golden Gate Bridge.
4:40
And here are those hotel
embeds I mentioned earlier.
4:44
But let's try another example.
4:51
Let's say my printer is
giving me an error message.
4:53
It's saying my print head isn't installed.
4:55
Let's try asking Bard to help.
4:58
I'll paste the following prompt.
4:59
My OfficeJet 8600 printer
won't install the print head,
5:01
can you show me how to fix it?
5:05
It looks like Bard is trying to
suggest tips to fix my printer itself.
5:10
I'll just re-prompt it.
5:15
This time I'll use the at
symbol to pick my extensions,
5:19
just like tagging someone
on a social media platform.
5:22
I'll select YouTube and
resubmit my prompt.
5:26
And just like that, Bard has
provided me with several videos to
5:36
help me fix the issue with my printer and
its printhead.
5:40
The real value here is when you start
to use this with your phone out in your
5:44
everyday life.
5:47
You can quickly see how this can change
from how we traditionally search for
5:48
information on the web.
5:52
But let's move on it to Bing AI so
we can check out its greatest strength.
5:53
Okay, we're back in Bing AI, and
in this demo, I wanna showcase one of
5:57
Bing's greatest strength, the ability to
generate images within the conversation.
6:00
Bing is utilizing the DALI 3 model and
in doing so is able to produce some high
6:05
quality images, some of which even
include text that the AI got right.
6:10
So let's give it a shot.
6:14
I'm gonna keep our conversation
style set to creative for this demo.
6:16
And then I'll paste my prompt.
6:19
Can you create a flat icon illustration
of a wolf with the word wolf underneath?
6:21
All right, and after a few moments,
Bing AI has generated two wolf icons.
6:41
Both spell wolf, although the F on
the icon to the left has some issues, but
6:45
it's amazing to see how far these image
generation models have come in so
6:50
little time.
6:54
Let's try another example.
6:57
Create an icon of a fish with
the words bait and tackle underneath.
6:58
Bing says it will try to create it.
7:07
So we'll just wait a few seconds.
7:09
And we now have three images of
a fish with a bait and tackle sign.
7:18
Pretty cool.
7:21
Now, none of these images
got the text perfect, but
7:23
we could keep trying to get it perfect.
7:25
In my experience, the shorter the text,
the better your chances for
7:27
a perfect generation.
7:30
Let's try another version with Bait Store.
7:31
Okay, awesome.
7:45
See, these aren't perfect either,
but they're getting much closer.
7:46
Let's try one more example.
7:55
This time I'll paste my prompt that says,
7:56
create an image of a convenient
store named Ben's Bargains.
7:59
There we have it, four images of
a convenient store named Ben's Bargains.
8:09
Some of the images got the text
better than others, but again,
8:13
that's the luck of the draw
with generative AI.
8:16
Lastly, I'll click the preset response
to change the name to Ben's Groceries.
8:19
Let's see what I get.
8:23
Neat, I got two images of
stores named Ben's Groceries.
8:29
This technology is truly amazing and
it's evolving at the speed of light.
8:32
It's seriously a game changer when it
comes to creating content for projects,
8:37
and I would encourage you to check it out.
8:40
But let's move on to our next chatbot,
Claude.
8:42
For the last demo of this course,
I'm back in Claude AI and
8:46
I'm here to showcase Claude's greatest
strength, which, in my opinion,
8:49
is its ability to
summarize large documents.
8:53
This goes hand in hand with its ability
to handle extremely long prompts.
8:56
Now, of course, Claude isn't the only chat
bot that can handle this type of task,
9:01
ChatGPT, especially the advanced
data analysis model,
9:05
is quite good at this as well.
9:08
But at this time,
I think this is where Claude truly shines.
9:10
In this example,
I have another white paper from Microsoft.
9:14
This time, one from their Azure
team to serve as a guide for
9:17
creating proof of concepts for developers.
9:20
As you can see, there is a lot
of information in this document.
9:23
It's text heavy and
has multiple diagrams and graphs.
9:25
Plus, it's 54 pages in total.
9:29
So, let's see if Claude can review this
white paper for us and provide us with
9:32
a summary, saving us time and
energy to focus on other important tasks.
9:36
I'll copy all of the text
using Command+A to select, and
9:40
right clicking to select copy.
9:44
Then I'll head back to Claude and
paste that in.
9:46
Next I'll use my prompt, which is a bit
longer to ensure higher accuracy.
9:48
The prompt is, please summarize the key
points of this article in simple terms
9:52
that are easy to understand.
9:56
Focus on explaining the main purpose or
goal of the article.
9:58
The key topics, concepts,
or ideas covered.
10:02
Any major advantages or
benefits discussed.
10:05
Any potential limitations,
disadvantages or considerations mentioned.
10:07
Aim for a summary that captures the
essence of the article using clear concise
10:12
language suitable for general
audience without technical expertise.
10:16
Please also provide some specific details
on the pros and cons or advantages and
10:20
disadvantages of the main topic.
10:24
And Claude is working hard
to generate its response.
10:27
It has created several sections for the
summary such as main purpose, key topics,
10:29
benefits of POCs and considerations.
10:34
This is looking great.
10:37
All right, let's review Claude's response.
10:38
Claude starts with a summary of
key points from the article.
10:45
It says the article provides guidance for
developers on how to plan and execute
10:48
a successful proof of concept project to
evaluate new technologies or concepts.
10:52
The guide's main purpose is to help
developers learn best practices of proof
10:58
of concept projects to validate
feasibility before larger investments.
11:02
It covers key topics such as defining
goals, metrics, timeline, budget and more.
11:08
Based on the article, Claude has created
a benefits of POC section that includes,
11:12
experiment with technologies
in a low risk way,
11:17
get insights and validate concepts
quickly before major investment, and
11:20
identify technical and
logistical issues early.
11:24
This is perfect.
11:27
This summary is helping me get a general
idea of the concepts taught in
11:28
the article.
11:31
But let's say I wanted a bit more detail,
especially on some of the subtopics
11:33
mentioned, like under POC Project
Planning, I'll ask Claude to help with
11:37
this by asking, can you expand
on the subtopics a bit, please?
11:41
We'll give Claude a few moments
to generate its response.
11:48
And it looks like Claude has finished.
11:58
Let's head back up top
to review the response.
12:00
Under the Planning a POC Project
section we have more details such as,
12:03
define goal and success criteria upfront.
12:07
Be specific on what you want to evaluate.
12:09
Set a timeline and
fixed budget to prevent scope creep.
12:12
And scope should be minimal
to prove the concept quickly.
12:15
It also included more details
under the Azure overview section,
12:18
including the different cloud models
Azure offers, as well as a sample project
12:22
section where the white paper showcased
examples of proof of concept projects.
12:26
All in all, this is an excellent
way to research topics and
12:31
get brought up to speed quickly.
12:34
This concludes our course for
AI chatbot comparisons.
12:36
What did you think?
12:40
Did you learn any new techniques for
prompting and working with chatbots?
12:41
Did you learn about any new
bots such as Bard or Claude?
12:44
And were you surprised by what all you
can accomplish with help from these AI
12:47
assistants?
12:51
I know I was.
12:52
AI chatbots are evolving every day and
we've only scratched the surface of
12:54
what these assistants will be
able to help us do in the future.
12:57
But thank you for watching, and
I hope to see you in the next one.
13:01
Happy coding and happy prompting.
13:04
See you later.
13:06
You need to sign up for Treehouse in order to download course files.
Sign upYou need to sign up for Treehouse in order to set up Workspace
Sign up