[00:00] (0.08s)
Google just released what might be the
[00:01] (1.76s)
most insane AI coding tool ever. But the
[00:04] (4.64s)
number one question on everyone's mind
[00:06] (6.56s)
is, does it beat the king of AI coding,
[00:09] (9.60s)
Claude Code? What if I told you one of
[00:11] (11.92s)
these is clearly the king of AI coding?
[00:14] (14.72s)
But it's not for the reason you think.
[00:16] (16.64s)
In this video, I'll put Gemini CLI and
[00:19] (19.44s)
Claude Code head to head to figure out
[00:22] (22.24s)
which is the best AI coding tool ever.
[00:24] (24.72s)
We'll build an app side to side on both.
[00:26] (26.96s)
You'll learn how to use each and by the
[00:29] (29.20s)
end of the video, you'll know which tool
[00:30] (30.88s)
is king and which you can throw away
[00:32] (32.56s)
forever. Let's get into it. So, a few
[00:34] (34.72s)
days ago, Google released their claw
[00:36] (36.56s)
code competitor, Gemini CLI. Google went
[00:39] (39.68s)
full mask off. This is a complete clone
[00:42] (42.32s)
of Claude Code. It works the same. It
[00:44] (44.16s)
behaves the same. It looks the same. It
[00:46] (46.40s)
is basically the same tool. For those
[00:48] (48.24s)
who haven't been paying attention,
[00:49] (49.60s)
there's a reason why they copied Claw
[00:51] (51.76s)
Code so blatantly. Clawed Code is by far
[00:55] (55.60s)
the best AI coding tool ever. It's been
[00:57] (57.68s)
taking the entire AI space by storm the
[01:00] (60.00s)
last few weeks. But here's the minor
[01:02] (62.32s)
difference though, and this is where
[01:03] (63.60s)
Google went savage mode. They made the
[01:06] (66.40s)
first,000 requests every day free on
[01:09] (69.60s)
Gemini CLI. With Clawed Code, you can
[01:12] (72.48s)
pay up to $200 a month if you're
[01:15] (75.44s)
building any sort of serious app. This
[01:17] (77.44s)
$17 tier is basically useless. You get
[01:19] (79.76s)
like five requests. you really want to
[01:21] (81.68s)
be using the $200 plan, which is pretty
[01:24] (84.56s)
expensive for most people. So, the fact
[01:26] (86.48s)
that Gemini CLI is free for the first
[01:29] (89.20s)
thousand requests is mind-blowing. But
[01:31] (91.60s)
none of that matters though if Gemini
[01:34] (94.00s)
CLI stinks. So, here's the test we're
[01:36] (96.64s)
going to do. On your screen here is two
[01:38] (98.72s)
instances of cursor. On the left, we're
[01:40] (100.96s)
going to load up Claude Code and cursor.
[01:42] (102.64s)
On the right, we're going to load up
[01:43] (103.84s)
Gemini CLI and cursor. We're going to
[01:46] (106.24s)
give them the same app to build and
[01:48] (108.56s)
we're going to see how it plays out. As
[01:50] (110.64s)
each AI agent finishes their task, I'm
[01:53] (113.60s)
going to immediately give them another
[01:54] (114.96s)
task so we can see which one works
[01:56] (116.72s)
faster. We're going to build out a full
[01:58] (118.80s)
complex stock tracking app. This isn't
[02:00] (120.72s)
going to be your mama's to-do list or
[02:03] (123.36s)
habit track or anything like that. This
[02:04] (124.56s)
is going to be a complex stock tracking
[02:06] (126.00s)
app to see how well these two perform.
[02:07] (127.84s)
Then we're going to look at the results
[02:09] (129.20s)
and we're going to rate each AI tool
[02:11] (131.36s)
based on speed, accuracy, taste, and
[02:15] (135.12s)
ease of use. We're going to see which
[02:17] (137.04s)
tool earns your hard-earned money and
[02:19] (139.68s)
which is the best. So, let's do this.
[02:21] (141.44s)
Let's load up both. If you haven't
[02:23] (143.12s)
installed Claude Code in Gemini yet,
[02:25] (145.28s)
I'll put the commands down below that
[02:27] (147.04s)
you can just copy and paste in. Feel
[02:28] (148.80s)
free to follow along with me here so
[02:30] (150.16s)
that you have your own app at the end of
[02:31] (151.76s)
this too that you can build out. So, I
[02:33] (153.76s)
have them both installed already. Let's
[02:35] (155.20s)
just load them up. For Claude, all you
[02:36] (156.80s)
do is type in Claude into your terminal,
[02:39] (159.04s)
hit enter. That's all loaded up and good
[02:40] (160.88s)
to go. And then for Gemini, it's
[02:42] (162.32s)
basically the same exact thing. We type
[02:43] (163.84s)
in Gemini, I hit enter, and that's
[02:46] (166.16s)
loaded up as well. So, what I did was
[02:48] (168.00s)
this. I went into Claude. I had it build
[02:50] (170.08s)
me a prompt for a stock tracking tool
[02:53] (173.28s)
that I can now just copy. And we're
[02:55] (175.36s)
going to put it in both. Basically, what
[02:57] (177.28s)
this prompt is going to do is build out
[02:59] (179.28s)
the stock tracking tool. It's going to
[03:01] (181.36s)
have a stock chart in it, the price
[03:03] (183.20s)
chart. It's going to have an AI company
[03:04] (184.80s)
summary. It's going to have full
[03:06] (186.00s)
sentiment analysis. It's going to have a
[03:07] (187.52s)
watch list in it. We're going to take
[03:09] (189.20s)
this prompt, put it into both, and we're
[03:11] (191.04s)
going to get to work on each to see what
[03:12] (192.56s)
kind of app they build out. So, I'm
[03:14] (194.24s)
going here. We'll paste this into Claude
[03:16] (196.40s)
and we'll paste this into Gemini. Uh,
[03:20] (200.64s)
now let's hit enter on both and see what
[03:23] (203.36s)
kind of results we get. So, here we go.
[03:25] (205.04s)
Enter. Enter. So, 0.1 second advantage
[03:28] (208.16s)
to Claude Code because I hit enter on it
[03:29] (209.68s)
first, but let's see what happens. So
[03:30] (210.88s)
they're both Claude Code is
[03:32] (212.64s)
pontificating while Gemini CLI is
[03:35] (215.36s)
planning the stock tracker. I think I
[03:37] (217.12s)
prefer pontificating a little bit more.
[03:39] (219.44s)
Claude Code is going in and updating a
[03:41] (221.36s)
to-do list. So it's building out its
[03:42] (222.64s)
to-do list for us so it can do it step
[03:44] (224.24s)
by step. Gemini is just giving us the
[03:46] (226.72s)
plan which it's already done. The plan
[03:48] (228.16s)
is already done. Of course I will build
[03:49] (229.84s)
out a sleek AI powered stock tracking
[03:51] (231.84s)
application. Here is my plan. Claude
[03:54] (234.08s)
code is still going here. So it has the
[03:55] (235.60s)
front end uh styling AI integration
[03:58] (238.72s)
interactive search dynamic price. Okay,
[04:00] (240.96s)
so it's doing that and then over here
[04:03] (243.04s)
claude code just finished. It already
[04:04] (244.64s)
started writing some code and it has a
[04:06] (246.72s)
to-do list. So I'll build an AI powered
[04:08] (248.48s)
stock tracking app with all the
[04:09] (249.60s)
features. I'm going to let it go into
[04:12] (252.08s)
YOLO mode here. So it can just keep
[04:13] (253.92s)
working and I will do the same with
[04:15] (255.20s)
Gemini CLI. So it looks like they're
[04:18] (258.16s)
both now installing the technology
[04:20] (260.88s)
behind our app. Uh, it looks like Claude
[04:22] (262.56s)
Code is now flipping,
[04:25] (265.52s)
which I think is their attempt at humor,
[04:27] (267.68s)
which I don't mind. You know, let's make
[04:29] (269.52s)
it so the AIS are fun to interact with.
[04:31] (271.20s)
So, I don't I don't hate on it. Little
[04:32] (272.80s)
tryhard, but I don't hate on it. I think
[04:34] (274.72s)
we're going to have two pretty cool
[04:35] (275.76s)
looking apps by the end of this. It's
[04:36] (276.88s)
going to be interesting to see. I'm most
[04:38] (278.64s)
interested in taste. I'm most interested
[04:40] (280.88s)
to see which one's going to have the
[04:42] (282.48s)
better tasteful app. Which one's going
[04:44] (284.64s)
to feel better to use? Which one's going
[04:46] (286.00s)
to look better to use? I'm I'm
[04:47] (287.76s)
interested to see which one's better at
[04:49] (289.28s)
that. Claude code now went into
[04:50] (290.80s)
implementing the stock search
[04:52] (292.00s)
functionality. So it already set up the
[04:53] (293.68s)
project structure in HTML foundation and
[04:55] (295.68s)
created the CSS styling. So now it's
[04:57] (297.92s)
implementing stock search. Both are
[04:59] (299.92s)
writing out code as we speak. Uh it
[05:02] (302.08s)
looks like Claude code has gone and just
[05:04] (304.48s)
done the most basic HTML app possible
[05:08] (308.16s)
while over on the right Gemini went with
[05:10] (310.24s)
Nex.js which I'd prefer the Nex.js so
[05:13] (313.28s)
that you can do more with it. So Gemini
[05:14] (314.96s)
might be in the lead a bit here so far.
[05:17] (317.44s)
While Claude Code is on straight up
[05:19] (319.12s)
HTML, JavaScript CSS, which a little old
[05:22] (322.32s)
school, uh, which is interesting. Okay,
[05:24] (324.40s)
so it looks like it's ready to run. All
[05:25] (325.92s)
we need to do is open index.html. So
[05:29] (329.68s)
let's do that. This reminds me of my
[05:31] (331.20s)
college days, just spilling out
[05:32] (332.72s)
JavaScript, HTML, and CSS files. All
[05:35] (335.04s)
right, let's run index.html. Let's see
[05:37] (337.12s)
what we got here. Okay, this actually
[05:39] (339.52s)
looks pretty nice. It's pretty clean.
[05:41] (341.36s)
What happens if I is this just going to
[05:42] (342.88s)
work? Search. What's going to happen?
[05:44] (344.40s)
Apple 289. Is that Is this fake data or
[05:46] (346.72s)
is this real? Are we looking at real
[05:48] (348.40s)
data here? Let me check my phone. I
[05:50] (350.32s)
mean, we're literally, I think, a minute
[05:52] (352.32s)
and a half into this. If this is real,
[05:54] (354.88s)
I'm going to be very impressed. Apple.
[05:57] (357.52s)
Nope. This is fake data. Uh, Apple is
[06:00] (360.00s)
currently at 2011. Okay, so it put in
[06:02] (362.16s)
some mock data. I think what it probably
[06:04] (364.00s)
needs to do next is hook up the API so
[06:06] (366.08s)
that it has some real data. But this is
[06:08] (368.00s)
what the a the UI is going to look like.
[06:10] (370.32s)
And honestly, overview, market
[06:12] (372.56s)
sentiment, this is pretty good from a
[06:15] (375.04s)
taste perspective. This is pretty nice.
[06:16] (376.72s)
Go back in and see how both of these are
[06:18] (378.16s)
working. Okay, looks like it's running.
[06:20] (380.16s)
Let's boot that up and see what we got
[06:21] (381.44s)
here. So, let's compare the two so far.
[06:23] (383.60s)
So, this is V1. Here we go. Uh, from a
[06:26] (386.24s)
looks perspective, I mean, it's
[06:28] (388.32s)
comparable. I think I lean a little bit
[06:30] (390.24s)
more towards the Cloud Code one. From a
[06:32] (392.00s)
looks perspective, let's see from a data
[06:33] (393.68s)
perspective, that's accurate. 201108. It
[06:36] (396.80s)
oneshot the functionality. It oneshot
[06:39] (399.20s)
the AI fun. So 2011, boom, that's the
[06:42] (402.40s)
actual price of Apple at the moment. So
[06:44] (404.16s)
this is accurate. That's accurate.
[06:46] (406.40s)
That's using real data. So oneshot real
[06:48] (408.96s)
data. It didn't implement the AI
[06:51] (411.44s)
functionality for the summary yet uh or
[06:54] (414.08s)
the market sentiment. But the price
[06:56] (416.40s)
perspective, it had in one shot, which
[06:58] (418.72s)
is pretty amazing. From a look and feel
[07:00] (420.88s)
perspective, I think it's quite a bit
[07:02] (422.40s)
weaker than Claude Code. Claude Code won
[07:05] (425.36s)
nicer colors, nicer feel. This one feels
[07:07] (427.52s)
pretty raw, like I'm in like a Bloomberg
[07:09] (429.44s)
terminal or something. V1 going to have
[07:11] (431.36s)
to give the Gemini CLI the first kind of
[07:13] (433.84s)
go through here. Mostly because the
[07:15] (435.60s)
functionality seems to work. It's just a
[07:17] (437.92s)
slight edge though because everything
[07:19] (439.44s)
else I'd say is weaker, but the fact
[07:20] (440.88s)
that it was able to get the price done
[07:22] (442.08s)
in V1 is pretty good. So in Claude Code,
[07:24] (444.40s)
I'm going to say that seems to work.
[07:25] (445.68s)
What are next steps? I'm going to assume
[07:27] (447.44s)
it's pulling in real data is the next
[07:28] (448.88s)
step. And then in Gemini CLI, I'm going
[07:30] (450.48s)
to say that works. The UI works, the
[07:32] (452.32s)
pricing works, but there's no AI data to
[07:34] (454.32s)
see what it does from there. Okay, so
[07:36] (456.08s)
Claude Code shows the updated to-do
[07:37] (457.84s)
list, which I love. I love that Claude
[07:39] (459.52s)
code works this way. It works in kind of
[07:41] (461.52s)
a to-do fashion where it builds and
[07:43] (463.92s)
maintains its own to-do list as it goes.
[07:46] (466.24s)
Uh, Gemini CLI doesn't seem to be
[07:48] (468.32s)
working kind of with to-do list. It kind
[07:49] (469.76s)
of just says, "Hey, this is what we're
[07:50] (470.80s)
going to work on," which I think I
[07:52] (472.40s)
leaned the user experience on claude
[07:54] (474.32s)
code a little bit more in that regard.
[07:56] (476.24s)
So immediate next steps it is going to
[07:58] (478.48s)
implement the real-time pricing on cloud
[08:00] (480.40s)
code give us pricing alerts have
[08:02] (482.24s)
portfolio tracking which tracks buy sell
[08:05] (485.28s)
transactions in and calculates P&L okay
[08:08] (488.64s)
and it's going to have news integration
[08:10] (490.48s)
so claude code's kind of going on here
[08:13] (493.20s)
and just adding a whole bunch of
[08:14] (494.80s)
functionality I didn't really ask for I
[08:17] (497.28s)
don't mind it to be quite honest with
[08:18] (498.80s)
you but it's something to take it's
[08:20] (500.96s)
something to note Gemini CLI on the
[08:23] (503.28s)
other end says now let's bring in the AI
[08:25] (505.68s)
feed features. So, it's sticking exactly
[08:27] (507.36s)
to what I'm asking for. So, it's going
[08:28] (508.96s)
to pull in the company summary and add
[08:30] (510.56s)
the AI features. So, they're both going
[08:32] (512.56s)
off. So far, Claude Code seems to have
[08:35] (515.44s)
more of a mind of its own, like it's a
[08:37] (517.60s)
human being, and it's adding other
[08:39] (519.04s)
features it finds interesting, which I
[08:41] (521.12s)
don't absolutely love, but to be honest
[08:42] (522.56s)
with you, I don't hate. It's very uh
[08:45] (525.44s)
Claude 3.7 for those who live through
[08:47] (527.60s)
the claude 3.7 days where it just go off
[08:49] (529.68s)
and do a billion things on its own. And
[08:51] (531.44s)
over on the right hand side, Gemini CLI
[08:53] (533.60s)
seems to be sticking to exactly what
[08:55] (535.20s)
we're doing. It's very important just as
[08:57] (537.28s)
a tip for each of these here. As you're
[08:59] (539.28s)
using Claude, as you're using Gemini
[09:01] (541.12s)
CLI, the higher the quality your first
[09:04] (544.00s)
prompt, the better they're both going to
[09:06] (546.08s)
work and the more time you'll save in
[09:08] (548.00s)
the back end. Right? If you spend a lot
[09:10] (550.08s)
more time in the front end building out
[09:11] (551.92s)
really good prompts like we did here
[09:13] (553.92s)
with this really big detailed prompt
[09:17] (557.12s)
from Claude, you're going to be able to
[09:18] (558.72s)
save a lot more time and you don't need
[09:20] (560.16s)
to handhold the AI every step of the
[09:22] (562.08s)
way. You just need to approve the next
[09:23] (563.52s)
steps, right? Like you've seen so far
[09:25] (565.28s)
each of these all I'm doing is saying
[09:26] (566.64s)
yes, continue, yes, continue, yes,
[09:28] (568.08s)
continue. But if you have kind of a
[09:29] (569.84s)
sloppy prompt up front, you're going to
[09:32] (572.00s)
have to do a lot more handholding and
[09:33] (573.68s)
fixing of errors. So, my little tip here
[09:35] (575.84s)
as they're both building and going at it
[09:37] (577.44s)
at the moment in our competition is make
[09:39] (579.76s)
sure before you start building any app,
[09:42] (582.08s)
run your ideas through an LLM. I find
[09:44] (584.64s)
Claude to be the most creative out of
[09:46] (586.48s)
all the LLMs and ask for a first prompt.
[09:49] (589.36s)
Then you'll get a really nice first
[09:50] (590.64s)
prompt like this that I showed you here.
[09:52] (592.64s)
That'll save you a lot more time as your
[09:54] (594.64s)
AIS are building. Okay, so both have
[09:56] (596.80s)
finished what they were doing. Let's see
[09:58] (598.08s)
where we're at. Okay, it looks improved.
[10:00] (600.56s)
Okay, so we have now a set alert button.
[10:02] (602.80s)
What does a set alert button do? We can
[10:04] (604.40s)
get wow we can get alerts when we go
[10:06] (606.88s)
above or below a certain price. I didn't
[10:08] (608.96s)
ask for that but that's still kind of
[10:10] (610.40s)
cool a thought of that. I kind of wish
[10:12] (612.40s)
it would focus on the core features
[10:13] (613.84s)
first before adding all these extra
[10:15] (615.68s)
features but that's fine. Whatever
[10:17] (617.76s)
company overview okay so it's getting
[10:19] (619.28s)
the real time news. It has like the
[10:21] (621.68s)
company overview. Uh is a multinational
[10:24] (624.16s)
technical company that has many okay so
[10:25] (625.76s)
it has that's the real news. That's real
[10:27] (627.92s)
Apple news. Uh market sentiment mixed
[10:30] (630.08s)
signals and price action market
[10:31] (631.36s)
indicators. has a balanced wait and see
[10:33] (633.20s)
approach. That's not really telling me
[10:35] (635.12s)
anything, but it looks good, which
[10:36] (636.96s)
counts for something. Does not look like
[10:38] (638.88s)
the price is real time yet, though. It
[10:40] (640.96s)
still looks like it is mock data. Yep.
[10:42] (642.88s)
So, it's switching every time I click
[10:44] (644.16s)
it. So, did not figure out that we're
[10:46] (646.16s)
two steps into cloud code. Did not
[10:47] (647.76s)
figure out the real data yet, but it
[10:49] (649.52s)
looks really nice. Let's go over to
[10:51] (651.92s)
Google CLI and see what they got over
[10:53] (653.84s)
here. So, I'm going to pull open the
[10:55] (655.36s)
Gemini CLI, see what we got. Okay, the
[10:57] (657.76s)
chart looks nicer now. uh about Apple.
[10:59] (659.92s)
Based on the information available,
[11:01] (661.04s)
Apple is a company that operates in the
[11:02] (662.88s)
following sectors. It's a multi. Okay,
[11:04] (664.48s)
so it's got the real-time news. Market
[11:06] (666.24s)
sentiment, uh it doesn't mean the
[11:08] (668.80s)
sentiment doesn't tell me as much.
[11:10] (670.32s)
Somewhat bullish. Doesn't really mean
[11:11] (671.92s)
anything. And the watch list, can we add
[11:13] (673.84s)
other stocks to the watch list? Yes.
[11:15] (675.92s)
Okay. So, it has the real-time stocks
[11:17] (677.68s)
from the watch list, which is pretty
[11:19] (679.68s)
cool. And the chart looks pretty nice as
[11:21] (681.68s)
well. I mean, out of all the things
[11:23] (683.60s)
requested, really the only thing this
[11:26] (686.24s)
Gemini CLI is missing is an explanation
[11:29] (689.36s)
on market sentiment. It says somewhat
[11:31] (691.60s)
bullish, but we asked for in our
[11:33] (693.20s)
original prompt some sort of
[11:34] (694.72s)
explanation. So, so far through two
[11:37] (697.12s)
prompts here. Claude Code's looking
[11:39] (699.28s)
really good and they're adding a bunch
[11:40] (700.72s)
of features. They haven't really done
[11:42] (702.40s)
the core of what we're looking for yet.
[11:44] (704.72s)
Gemini CLI, on the other hand, is going
[11:47] (707.04s)
just straight to what we asked for. It's
[11:48] (708.96s)
not really adding any of the window
[11:50] (710.56s)
dressings or like things that make apps
[11:53] (713.44s)
fun or interesting to use, right? Like
[11:55] (715.12s)
this is as plain Jane as it gets, but
[11:57] (717.12s)
it's doing what we asked for, which is
[11:58] (718.72s)
really helpful and nice. So, so far I'm
[12:00] (720.72s)
seeing this in this competition. Clawed
[12:03] (723.28s)
Code seems to be building better code.
[12:06] (726.48s)
Like the the app looks nicer. It looks
[12:10] (730.16s)
more modern. It looks more usable. It
[12:12] (732.80s)
seems like, and this is pretty expected
[12:14] (734.48s)
with Claude, just based on everything
[12:15] (735.60s)
we've seen over the years, it's able to
[12:18] (738.08s)
come up with context on its own and like
[12:20] (740.96s)
it's able to do more than you ask. The
[12:23] (743.20s)
number one complaint with Claude 3.7 was
[12:26] (746.08s)
that it just did too much, right? It
[12:27] (747.76s)
would you give it a command then do a
[12:28] (748.96s)
hundred other things. It still has
[12:30] (750.88s)
elements of that in here, but I actually
[12:33] (753.44s)
like it, right? Like it did extra things
[12:35] (755.04s)
like make this look pretty, right? Like
[12:36] (756.64s)
make the chart look nice, use nice
[12:39] (759.04s)
looking colors. I like that it kind of
[12:41] (761.52s)
has a brain and a taste of its own. The
[12:44] (764.64s)
Gemini CLI side, it's doing what we're
[12:47] (767.36s)
asking very well, but it's not doing it
[12:50] (770.08s)
in a way where like like this is just
[12:52] (772.56s)
not appealing. No one would use this app
[12:54] (774.56s)
because it's very ugly and simple.
[12:56] (776.64s)
There's no taste to it. It doesn't feel
[12:58] (778.80s)
like it has its own brain. It feels like
[13:00] (780.40s)
it's just doing what I'm telling it to
[13:02] (782.16s)
do, which don't get me wrong, it's not
[13:04] (784.48s)
necessarily a negative thing. Like from
[13:06] (786.80s)
a functionality perspective, it seems
[13:08] (788.56s)
like a step ahead of Claude. But like so
[13:11] (791.28s)
far, I think I prefer what Claude's
[13:13] (793.28s)
building a little bit more. And over on
[13:14] (794.80s)
the Claude side, let's test out the
[13:16] (796.24s)
real-time pricing to see if they finally
[13:18] (798.32s)
got that. So I'm going to refresh here.
[13:20] (800.40s)
I'm going to type in a stock Google.
[13:22] (802.80s)
It's going to load the stock data 178.
[13:25] (805.04s)
Let's see if that's the real time price.
[13:26] (806.64s)
Loading up the stock chart here. 178. So
[13:28] (808.88s)
that's accurate. It has the real-time
[13:30] (810.24s)
data. That's awesome. Can we add that to
[13:32] (812.16s)
watch list? Yep. And it has the real
[13:33] (813.76s)
chart there. Let's see if we can get
[13:35] (815.12s)
Apple to update. Yep, it has the real
[13:37] (817.04s)
Apple chart. Let's see if we can add
[13:38] (818.24s)
that to the watch. It's already in the
[13:39] (819.36s)
watches. All right, so it has the mock
[13:40] (820.64s)
data say. Oh, nope. It updated. There we
[13:42] (822.40s)
go. Okay, so now the functionality works
[13:44] (824.48s)
here. So now it's caught up to Gemini
[13:46] (826.24s)
CLI. It just took one extra command to
[13:48] (828.40s)
make it catch up from a functionality
[13:49] (829.92s)
perspective. But again, from a usability
[13:52] (832.24s)
perspective, I like this a lot more.
[13:54] (834.64s)
That looks great. What are next steps?
[13:57] (837.12s)
Oh, and while I'm doing that, it says
[13:58] (838.88s)
Gemini has completed phase one and phase
[14:01] (841.52s)
two. So, usability should be improved
[14:03] (843.20s)
and it should have better AI analysis.
[14:05] (845.20s)
Let's see what we got here. Okay, so
[14:07] (847.20s)
looks like we have our first bug. We
[14:08] (848.88s)
have our first We have our first bug.
[14:11] (851.36s)
Gemini messed up. It's not loading and
[14:13] (853.84s)
we got an error. So, let's give this
[14:15] (855.04s)
back to Gemini and see what we got here.
[14:17] (857.52s)
One of the great things about Claude
[14:19] (859.28s)
Code I've been seeing over the last few
[14:20] (860.80s)
weeks as I've been using it every day
[14:22] (862.16s)
for like 10 hours a day, it does not
[14:24] (864.64s)
mess up, right? Like out of the hours
[14:26] (866.80s)
and hours and hours I've spent in clawed
[14:28] (868.64s)
code, I haven't gone down a single error
[14:32] (872.16s)
rabbit hole, which is amazing. Back in
[14:34] (874.56s)
the old days, the wild west days of AI
[14:36] (876.64s)
coding in like September of last year, I
[14:40] (880.24s)
mean, you would go in rabbit hole after
[14:42] (882.16s)
rabbit hole after rabbit hole and I'd be
[14:44] (884.00s)
pounding my desk. That doesn't happen
[14:45] (885.36s)
with cloud code. It doesn't get errors.
[14:46] (886.88s)
So, that's amazing. Looks like Gemini
[14:48] (888.56s)
got some errors. So, let's just put that
[14:50] (890.88s)
in. Got an error and I paste it in. This
[14:53] (893.60s)
is what you do when working with your
[14:54] (894.96s)
AIS, by the way. You get errors, you
[14:56] (896.72s)
keep it simple. Hey, got an error. You
[14:58] (898.80s)
copy and paste the error from the
[15:00] (900.16s)
console, right? Right click console in
[15:01] (901.92s)
your browser and you're able to give the
[15:03] (903.76s)
error and it should be able to fix it
[15:05] (905.12s)
for you. So, as Gemini is fixing that,
[15:08] (908.88s)
we are at this point Claude has done
[15:11] (911.60s)
pretty much what we asked. The market
[15:14] (914.32s)
sentiment analysis isn't great. Let's
[15:16] (916.40s)
test the kind of taste a little bit more
[15:18] (918.48s)
cloud code and let's just see what it
[15:20] (920.64s)
wants to build next and kind of give it
[15:22] (922.08s)
a little bit of freedom to add a couple
[15:23] (923.60s)
features. Then we'll call it there. See
[15:25] (925.84s)
what it ends up with and then Gemini
[15:27] (927.36s)
will fix the errors and see where that
[15:28] (928.96s)
ends up with as well. So I asked Claude
[15:30] (930.88s)
Code what it wants to add next. It wants
[15:32] (932.72s)
to add portfolio tracking. So it tracks
[15:34] (934.80s)
your real-time portfolio, buy and sell
[15:37] (937.12s)
transactions, P&L, advanced technical
[15:39] (939.52s)
indicators, and user authentication. So
[15:42] (942.24s)
let's do this. Let's take a couple of
[15:43] (943.36s)
these and have it just build it out. Can
[15:45] (945.84s)
you please add the technical indicators
[15:50] (950.72s)
and news feed? So I took a couple of
[15:52] (952.88s)
things it came up with. I said just feel
[15:54] (954.64s)
free to add that. We'll leave this as
[15:56] (956.40s)
last steps of claude code. And then on
[15:58] (958.08s)
the Gemini side, we will try to see what
[16:00] (960.72s)
is going on here. Okay, so it says I've
[16:02] (962.80s)
corrected the error. Let's pull it open.
[16:04] (964.40s)
Uh does not appear like it corrected the
[16:06] (966.24s)
error. Looks like we're still getting
[16:07] (967.92s)
errors here. So we're in our first
[16:09] (969.52s)
rabbit hole here. Let's do this. Let's
[16:12] (972.48s)
take this. Let's paste this error in. To
[16:14] (974.80s)
Claude's credit, it did not get any
[16:16] (976.40s)
errors, which is expected because I've
[16:17] (977.92s)
been using Claude Code for a while and I
[16:19] (979.84s)
never get errors with it. More errors. I
[16:22] (982.72s)
paste in the errors to Gemini. I say,
[16:24] (984.56s)
"Go at it. Let's fix it." Claude Code's
[16:26] (986.96s)
building out its other functionality
[16:28] (988.40s)
that it kind of just came up with on the
[16:30] (990.00s)
fly. We already met the requirements on
[16:32] (992.08s)
Claude Code. Now, it's just building out
[16:33] (993.84s)
extra stuff. It's just styling on us
[16:35] (995.52s)
now, while Gemini is trying to fix.
[16:38] (998.08s)
While Gemini did get to the basic
[16:40] (1000.32s)
functionality quicker, it wasn't as
[16:42] (1002.72s)
pleasant of an experience. Now we're
[16:44] (1004.24s)
running into errors, unfortunately. All
[16:45] (1005.84s)
right, Gemini says it fixed. Let's see
[16:47] (1007.68s)
what we got here. Let's refresh. Oh, it
[16:50] (1010.08s)
almost showed something. Now we're
[16:51] (1011.28s)
getting more errors. Okay, uncaught
[16:52] (1012.80s)
reference dates is not defined. All
[16:54] (1014.88s)
right, we're getting a little fed up
[16:56] (1016.88s)
with Gemini here. Claude code just
[16:58] (1018.56s)
doesn't do error rabbit holes. Funny
[17:00] (1020.24s)
because a lot of people say Gemini 2.5
[17:02] (1022.48s)
Pro is the best coding model out there.
[17:04] (1024.40s)
It seems to be very highly intelligent
[17:06] (1026.88s)
like as we kind of saw it was able to
[17:08] (1028.88s)
get to the final results quick but it it
[17:11] (1031.36s)
doesn't seem it seems to be kind of un
[17:13] (1033.84s)
nonironed out right like claude code's
[17:16] (1036.72s)
very ironed out dependable reliable
[17:19] (1039.36s)
which I think is very important Gemini
[17:21] (1041.92s)
is not very reliable if you use Gemini
[17:23] (1043.76s)
and like cursor before you know it
[17:25] (1045.20s)
messes up with tool calls and things
[17:26] (1046.88s)
like that you're right my apologies I
[17:28] (1048.88s)
introduced a reference error at my last
[17:30] (1050.64s)
fix okay so this is your fullon rabbit
[17:33] (1053.36s)
hole here And over here, Cloud Code
[17:35] (1055.12s)
finished with its bonus features. It
[17:37] (1057.20s)
finished with its bonus features. So,
[17:38] (1058.72s)
let's see. So, I'm going to refresh
[17:40] (1060.24s)
here. Let's put in Tesla. Shout out
[17:42] (1062.88s)
Tesla. I got a Cybert truck and it's
[17:45] (1065.20s)
awesome. Okay, so we got Tesla. Let's
[17:48] (1068.24s)
see what we got here. We got technical
[17:49] (1069.52s)
indicators. So, it has the RSI, the
[17:51] (1071.12s)
MACD, the Bowlinger bands. That's pretty
[17:53] (1073.52s)
good. Latest news. Okay, so this is
[17:55] (1075.60s)
fake. This is fake data. I'm sure this
[17:57] (1077.92s)
is fake data because it didn't actually
[18:00] (1080.24s)
implement the API yet. I'm sure I have
[18:02] (1082.88s)
to put in like an API key, but it has
[18:05] (1085.12s)
the features built out. And just like
[18:06] (1086.64s)
the real- time pricing, probably all I
[18:08] (1088.80s)
need to do to get that to work is put in
[18:10] (1090.40s)
an API key, right? So, let's look at
[18:12] (1092.48s)
what Claude Code has for us so far. It
[18:14] (1094.88s)
has, I think you'd probably agree
[18:16] (1096.40s)
watching this video, the best UX, the
[18:18] (1098.32s)
best UI. This is nice. This is an app I
[18:20] (1100.80s)
would use every day. The watch list
[18:22] (1102.48s)
looks great. I mean, it has little taste
[18:24] (1104.32s)
features like the hover state, things
[18:26] (1106.16s)
like that. Like, you hover over things,
[18:27] (1107.44s)
it looks nice. I click it, it loads it
[18:29] (1109.52s)
in, right? Boom. It updates the price.
[18:31] (1111.52s)
Nice. It has the I I like that it has
[18:33] (1113.60s)
like the color coding. The colors are
[18:35] (1115.60s)
nice. I like that. It has the overview,
[18:38] (1118.40s)
which is good. It has technical
[18:40] (1120.16s)
indicators, which is cool. That's just
[18:41] (1121.68s)
bonus features it came up with. And the
[18:43] (1123.60s)
latest news and I'm sure I just put in
[18:44] (1124.96s)
an API key and it'll get the latest
[18:46] (1126.48s)
news. So, this is good. This is clearly
[18:49] (1129.12s)
like Claude understands human beings and
[18:51] (1131.76s)
made like a very human specific app. Was
[18:54] (1134.56s)
it as fast as Gemini? It took a little
[18:57] (1137.12s)
bit while to get to that baseline usage,
[18:59] (1139.84s)
right? But it eventually got there.
[19:02] (1142.08s)
Gemini. Let's see what we got going on
[19:03] (1143.76s)
in Gemini. So, it finished. Let's see
[19:05] (1145.76s)
what we got here. More failed to fetch
[19:07] (1147.84s)
data. More failed to fetch. Let's see
[19:09] (1149.92s)
what happens if I put in a stock here.
[19:11] (1151.76s)
Nope. Failed to fetch. So, we've run
[19:14] (1154.00s)
into a full-on rabbit hole of errors
[19:16] (1156.56s)
with Gemini. And to be honest, this is
[19:19] (1159.28s)
kind of this is kind of predictable
[19:21] (1161.12s)
using both models. I've used both models
[19:23] (1163.20s)
a ton. There was a lot of hype when
[19:25] (1165.36s)
Gemini 2.5 Pro came out. People were
[19:28] (1168.08s)
saying, "Oh, it's the smartest ever."
[19:29] (1169.60s)
You know what? It is pretty damn smart.
[19:31] (1171.44s)
But to be honest with you, it's just
[19:33] (1173.36s)
kind of like unpredictable and
[19:35] (1175.52s)
unfocused, right? It runs into errors.
[19:38] (1178.64s)
It doesn't know when to use tools,
[19:40] (1180.00s)
right? It'll do things that blow you
[19:42] (1182.64s)
away. Like it kind of oneshot the basic
[19:44] (1184.96s)
functionality of the app, but the moment
[19:46] (1186.72s)
we started trying to change anything, it
[19:48] (1188.48s)
tripped over itself. So, let's let's do
[19:50] (1190.72s)
this. Let's score these and determine
[19:53] (1193.76s)
the winner. All right. So, let's rate
[19:55] (1195.60s)
these. We'll rate them all one to 10 in
[19:58] (1198.00s)
each one of these categories. Let's do
[20:00] (1200.00s)
this. Claude Code speed. Claude Code was
[20:02] (1202.08s)
a little slower, so it didn't execute
[20:03] (1203.92s)
quite as fast, but it was a good speed.
[20:06] (1206.40s)
If we're going to go 1 to 10, I'll
[20:07] (1207.76s)
probably give it a seven. With Gemini
[20:09] (1209.92s)
speed, I'll probably give it an eight.
[20:11] (1211.44s)
Slightly faster. Accuracy. I define
[20:13] (1213.76s)
accuracy as does it get it done without
[20:16] (1216.48s)
many errors or bugs or messing up at
[20:18] (1218.48s)
all. Claude Code from an accuracy
[20:21] (1221.20s)
perspective had zero bugs. It did do a
[20:24] (1224.32s)
little couple extra things, but to be
[20:26] (1226.00s)
honest with you, I don't mind it. Claude
[20:28] (1228.24s)
4 clearly is a little bit more focused
[20:30] (1230.48s)
than 37. Like if it was 37 running this,
[20:33] (1233.04s)
then we'd have a 100 features that were
[20:34] (1234.96s)
nothing that we requested. So, I'm going
[20:37] (1237.12s)
to since it had no bugs, I'll give this
[20:38] (1238.80s)
a nine. The accuracy of Gemini CLI, I
[20:42] (1242.08s)
mean, it had all those bugs at the end.
[20:44] (1244.48s)
To its credit, it was able to get the
[20:46] (1246.24s)
core functionality done much faster, but
[20:48] (1248.48s)
the bugs, I mean, if you're a beginning
[20:50] (1250.56s)
programmer, you're new to AI vibe
[20:52] (1252.08s)
coding, that would have probably had you
[20:53] (1253.60s)
frustrated wanting to quit. So, I'm
[20:55] (1255.36s)
going to have to give that a six. From a
[20:57] (1257.20s)
taste perspective, Claude Code had
[20:59] (1259.04s)
fantastic taste. It was an easy to use
[21:01] (1261.44s)
app. It looked beautiful. It was very
[21:03] (1263.60s)
pleasant to use. Honestly, I'm going to
[21:05] (1265.44s)
have to give it a 10 from a taste
[21:07] (1267.20s)
perspective. Taste perspective, a Gemini
[21:09] (1269.84s)
CLI, it was ugly. Uh, it was kind of
[21:12] (1272.80s)
gross. I didn't feel motivated to use
[21:15] (1275.20s)
it. And like when you're building apps
[21:17] (1277.04s)
out, taste is so important. The emotions
[21:18] (1278.96s)
you feel when you're using the app are
[21:21] (1281.20s)
so important. So, I'm going to have to
[21:22] (1282.72s)
give that a five. Ease of use. Just how
[21:25] (1285.60s)
easy it was to use. I mean, honestly,
[21:27] (1287.60s)
both are pretty easy to use. If
[21:29] (1289.68s)
anything, Gemini CLI might have been a
[21:32] (1292.00s)
little easier because I had to ask
[21:33] (1293.52s)
Claude Code to kind of do some of the
[21:35] (1295.84s)
little tiny things like implementing the
[21:37] (1297.52s)
API keys or things like that. Gemini
[21:39] (1299.52s)
went out of its way and just implemented
[21:41] (1301.36s)
the APIs and got that done for me.
[21:44] (1304.32s)
Clawed Code, it didn't really say, "Hey,
[21:46] (1306.40s)
you need to put your API keys in here or
[21:48] (1308.32s)
do anything like that." It kind of just
[21:49] (1309.44s)
assumed I would figure it out. So, from
[21:51] (1311.36s)
an ECU's perspective, Gemini is a little
[21:53] (1313.68s)
easier. I'd say probably an eight for
[21:55] (1315.76s)
Gemini CLI and a seven for Claude Code.
[21:59] (1319.04s)
So, that brings us to the totals. Claude
[22:01] (1321.28s)
Code, that's a 33 total for Claude Code
[22:04] (1324.00s)
and a 27 for Gemini CLI. Now, I'll say
[22:07] (1327.44s)
this. This backs up my experiences with
[22:10] (1330.64s)
both. Gemini CLI is an amazing tool and
[22:14] (1334.40s)
if Claude Code did not exist, Gemini CLI
[22:17] (1337.68s)
would be blowing minds, right? But
[22:19] (1339.68s)
Claude Code just seems a little bit more
[22:21] (1341.92s)
humanlike, more lifelike, a little bit
[22:24] (1344.24s)
more intelligent. It just seems to
[22:26] (1346.56s)
operate like a human, right? Gemini CLI,
[22:29] (1349.52s)
it does pretty incredible things, but
[22:31] (1351.44s)
every once in a while it just goes off
[22:32] (1352.72s)
the rails a little bit and messes up,
[22:34] (1354.32s)
which was what my experience was with
[22:36] (1356.72s)
with Gemini 2.5 Pro. That was my
[22:39] (1359.12s)
experience. So, it matches up with that.
[22:40] (1360.96s)
Here's what it comes down to. If you're
[22:42] (1362.96s)
tight on budget, Gemini CLI is pretty
[22:45] (1365.68s)
amazing for its cost, which is uh free.
[22:48] (1368.00s)
Free for the first thousand requests.
[22:50] (1370.08s)
So, if you don't have money to spend,
[22:51] (1371.84s)
here's what I would do. I'd go out, I'd
[22:53] (1373.76s)
use Gemini CLI. I'd build your first
[22:56] (1376.48s)
app. I'd put it online. I'd start making
[22:58] (1378.72s)
money. If you want tips on building apps
[23:00] (1380.40s)
and making money, I have a link for a
[23:02] (1382.08s)
video down below you can watch and how I
[23:03] (1383.76s)
monetize my app. That's down below. You
[23:06] (1386.56s)
start making money by with your app on
[23:08] (1388.56s)
Gemini CLI and then you take that money
[23:10] (1390.96s)
and then you get on the $200 a month
[23:12] (1392.72s)
claude code. Right? If you have extra
[23:14] (1394.56s)
money, you invest in yourself and you go
[23:16] (1396.00s)
straight to Claude Code. But what Google
[23:18] (1398.16s)
did with Gemini CLI is amazing making it
[23:20] (1400.64s)
free, right? The more competition, the
[23:23] (1403.12s)
better it is for all of us. Claude Code
[23:25] (1405.44s)
though does appear smarter. does appear
[23:27] (1407.68s)
more focused and just has better taste
[23:29] (1409.92s)
overall. The apps that come out of
[23:31] (1411.44s)
Claude Code just feel like better apps,
[23:34] (1414.08s)
more usable apps, and this does back up
[23:36] (1416.56s)
my personal experiences using both for
[23:38] (1418.80s)
hours and hours. If this was helpful,
[23:40] (1420.64s)
make sure to subscribe, turn on
[23:42] (1422.24s)
notifications. All I do is make amazing
[23:44] (1424.48s)
videos about AI. Leave a like as well.
[23:46] (1426.64s)
Also, every Thursday I do a live stream
[23:48] (1428.24s)
and I build apps live, so watch out for
[23:50] (1430.08s)
that as well. See you in the next video.