1
00:00:06,400 --> 00:00:12,400
Hey guy, you look like you could use an

2
00:00:09,440 --> 00:00:18,160
upgrade. I can get YOU NVIDIA'S LATEST $1,000 GPU FOR JUST 20 BUCKS.

3
00:00:15,519 --> 00:00:23,840
>> REALLY? What's the catch? >> THERE'S NO CATCH, FRIEND. With GeForce

4
00:00:21,039 --> 00:00:27,279
Now, you don't need to own a GPU. You can use ONE OF NVIDIA'S. THEY'VE GOT

5
00:00:26,400 --> 00:00:32,399
LOTS. >> BUT cloud gaming? Come on. Everyone

6
00:00:30,320 --> 00:00:36,880
knows the input latency sucks and compressed video looks super low res.

7
00:00:34,719 --> 00:00:41,760
>> No way, buddy. With NVIDIA's new Blackwell 5080 Super Pods, those

8
00:00:39,440 --> 00:00:46,239
problems ARE IN [screaming] THE PAST. IF THAT'S TRUE, it sounds like a pretty

9
00:00:44,000 --> 00:00:49,600
good deal. I mean, it would take 4 years to spend the same money on my

10
00:00:47,680 --> 00:00:55,520
subscription. And the service includes the whole gaming PC, not just that GPU.

11
00:00:52,960 --> 00:00:58,960
Besides, I already rent my games FROM MICROSOFT ANYWAY.

12
00:00:56,879 --> 00:01:04,159
>> YEAH. And the best part is when new generations of GPUs COME OUT, YOU'LL

13
00:01:01,039 --> 00:01:05,439
AUTOMATICALLY HAVE THEM. WAIT, NVIDIA

14
00:01:04,159 --> 00:01:10,799
has raised the prices with new generations before. Can I really count on that value? And what if my internet

15
00:01:08,799 --> 00:01:13,280
goes down? Then I can't game at all. Hey, where is the other guy that's

16
00:01:12,159 --> 00:01:18,720
supposed to bring up all the counter points? >> Oh, sorry. Sorry. Oh, I'm sorry. I was

17
00:01:17,840 --> 00:01:23,200
with our sponsor. >> Oh, you. Delete Me. Data brokers are

18
00:01:21,759 --> 00:01:28,320
getting their grubby little hands all over your personal data, and they're selling it to the highest bidder. Learn

19
00:01:26,240 --> 00:01:32,320
how delete me can protect your online identity and get that data removed at

20
00:01:30,080 --> 00:01:37,960
joindeme.com/lt and save 20% off a plan that fits your

21
00:01:34,960 --> 00:01:37,960
needs.

22
00:01:39,561 --> 00:01:47,200
[music] Sponsor sponsor

23
00:01:44,560 --> 00:01:52,320
sponsor. While the sales pitch for cloud gaming

24
00:01:49,920 --> 00:01:56,960
is affordability, performance, and convenience, the reason that NVIDIA

25
00:01:54,720 --> 00:02:01,840
actually invested untold amounts of money in cloud gaming is, of course,

26
00:01:59,119 --> 00:02:06,560
economics. See, the old deal was you give NVIDIA a thousand of your dollars.

27
00:02:04,240 --> 00:02:11,520
They buy a nice leather jacket, and you get a GPU that you can keep for a very,

28
00:02:09,360 --> 00:02:18,480
very long time. But then they thought, "Hold on a second. What if we lease that

29
00:02:14,160 --> 00:02:20,560
same GPU to gamers for just $20 a month?

30
00:02:18,480 --> 00:02:24,480
Then we could have leather jackets forever. It's the same you'll own

31
00:02:22,720 --> 00:02:29,360
nothing and be happy strategy that has led to inflated prices in other markets

32
00:02:26,480 --> 00:02:33,360
because for the owner class, renting is more profitable than selling over time.

33
00:02:31,680 --> 00:02:37,680
But I don't know, maybe I'm being too harsh. Game journalists are raving about

34
00:02:35,599 --> 00:02:42,879
the iniccernable experience of the latest GeForce Now compared to local

35
00:02:40,959 --> 00:02:47,440
hardware. And it is certainly impressive what NVIDIA is doing. Taking input from

36
00:02:45,440 --> 00:02:51,280
your local peripherals, then running the games in a virtual machine in their data

37
00:02:49,680 --> 00:02:55,519
center, compressing that gameplay footage, and shooting it over the

38
00:02:53,040 --> 00:03:01,200
internet to your screen at up to 5K resolution or up to 360 frames per

39
00:02:58,800 --> 00:03:04,720
second at 1080p. This support for higher resolutions and higher frame rates is a

40
00:03:03,120 --> 00:03:10,080
big improvement since we last looked at the service 4 years ago. And there's

41
00:03:07,599 --> 00:03:14,319
more. There's a lot more titles you can play if you own them on Steam or other

42
00:03:12,159 --> 00:03:19,040
platforms, including [music] Xbox Game Pass now. So,

43
00:03:16,959 --> 00:03:22,640
I'm going to go into this with an open mind. It's a problem with having an open

44
00:03:20,720 --> 00:03:26,480
mind, though. Sometimes your brain falls out.

45
00:03:24,480 --> 00:03:31,680
For our test machine, we targeted the type of system that you'd probably build

46
00:03:28,400 --> 00:03:34,799
around an RTX 5080, except we're running

47
00:03:31,680 --> 00:03:37,040
an older, more affordable GPU, the RTX

48
00:03:34,799 --> 00:03:41,680
3060. We're going to put the 5080 back in later to compare the experience of

49
00:03:39,120 --> 00:03:46,400
GeForce Now to running our games locally on the same hardware. Let's kick things

50
00:03:44,000 --> 00:03:50,959
off with Indiana Jones and the Great Circle. Now, our original plan was to

51
00:03:48,959 --> 00:03:56,000
run our tests on our CPU's integrated graphics, and you can do that. You don't

52
00:03:52,959 --> 00:03:58,640
need a GPU, but in our case, our onboard

53
00:03:56,000 --> 00:04:02,239
graphics is limited to HDMI 2.0, O, meaning we would have been limited to

54
00:03:59,732 --> 00:04:08,247
[music] 60 Hz 4K and couldn't have experienced the full power of the latest

55
00:04:05,200 --> 00:04:10,959
GeForce Now, which can do up to 4K 120

56
00:04:08,247 --> 00:04:15,360
[music] frames per second. Uh, well, it's already not completely bugree.

57
00:04:12,879 --> 00:04:19,759
Second attempt. Oh, this is fun. If you interact with the Steam interface before

58
00:04:17,440 --> 00:04:27,919
the VM finishes launching and then opens up the game, you can actually see system

59
00:04:23,280 --> 00:04:29,919
information. Oh, wow. Eight physical CPU

60
00:04:27,919 --> 00:04:33,759
cores gives us a [music] clock speed, but that doesn't tell us how much it

61
00:04:31,360 --> 00:04:41,040
boosts to. Yes, sir. Indiana Jones on the Great Circle gives us an RTX 5080 as

62
00:04:36,720 --> 00:04:44,240
advertised, except what? With 48 gigs of

63
00:04:41,040 --> 00:04:47,040
VRAM. Clearly, this is not a 5080. Then

64
00:04:44,240 --> 00:04:52,560
it's [music] probably an RTX Pro 5000 Blackwell, which is obviously a much

65
00:04:49,520 --> 00:04:54,160
more expensive GPU, except for NVIDIA.

66
00:04:52,560 --> 00:04:59,280
You know, they make them, so they can pay however much they feel like for them, I guess. detects our monitor

67
00:04:56,880 --> 00:05:03,120
refresh rate at 120 Hz. [music] Windows 11. Yeah, I guess they could if they

68
00:05:01,600 --> 00:05:07,600
want the best game support. Yeah, that makes sense. And the OS itself doesn't

69
00:05:05,919 --> 00:05:11,600
need to handle any kind of specific virtualization stuff. It just needs to

70
00:05:08,960 --> 00:05:16,240
run in a virtual environment. 57 gigs of RAM. Okay. No, that actually makes sense

71
00:05:14,720 --> 00:05:19,600
because they would reserve some of the system memory for the hypervisor and

72
00:05:17,759 --> 00:05:23,280
then whatever's left, they would divide up evenly for all of the VMs that are

73
00:05:22,000 --> 00:05:28,160
running. I believe is going to work perfectly this time.

74
00:05:25,680 --> 00:05:32,639
Hey. Oh, I can feel the latency already. In fairness though, it's a lot easier to

75
00:05:30,560 --> 00:05:37,440
feel on a mouse than it is with a camera. And in fairness, guys, this is

76
00:05:35,840 --> 00:05:41,199
the first game that was designed from the ground up to require ray tracing

77
00:05:39,840 --> 00:05:46,000
hardware for its lighting. And seeing this run on a couple generation old

78
00:05:43,520 --> 00:05:52,479
mainstream card at 120 frames per second, 4K, and wow, I mean, looks

79
00:05:49,680 --> 00:05:57,039
pretty damn good, doesn't it? Sharp even. I do notice some blockiness and a

80
00:05:55,440 --> 00:06:02,240
little bit of compression, especially when in motion, but I wouldn't say that

81
00:05:59,039 --> 00:06:04,319
it feels like playing at 120 frames per

82
00:06:02,240 --> 00:06:07,199
second locally, but it's [music] very usable. I think the most impressive

83
00:06:06,000 --> 00:06:13,360
thing about the image quality, though, is that we're not even at the highest setting. We're running at 60 megabit and

84
00:06:11,120 --> 00:06:17,360
we're running at 120 FPS. And my understanding is it can go as high as

85
00:06:14,960 --> 00:06:22,240
100 megabit if you're in the cinematic mode, whereas this is balanced. And I'd

86
00:06:20,639 --> 00:06:27,759
probably prefer to play in balanced. looks good enough and the latency is uh

87
00:06:25,360 --> 00:06:32,560
better. If you didn't tell me anything, I'd say I'm probably running at around

88
00:06:29,680 --> 00:06:37,280
75 80 FPS local because that additional latency feels like you a little bit

89
00:06:35,360 --> 00:06:41,600
lower frame rate. It's improved a lot since the last time I used it and I'm on

90
00:06:39,680 --> 00:06:45,840
a mouse. I should try with a controller where it's even harder to tell how much

91
00:06:43,520 --> 00:06:49,759
latency there is. Do we have one? I wouldn't notice. Damn impressive.

92
00:06:48,560 --> 00:06:53,440
>> How does this compare to your last impressions of GeForce Now? Even 4 years

93
00:06:51,680 --> 00:06:58,800
ago, it was pretty good. Now, it's even better. [music] And with a controller,

94
00:06:55,520 --> 00:07:00,800
from a reasonable distance,

95
00:06:58,800 --> 00:07:04,479
I don't think I would notice that this is not running on local hardware. I

96
00:07:02,639 --> 00:07:07,280
wouldn't think it's 120 frames per second, but I would think it's [music]

97
00:07:05,759 --> 00:07:12,479
running right on this computer right here. Our next game's going to be Borderlands 4. But before we do that, I

98
00:07:10,960 --> 00:07:16,400
want to experience cinematic mode for myself. And while one of the features of

99
00:07:14,479 --> 00:07:20,400
the ultimate tier is that you can go into the game settings menu and alter

100
00:07:18,560 --> 00:07:25,199
things, there are certain elements that you have to configure before you launch

101
00:07:22,160 --> 00:07:27,199
the VM. And this is one of them. Never

102
00:07:25,199 --> 00:07:32,000
mind. We got stuck on the dreaded loading user profile bug and we were not

103
00:07:29,440 --> 00:07:36,160
able to launch cinematic Indiana Jones. Seriously, we tried like many, many

104
00:07:34,080 --> 00:07:39,120
times. Borderlands it is. I think it's important to acknowledge these sorts of

105
00:07:37,520 --> 00:07:45,520
bugs as part of evaluating the experience. Now, this this I like. This

106
00:07:43,440 --> 00:07:51,199
is an art style that lends itself extremely well to compression. And don't

107
00:07:48,479 --> 00:07:55,440
mistake any of my lack of skill for the game not running correctly. This is kind

108
00:07:53,199 --> 00:07:59,280
of freaking awesome. Anything with a first person view, I'm going to be able

109
00:07:57,520 --> 00:08:06,240
to tell that the mouse movements have an extra 25 milliseconds total. Extra, I'm

110
00:08:04,319 --> 00:08:10,319
going to say 30. And so it should. I mean, this is one of the first games

111
00:08:07,919 --> 00:08:14,879
that was validated as being optimized for their new Blackwell servers, which

112
00:08:12,400 --> 00:08:20,080
means what exactly? I mean, obviously, a 5080 can run anything that a 4080 can.

113
00:08:17,520 --> 00:08:24,639
So, we think this list is likely less about working on the new servers and

114
00:08:22,240 --> 00:08:29,680
more about NVIDIA identifying which games they feel require a 5080 for an

115
00:08:27,520 --> 00:08:35,279
optimal experience. To test this, we fired up control on a 5080 enabled

116
00:08:32,560 --> 00:08:40,320
account and a 5080 enabled location and noticed that we were dropped down to an

117
00:08:37,440 --> 00:08:43,919
older 4080 server because that was all it really needed. On the subject of

118
00:08:41,839 --> 00:08:48,800
control, I mentioned before that NVIDIA gives us control over the in-game

119
00:08:45,600 --> 00:08:52,000
settings. So, what I wanted to do is set

120
00:08:48,800 --> 00:08:54,959
this to exactly Gearbox's recommended

121
00:08:52,000 --> 00:08:58,880
settings for an RTX 5080 and see if our VM is bottlenecked in some way where it

122
00:08:57,040 --> 00:09:05,440
doesn't give us an optimal experience. Here we go. And wow, that looks way

123
00:09:02,640 --> 00:09:09,920
better. Also, this is still running at over 200 frames per second in game,

124
00:09:07,680 --> 00:09:14,720
which is more than enough to stream it at 120 FPS. With that said, I can

125
00:09:12,800 --> 00:09:18,640
definitely see stuff that doesn't look that great with compression artifacts.

126
00:09:17,040 --> 00:09:22,720
Like, see what's happening as I go through these god rays.

127
00:09:20,800 --> 00:09:26,880
That's a yuck yuck yuck. I'm not even getting the benefit of the extra level

128
00:09:24,480 --> 00:09:31,760
of detail and stuff like this grass. It just ends up looking kind of gross when

129
00:09:29,120 --> 00:09:35,839
I'm in motion. Still, if I'm paying for Ultimate, I would appreciate the ability

130
00:09:33,839 --> 00:09:39,760
to be able to set this stuff. And you can even configure GeForce Now to

131
00:09:38,160 --> 00:09:45,200
remember your settings. So the next time you launch it, if there's something you like better, you can totally do that. It

132
00:09:43,200 --> 00:09:48,959
feels like the latency might be a touch higher here. So it's possible that

133
00:09:46,800 --> 00:09:53,279
NVIDIA had things turned down to ensure that the GPU doesn't get overburdened by

134
00:09:50,880 --> 00:09:58,800
the encoding and then end up running a little bit slower. Let's drop it down to

135
00:09:54,959 --> 00:10:00,399
1440p and see what happens. Oh boy.

136
00:09:58,800 --> 00:10:05,040
Oh, I still definitely suck at video games.

137
00:10:02,000 --> 00:10:06,959
Ooh, distant objects definitely fuzzier.

138
00:10:05,040 --> 00:10:10,959
If you have a 4K monitor, you're going to want to use it. And I gotta say, I

139
00:10:08,959 --> 00:10:15,360
don't think the latency is really much better. Like maybe we gain back another

140
00:10:13,040 --> 00:10:19,360
10 milliseconds, five. It's still very noticeable in a firstperson application

141
00:10:17,279 --> 00:10:22,720
like this. I would want to play at 4K if I could. Garbage in, garbage out when it

142
00:10:21,279 --> 00:10:28,000
comes to compression, man. So, the cleaner an image you can put into this thing, the better it's going to look

143
00:10:26,000 --> 00:10:31,200
once you transmit it over the internet. Time to find out if I'm right about the

144
00:10:29,360 --> 00:10:36,880
latency with a little device called an LDAT. This is a creation from NVIDIA. It

145
00:10:34,560 --> 00:10:41,519
basically automates mouse clicks and then it checks for the muzzle flash,

146
00:10:39,279 --> 00:10:47,680
measuring the time between the click and the weapon firing. So, this is adding

147
00:10:43,440 --> 00:10:49,279
more like 50 milliseconds plus. Damn. Of

148
00:10:47,680 --> 00:10:53,440
course, we won't know the exact amount until we throw a 5080 into this machine

149
00:10:51,680 --> 00:10:58,720
and actually run that. And we can't do that until we do one more game. The

150
00:10:55,279 --> 00:11:01,440
spiciest game yet. See, PC gamers said

151
00:10:58,720 --> 00:11:07,519
that they could not tell the difference between local hardware and GeForce Now

152
00:11:04,320 --> 00:11:09,760
while playing Apex Legends at 1440p on

153
00:11:07,519 --> 00:11:14,240
medium to low settings. That was before it was rolled out publicly and it was in

154
00:11:12,000 --> 00:11:18,160
a different geographical location. So, we're not sure that we can 100%

155
00:11:16,240 --> 00:11:23,279
replicate their test conditions, but what we can do is play the game and see

156
00:11:20,880 --> 00:11:28,240
if we agree. to try to replicate their experience. I've set everything to low

157
00:11:25,760 --> 00:11:31,760
to medium settings, whatever exactly it is that that means. It feels way better

158
00:11:30,000 --> 00:11:38,640
than Borderlands. Like way better than Borderlands, dude.

159
00:11:36,000 --> 00:11:42,240
I would guess that this is a solid like 25 milliseconds better than what I just

160
00:11:40,560 --> 00:11:48,399
played. Okay, I need the LDAT back. A very big

161
00:11:44,800 --> 00:11:50,800
difference. However, and this is kind of

162
00:11:48,399 --> 00:11:55,440
hard to know how to explain to you guys, the difference we're measuring right now

163
00:11:52,959 --> 00:11:59,279
could be because this game runs way better in GeForce Now. It could also

164
00:11:57,200 --> 00:12:03,200
just be runto- run variance because when they tested Borderlands at the same

165
00:12:01,200 --> 00:12:06,959
settings that I was using just now, when they were prepping for this video, they

166
00:12:04,880 --> 00:12:10,959
got more like 50 milliseconds, like way better than what I got. That's the thing

167
00:12:08,720 --> 00:12:16,160
about cloud gaming. The experience can be great and other times it can be not

168
00:12:12,800 --> 00:12:17,839
as great. Just ask Luke. He knows. Dude,

169
00:12:16,160 --> 00:12:23,680
this is way better. Like, we're averaging freaking 45 milliseconds. This

170
00:12:20,399 --> 00:12:26,160
is very playable. And we have seen only

171
00:12:23,680 --> 00:12:30,639
a fraction of its full power. I want to try the 360 Hz competitive mode. In

172
00:12:29,279 --> 00:12:36,800
order to get the most out of competitive mode though, we need to reconfigure our display. This is a dual mode display

173
00:12:33,839 --> 00:12:41,279
from LG. So, it was running at 4K 120 Hz before, and now we've reconfigured it to

174
00:12:39,279 --> 00:12:49,519
run at 1080p 360 Hz. Woo. The image quality is uh not

175
00:12:46,480 --> 00:12:51,920
as good. Like that's pretty rough. But

176
00:12:49,519 --> 00:12:56,880
the latency, if you told me this was local and just, you know, not that great

177
00:12:54,399 --> 00:13:00,480
of a computer, dude, I think I believe you. And not just cuz I'm on a

178
00:12:59,120 --> 00:13:03,519
controller or whatever, like with Indiana Jones, like in a Twitch shooter.

179
00:13:02,639 --> 00:13:09,839
But I'm going to guess. I'm going to guess. I'm going to say 25.

180
00:13:07,200 --> 00:13:14,399
Man, I'm good. That's really impressive. 25 milliseconds full change. I mean,

181
00:13:12,480 --> 00:13:18,000
back when we were gaming on 60 Hz monitors, that would have been

182
00:13:15,920 --> 00:13:22,240
impressive even, cuz that's up to 16 milliseconds right there, just from the

183
00:13:20,560 --> 00:13:26,880
monitor. What's interesting, though, is this time my results are the same as

184
00:13:24,720 --> 00:13:30,240
when my team tested it a week ago. Maybe I just got a more favored connection

185
00:13:28,480 --> 00:13:33,920
when I put it in competitive mode. That's a possibility. But another

186
00:13:31,760 --> 00:13:38,560
possibility is that by running at such a high frame rate, NVIDIA is able to pair

187
00:13:36,399 --> 00:13:43,519
down the total latency. You can really feel it. Now the question is, is it

188
00:13:41,200 --> 00:13:48,160
worth what we're giving up compared to if we just spend the money upfront and

189
00:13:45,760 --> 00:13:52,959
get a GPU? Let's check it in. Naturally, I'll be doing so with my trusty

190
00:13:50,000 --> 00:13:55,680
transparent LTT screwdriver. Where can we get it?

191
00:13:53,760 --> 00:14:00,639
>> lttstore.com. >> Now, I fully recognize that most people

192
00:13:58,720 --> 00:14:04,560
are buying GeForce now because they don't have a 5080 in their gaming PC,

193
00:14:02,959 --> 00:14:09,680
but we have the luxury of trying them side by side. So, the way that I see it,

194
00:14:06,720 --> 00:14:14,959
we should. And I got to tell you, yeah, it looks a lot better, but if I was my

195
00:14:13,120 --> 00:14:18,720
brother-in-law, sorry, Alec, I probably wouldn't be able to tell the difference.

196
00:14:16,240 --> 00:14:22,480
At 1080p, you're still getting a bunch of, you know, shimmering and aliasing

197
00:14:20,959 --> 00:14:27,040
all over everything. And in some ways, it looks worse here. So, you're kind of

198
00:14:24,399 --> 00:14:32,800
trading one version of low fidelity for another one. As for the latency,

199
00:14:30,320 --> 00:14:38,079
yeah, it's lower, but not in a way that my weak human body could probably

200
00:14:35,360 --> 00:14:42,320
perceive without like really focusing on it.

201
00:14:39,680 --> 00:14:47,600
Yes, it's lower. Can confirm not using the internet to stream your game video

202
00:14:44,800 --> 00:14:51,839
is faster. A lot faster. About 15 millconds, which is cool because that

203
00:14:49,279 --> 00:14:58,079
actually tells us best case scenario running at 360 hertz. How much latency

204
00:14:55,839 --> 00:15:02,480
is NVIDIA's encoding? and transmission of this video data actually adding and

205
00:14:59,760 --> 00:15:07,839
the answer is not that much like 15 milliseconds. Oregon the GPU that we

206
00:15:05,680 --> 00:15:12,720
were gaming on was now it's right here that only cost me 15 freaking

207
00:15:10,480 --> 00:15:17,360
milliseconds of course. Okay, this was at 1080p. So this is kind of the worst

208
00:15:15,360 --> 00:15:20,800
case scenario for the local experience. Let's see how we feel about it once

209
00:15:18,880 --> 00:15:25,360
we're running at, you know, 4K with all the eye candy turned on. I think we

210
00:15:23,040 --> 00:15:28,959
could just go straight to Indiana Jones. >> Sure. I don't know that we need

211
00:15:26,639 --> 00:15:36,639
Borderlands. Jumping into Borderlands, this immediately looks way better, but I

212
00:15:34,480 --> 00:15:42,000
got to admit something I didn't really think about before when we were testing

213
00:15:38,399 --> 00:15:45,360
GeForce Now is that a bunch of our

214
00:15:42,000 --> 00:15:47,600
latency was actually added by frame gen,

215
00:15:45,360 --> 00:15:51,839
not by video encoding, and not by our distance to the NVIDIA data center. So,

216
00:15:50,000 --> 00:15:57,040
even though I'm running this locally, it's a little bit on the higher side.

217
00:15:54,160 --> 00:16:01,759
The animation is super smooth cuz we're probably running at hundreds of frames

218
00:15:58,959 --> 00:16:07,040
per second. But because we're running at 3x frame gen, just like NVIDIA was,

219
00:16:04,560 --> 00:16:13,279
feels a little jelloy, if you kind of get what I mean

220
00:16:09,519 --> 00:16:14,880
by that, man. So, how do I evaluate what

221
00:16:13,279 --> 00:16:20,240
it feels like? Because there's sort of two different versions of it. It's

222
00:16:17,279 --> 00:16:29,600
definitely better. Looks so smooth, but it feels so spongy. Say 35. 30.

223
00:16:26,480 --> 00:16:31,519
Damn. [music] To its credit, Frame Gen

224
00:16:29,600 --> 00:16:36,160
bamboozled me here. I don't know what's more interesting here. The fact that

225
00:16:33,199 --> 00:16:43,920
Frame Gen seems to be adding almost as much latency as GeForce Now was or the

226
00:16:40,079 --> 00:16:45,920
fact that my perception of it is better

227
00:16:43,920 --> 00:16:50,160
than it actually is. It's definitely the kind of thing that like older TVs, if

228
00:16:48,399 --> 00:16:54,000
you're a good enough gamer, you can get used to it. When we tested locally

229
00:16:52,079 --> 00:16:57,279
without frame gen, it was 20 milliseconds, just under.

230
00:16:55,519 --> 00:17:00,959
>> So, it's adding like 20 milliseconds and then it was only what, like another 25

231
00:16:59,120 --> 00:17:06,559
to get to the GeForce Now results. Like, that's crazy. Here we go. And

232
00:17:04,319 --> 00:17:10,640
yeah, it's sharper. But one of the things that I didn't realize because I

233
00:17:08,240 --> 00:17:15,120
hadn't done a side by side yet is that part of the blockiness that I was

234
00:17:12,400 --> 00:17:19,679
observing was probably just the game's built-in film grain filter not being

235
00:17:17,679 --> 00:17:23,600
interpreted that well. to go back. I'd probably just turn that off and it would

236
00:17:21,360 --> 00:17:27,039
help at least a little bit. The latency is obviously better, though, and that's

237
00:17:25,280 --> 00:17:29,840
not something that I could fix just by, you know, turning off film grain,

238
00:17:28,400 --> 00:17:34,000
though. Man, this is a beautiful looking game. Either running locally or with

239
00:17:31,919 --> 00:17:38,160
GeForce Now. Clearly, you're getting more with local hardware in terms of the

240
00:17:36,320 --> 00:17:43,200
feel and in terms of the image quality, which totally makes sense. I mean,

241
00:17:40,240 --> 00:17:46,799
streaming by definition can only add overhead compared to bare metal. But

242
00:17:45,600 --> 00:17:52,000
that doesn't mean that the value proposition is total trash either. It's

243
00:17:49,600 --> 00:17:57,120
true what the devil said. If you commit and buy GeForce Now annually, you're

244
00:17:54,160 --> 00:18:03,120
paying $1,000, which is the RTX 580s MSRP, over 5 years. And during that

245
00:18:01,200 --> 00:18:08,160
time, you probably won't need to upgrade your other hardware. So, if you're on a

246
00:18:05,760 --> 00:18:12,559
5-year upgrade cycle for your computer, that's some pretty significant savings.

247
00:18:09,919 --> 00:18:17,200
And if it works as well on LG TVs and Chromebooks as NVIDIA says, you might

248
00:18:14,799 --> 00:18:21,360
not need a computer at all. But the Angel meant to make some pretty good

249
00:18:19,200 --> 00:18:26,799
points, too. You don't own your cloud gaming PC, which puts you completely at

250
00:18:23,840 --> 00:18:30,480
NVIDIA's mercy for price increases and hardware upgrades. At least you can

251
00:18:28,640 --> 00:18:34,640
install your own Steam cloud supported games now, but unless you pay extra to

252
00:18:32,799 --> 00:18:38,559
NVIDIA for persistent storage, you're going to need to install those fresh

253
00:18:36,160 --> 00:18:43,120
every time you fire up GeForce Now. And yeah, what if your internet goes down?

254
00:18:40,880 --> 00:18:49,120
Steam has an offline mode, but GeForce Now, [snorts] that makes every game

255
00:18:45,679 --> 00:18:51,280
always online all the time. And this

256
00:18:49,120 --> 00:18:54,240
entire conversation completely ignores that if you're in a rural area that

257
00:18:52,880 --> 00:18:58,080
doesn't have fiber, or if you just happen to be on a node that's unreliable

258
00:18:56,080 --> 00:19:05,840
or far away, you can just forget the whole thing. Uh, was there anything else

259
00:18:59,600 --> 00:19:07,760
you wanted to add? Oh, uh, sponsor,

260
00:19:05,840 --> 00:19:11,520
delete me. Even if you're super safe online, your data might still be

261
00:19:09,760 --> 00:19:14,880
compromised without you even knowing. See, sometimes data brokers can get

262
00:19:13,280 --> 00:19:18,160
their hands on your email addresses, phone numbers, and they will sell that

263
00:19:16,640 --> 00:19:22,160
information to the highest bidder. Delete Me can help protect your identity

264
00:19:20,000 --> 00:19:25,360
online by reaching out to data brokers on your behalf with requests for your

265
00:19:24,080 --> 00:19:28,960
info to be removed. And this doesn't have to just be a you thing. Delete Me

266
00:19:27,280 --> 00:19:32,559
also offers family plans where three other loved ones can receive protection,

267
00:19:30,960 --> 00:19:36,799
too. Each member under our plan will get a personalized data sheet showing what

268
00:19:34,320 --> 00:19:41,360
info was removed and from where. So, head on over to joindeleteme.com/ltt20

269
00:19:39,919 --> 00:19:45,679
or click our link in the description to get 20% off today. If you guys enjoyed

270
00:19:43,679 --> 00:19:50,880
this video, why not check out the first time that I checked out GeForce Now. It

271
00:19:47,760 --> 00:19:53,840
was impressive, but man, has it ever

272
00:19:50,880 --> 00:19:53,840
come a long way.