WEBVTT

1
00:00:00.000 --> 00:00:03.920
<v 0>I've laid low, right. You know, until now I could be canceled after this. Hello,</v>

2
00:00:03.980 --> 00:00:04.960
I'm Robert Sachs.

3
00:00:04.960 --> 00:00:08.440
I'm a patent attorney and I've been practicing in the Silicon valley for over 30

4
00:00:08.440 --> 00:00:11.120
years. I've represented many of the companies in the valley that, you know,

5
00:00:11.120 --> 00:00:11.680
the names of,

6
00:00:11.680 --> 00:00:15.560
and some you don't as they were startups and they didn't get to their IPO today.

7
00:00:15.560 --> 00:00:17.720
We're gonna be looking at some interesting patents from Google.

8
00:00:17.720 --> 00:00:21.760
I'm gonna be talking about what the technology covers and how this may be useful

9
00:00:21.760 --> 00:00:23.160
or not in your life.

10
00:00:25.580 --> 00:00:30.160
The first patent we're gonna look at is from 2015 and it's for an odor

11
00:00:30.160 --> 00:00:32.840
removing device. Now, this is a very interesting contraption.

12
00:00:32.840 --> 00:00:37.160
What we see here on the left side of the screen is a fragrance submitting

13
00:00:37.160 --> 00:00:41.160
device. It's a small unit about this big, it's got a display screen on it,

14
00:00:41.160 --> 00:00:43.320
as you can see in the front, and it has a chamber,

15
00:00:43.370 --> 00:00:45.200
it has a fragrance material in there.

16
00:00:45.200 --> 00:00:49.880
And this thing can turn on and off and emit a fragrance because you may

17
00:00:49.880 --> 00:00:52.760
smell bad. And that's the key insight of this pattern.

18
00:00:52.940 --> 00:00:57.720
You may smell bad and you don't want to smell bad around people. You know,

19
00:00:57.720 --> 00:01:01.040
you're on your way. Home from the gym, you're all stinking and sweaty.

20
00:01:01.040 --> 00:01:04.840
And you gotta go from here to your office. And it turns out that someone,

21
00:01:04.840 --> 00:01:07.640
you know, in your social is gonna be on your route.

22
00:01:07.640 --> 00:01:10.800
That's this little guy here in red, that's a defined person.

23
00:01:10.800 --> 00:01:13.320
The patent talks about figuring out where you are,

24
00:01:13.320 --> 00:01:16.400
figuring out where that person is and giving you an alternate route.

25
00:01:16.400 --> 00:01:19.600
So you avoid going anywhere near anybody, you know,

26
00:01:19.600 --> 00:01:22.960
while you smell well, you know,

27
00:01:22.960 --> 00:01:27.080
I'm not so certain about this technology. If you know, you have this issue,

28
00:01:27.080 --> 00:01:31.360
you probably would be wearing deodorant to begin with and your friends would

29
00:01:31.600 --> 00:01:35.880
probably already be tolerant of you. What kind of world is it? We live in that.

30
00:01:35.880 --> 00:01:38.440
We're so worried about what we smell like,

31
00:01:38.440 --> 00:01:41.680
that we're willing to offend other people that we don't know,

32
00:01:41.780 --> 00:01:44.680
but not willing to offend the people we know and care about.

33
00:01:44.680 --> 00:01:47.200
This seems something out of black mirror. I don't know about you,

34
00:01:47.260 --> 00:01:49.360
but I don't think this one is gonna make it to market.

35
00:01:52.130 --> 00:01:56.200
As you may know, Google spun off Waymo to develop autonomous vehicles.

36
00:01:56.340 --> 00:01:59.120
One of the issues with autonomous vehicles is without a driver,

37
00:01:59.120 --> 00:02:01.480
they're sometimes gonna possibly hit somebody.

38
00:02:01.540 --> 00:02:03.200
And what are you gonna do about that?

39
00:02:03.200 --> 00:02:07.680
Google came up with a very interesting solution and that's basically what you

40
00:02:07.680 --> 00:02:10.160
could call sticky car adhesive,

41
00:02:10.160 --> 00:02:13.720
vehicle front end for mitigation of secondary pedestrian impact.

42
00:02:13.820 --> 00:02:16.200
The problem being is that when a car hits a person,

43
00:02:16.220 --> 00:02:19.000
the person bounces off and hits something else.

44
00:02:19.490 --> 00:02:21.520
So let's not have them bounce off.

45
00:02:21.520 --> 00:02:24.600
What we're gonna do is we're gonna have a sticky layer.

46
00:02:24.600 --> 00:02:29.080
This is some adhesives on the front end of the car. So the patent shows these,

47
00:02:29.080 --> 00:02:32.000
this front end, the car, those little dots are some adhesive material,

48
00:02:32.260 --> 00:02:34.760
and thankfully that's not exposed.

49
00:02:34.760 --> 00:02:38.320
You don't see that adhesive material it's behind a coating. What?

50
00:02:38.320 --> 00:02:41.120
Google calls an eggshell light coating. Okay.

51
00:02:41.120 --> 00:02:45.120
So you've got this eggshell light coating on the car that is meant to be

52
00:02:45.460 --> 00:02:50.040
instantaneously breakable. So when the car hits a person, the coating is broken,

53
00:02:50.230 --> 00:02:54.200
exposing the layer to adhere the colliding object upon that initial.

54
00:02:54.200 --> 00:02:59.120
So basically you're gonna stick to the cough so you don't go flying off. Okay.

55
00:03:00.310 --> 00:03:03.360
That's a little bit odd. I think there's some serious problems with that.

56
00:03:03.360 --> 00:03:07.360
First of all, if this layer is so instantaneously breakable,

57
00:03:07.390 --> 00:03:10.680
it's gonna break all the time, right? You're gonna bump into something.

58
00:03:10.680 --> 00:03:13.120
You're gonna break it. I can imagine kids walking down the streets,

59
00:03:13.120 --> 00:03:16.920
looking at people's hoods and slapping cats and dogs and sneakers on there.

60
00:03:16.920 --> 00:03:20.760
So stuff is gonna get on there all the time. It's gonna be really expensive.

61
00:03:20.760 --> 00:03:24.240
And then what happens when you actually are hit by this? Are you now stuck here?

62
00:03:24.240 --> 00:03:28.520
Do you have to take off of your clothes to get off of this car? I don't know.

63
00:03:28.530 --> 00:03:31.200
Doesn't seem like a very practical solution. And again,

64
00:03:31.200 --> 00:03:34.240
something that's just gonna make repairing cars all the more expensive.

65
00:03:34.750 --> 00:03:37.920
I wouldn't wanna be stuck on anybody's hood. I rather just go flying off.

66
00:03:37.920 --> 00:03:41.920
Thank you. This is an interesting technology. Yeah, not so much.

67
00:03:45.020 --> 00:03:49.160
Now we're gonna look at some more car related safety technology. And again,

68
00:03:49.160 --> 00:03:53.400
I think this reflects a certain lack of understanding about what people are

69
00:03:53.400 --> 00:03:55.280
looking for in a car here.

70
00:03:55.280 --> 00:03:58.880
What we have is reducing the surface rigidity of a vehicle.

71
00:03:58.880 --> 00:04:01.640
So now we actually have an illustration of a car. It looks like a car,

72
00:04:01.640 --> 00:04:04.880
but you can see the top section here is the view of the bumper.

73
00:04:04.880 --> 00:04:09.760
It's this protruding section and behind it are cables

74
00:04:09.810 --> 00:04:12.560
or some kind of tension elements on the top one.

75
00:04:12.560 --> 00:04:15.720
You've got a curve surface and you've got these cables and winches and the

76
00:04:15.720 --> 00:04:17.080
cables and the winters work together.

77
00:04:17.100 --> 00:04:21.720
And they can change the shape of that service and they could make it less rigid

78
00:04:21.760 --> 00:04:25.320
under impact collision. On the bottom one, we have a corrugated bumper,

79
00:04:25.670 --> 00:04:28.280
same idea. When the car senses an impact,

80
00:04:28.450 --> 00:04:31.480
it will reduce the tension in those tension members.

81
00:04:31.810 --> 00:04:34.560
So the rigidity of that portion of the vehicle is reduced.

82
00:04:34.560 --> 00:04:37.840
What you could call soft car car, slams into something.

83
00:04:37.840 --> 00:04:40.960
The car detects that there's a collision and makes it softer,

84
00:04:40.960 --> 00:04:45.120
reduces the rigidity of that. Now that may work, boy,

85
00:04:45.120 --> 00:04:49.000
it's gonna make for some bulky, ugly looking cars,

86
00:04:49.350 --> 00:04:52.120
just the additional mass and weight from this,

87
00:04:52.230 --> 00:04:55.560
that alone is gonna undermine the whole utility of it.

88
00:04:55.750 --> 00:04:59.440
A heavier car is the more energy it has when it hits something.

89
00:04:59.690 --> 00:05:03.200
So you may make the car softer, but now it's a heavier car.

90
00:05:03.200 --> 00:05:07.600
So it's gonna do as much damage as it did before. So it doesn't really pan out.

91
00:05:07.600 --> 00:05:09.120
I don't think this is one's coming to market.

92
00:05:12.580 --> 00:05:15.280
One of the truths of our times is that a service is free.

93
00:05:15.280 --> 00:05:16.800
It means you are the product.

94
00:05:16.800 --> 00:05:21.000
Google has taken this to the next level in a smart home system.

95
00:05:21.000 --> 00:05:24.800
What they call privacy aware, personalized content for smart home.

96
00:05:25.260 --> 00:05:27.920
And the idea is this you're in your house.

97
00:05:27.980 --> 00:05:31.080
You have books and media and other things lying around.

98
00:05:31.080 --> 00:05:34.800
You've got a camera system and that camera system is observing what you're

99
00:05:34.800 --> 00:05:36.040
doing. So when you pick up a book,

100
00:05:36.570 --> 00:05:39.640
it sees that you picked up a book and it can scan the title of that book,

101
00:05:39.640 --> 00:05:40.170
for example,

102
00:05:40.170 --> 00:05:43.720
or a magazine that you're reading or something that's on television.

103
00:05:43.720 --> 00:05:48.680
And it will train a neural network to learn your interactions between that

104
00:05:48.680 --> 00:05:52.960
type of media and how you consume it. So what time of day, where you are,

105
00:05:52.960 --> 00:05:53.840
what room you're in,

106
00:05:54.020 --> 00:05:58.640
all of the content that you interact goes in a neural network to learn your

107
00:05:58.920 --> 00:06:02.360
behaviors as time goes on, that thing is smart.

108
00:06:02.580 --> 00:06:04.840
And it determines where you are in your house,

109
00:06:05.240 --> 00:06:10.000
looks at the content that you have near you, and then makes recommendations.

110
00:06:10.000 --> 00:06:12.640
So what we see here in this illustration is it says,

111
00:06:12.830 --> 00:06:15.960
I noticed you have a copy of the godfather by your bed.

112
00:06:16.140 --> 00:06:19.920
The movie based on this novel is showing tonight on channel.

113
00:06:20.150 --> 00:06:21.440
What have you coming.

114
00:06:21.440 --> 00:06:23.400
<v 3>To my house that I cannot do?</v>

115
00:06:23.670 --> 00:06:25.720
<v 0>This is smart advertising.</v>

116
00:06:25.740 --> 00:06:30.320
And the idea is it selects that content without communicating that information

117
00:06:30.330 --> 00:06:33.080
to a content server in a distant network,

118
00:06:33.220 --> 00:06:37.440
the patent is full of examples of advertising,

119
00:06:37.440 --> 00:06:40.880
how the system could detect, for example, you're playing guitar.

120
00:06:40.880 --> 00:06:44.640
And so it shows you an advertisement for guitar on the television or that you

121
00:06:44.640 --> 00:06:46.280
are interested in basketball,

122
00:06:46.280 --> 00:06:49.680
cuz you're reading a basketball magazine and shows you an advertisement for

123
00:06:49.910 --> 00:06:53.200
basketball camp. This is all about monetizing,

124
00:06:53.200 --> 00:06:55.440
what people are doing in their personal lives.

125
00:06:55.580 --> 00:07:00.400
And that it's privacy aware because it doesn't transmit the OCR information

126
00:07:00.400 --> 00:07:03.600
about what it has seen to a network offsite.

127
00:07:03.600 --> 00:07:07.320
So it all lives locally on your house, but what the patent doesn't address.

128
00:07:07.320 --> 00:07:10.960
And I think this is of concern is what happens to the information about your

129
00:07:10.960 --> 00:07:13.880
actual behaviors. It has to send that information somewhere,

130
00:07:13.880 --> 00:07:16.720
particularly to be able to promote your advertising information.

131
00:07:16.720 --> 00:07:19.880
And there's nothing in the patent that says that this information about your

132
00:07:20.080 --> 00:07:22.680
behaviors is not transmitted. For example, to third parties,

133
00:07:22.930 --> 00:07:26.640
to people who are advertising your content to you. So yes,

134
00:07:26.640 --> 00:07:28.560
it's privacy aware in one side,

135
00:07:28.650 --> 00:07:31.920
it doesn't send out the behaviors themselves that neural network has learned,

136
00:07:32.020 --> 00:07:35.920
but it doesn't really seem to protect the privacy of what you're actually doing

137
00:07:36.140 --> 00:07:39.400
and how that behavior is gonna be used by third party advertisers.

138
00:07:39.400 --> 00:07:41.360
That's not privacy protection in my mind.

139
00:07:43.580 --> 00:07:47.360
The last patent we're gonna look at from Google is a medical technology patent.

140
00:07:47.360 --> 00:07:50.840
I think this is the most interesting one and has the most potential for really

141
00:07:50.840 --> 00:07:55.480
benefiting people in the future. Google calls, this interocular device,

142
00:07:55.830 --> 00:07:57.760
I would call this active eye.

143
00:07:57.760 --> 00:08:01.280
What you see here in yellow is a part of the lens capsule.

144
00:08:01.280 --> 00:08:04.760
That's the part of your eye where actually has the lens and inside that's a

145
00:08:04.760 --> 00:08:05.920
fluid filled sack.

146
00:08:05.920 --> 00:08:10.480
You could say the fluid is removed and a polymeric soft material is placed in

147
00:08:10.480 --> 00:08:15.200
there along with an active device and active lens and that polymeric

148
00:08:15.400 --> 00:08:19.840
material, then bonds the lens to the inside of the lens capsule.

149
00:08:19.850 --> 00:08:24.840
There's a sensor that detects when you are attempting to

150
00:08:24.840 --> 00:08:26.200
focus, that's called accommodation.

151
00:08:26.340 --> 00:08:29.320
And so that bonds together with the active lens.

152
00:08:29.320 --> 00:08:33.320
And then there's a controller that detects when you are trying to focus and then

153
00:08:33.320 --> 00:08:35.640
will adjust the lens to do the focusing for you.

154
00:08:35.640 --> 00:08:39.400
So you can think of this as a it's an active eye or a bionic eye.

155
00:08:40.700 --> 00:08:42.640
Now this is very interesting technology.

156
00:08:42.730 --> 00:08:45.640
It can solve a lot of problems for people who are nearsighted,

157
00:08:45.640 --> 00:08:47.760
who have lost their vision, who have cataracts,

158
00:08:47.860 --> 00:08:52.560
all sorts of eye problems can be remedied by this technology that said

159
00:08:52.560 --> 00:08:57.000
it also opens up the prospect of our old friend GORT. Maybe he had that.

160
00:08:57.620 --> 00:08:58.453
<v 2>Wow.</v>

161
00:09:00.570 --> 00:09:04.240
<v 0>It certainly seems to make us a little bit closer to our robot overlords.</v>

162
00:09:06.080 --> 00:09:09.040
Thank you for watching experts react. If you like the content,

163
00:09:09.040 --> 00:09:11.640
please subscribe to C net's YouTube channels.

164
00:09:11.640 --> 00:09:15.280
If you have an opinion and I am sure you do leave them in the comments below.

165
00:09:15.430 --> 00:09:16.520
Thanks for watching.

