1
00:00:02,440 --> 00:00:06,800
Speaker 1: Bloomberg Audio Studios, podcasts, radio news.

2
00:00:07,320 --> 00:00:12,039
Speaker 2: Adobe makes its AI video debut, announcing new tools allowing

3
00:00:12,080 --> 00:00:14,960
Speaker 2: for video generation now. The company calls it Firefly, and

4
00:00:15,000 --> 00:00:17,639
Speaker 2: it's set to be implemented in its video editing software

5
00:00:17,640 --> 00:00:20,320
Speaker 2: Premiere Pro, taking on, of course, the likes of OpenAI,

6
00:00:20,440 --> 00:00:24,400
Speaker 2: Meta Google a place to say. Adobe CEO Shantanu Orian

7
00:00:24,560 --> 00:00:28,800
Speaker 2: joins us now on the back of the announcement in Miami. Shantanu,

8
00:00:29,000 --> 00:00:32,080
Speaker 2: what's interesting is you're getting into the hands of users

9
00:00:32,280 --> 00:00:36,440
Speaker 2: faster than rivals. Is it based on what it's trained upon?

10
00:00:36,600 --> 00:00:38,440
Speaker 2: How are you able to give it to us so quick?

11
00:00:40,280 --> 00:00:42,800
Speaker 1: Caroline, first, thanks for having me on the show.

12
00:00:42,920 --> 00:00:43,240
Speaker 3: Max.

13
00:00:43,280 --> 00:00:47,000
Speaker 1: As you know, is our biggest creativity event, and it's

14
00:00:47,040 --> 00:00:50,240
Speaker 1: the time that our engineers really take it upon themselves

15
00:00:50,880 --> 00:00:54,880
Speaker 1: to show their magic to the entire creative community. As

16
00:00:54,920 --> 00:00:58,160
Speaker 1: you point out, we have been hard at work ensuring

17
00:00:58,200 --> 00:01:01,520
Speaker 1: that our Firefly set of models works on things like

18
00:01:01,600 --> 00:01:05,040
Speaker 1: images and vector and design. But today's a big day

19
00:01:05,160 --> 00:01:10,280
Speaker 1: because we have been releasing our Firefly video models. You know,

20
00:01:10,319 --> 00:01:12,880
Speaker 1: our strategy has always been that it's not just about

21
00:01:12,880 --> 00:01:16,080
Speaker 1: the models that people develop, but it's the integration with

22
00:01:16,200 --> 00:01:19,200
Speaker 1: the apps where the real magic is, and so we've

23
00:01:19,200 --> 00:01:23,000
Speaker 1: had thirteen billion generations where people have used either Photoshop

24
00:01:23,080 --> 00:01:25,920
Speaker 1: or illustrator to do it, and today that magic comes

25
00:01:25,920 --> 00:01:29,360
Speaker 1: to video in Premiere Pro, will be in after effects

26
00:01:29,400 --> 00:01:33,040
Speaker 1: as well as on firefly dot com Chantanu.

27
00:01:33,840 --> 00:01:37,000
Speaker 3: The most interesting part is the accessibility. There are no

28
00:01:37,240 --> 00:01:41,479
Speaker 3: usage limits one, but beyond the basic subscription fees, you're

29
00:01:41,520 --> 00:01:46,000
Speaker 3: not charging for the AI tools. Will that change.

30
00:01:47,000 --> 00:01:50,040
Speaker 1: Well, Our focus really first is on ensuring that we

31
00:01:50,240 --> 00:01:53,600
Speaker 1: developed these great models. We're very differentiated, and these models

32
00:01:53,600 --> 00:01:56,920
Speaker 1: were designed to be commercially safe, so they've only been

33
00:01:56,960 --> 00:02:01,400
Speaker 1: trained on licensed and public domain content, and you know,

34
00:02:01,440 --> 00:02:03,520
Speaker 1: we want people to experiment with it. So I think

35
00:02:03,600 --> 00:02:07,840
Speaker 1: clearly on the imaging, vector and design front, we have

36
00:02:07,920 --> 00:02:09,960
Speaker 1: said as part of the subscription, we don't want you

37
00:02:10,000 --> 00:02:13,000
Speaker 1: to think about credits. We just want you to experience

38
00:02:13,040 --> 00:02:15,600
Speaker 1: the magic that exists. A video might be a little

39
00:02:15,639 --> 00:02:18,040
Speaker 1: bit different ed. What we've announced today is that it's

40
00:02:18,080 --> 00:02:21,720
Speaker 1: available in a beta form. The cost of producing video

41
00:02:21,880 --> 00:02:25,400
Speaker 1: is clearly a little bit more expensive, but first steps

42
00:02:25,440 --> 00:02:27,240
Speaker 1: is to just get it in the hands of people,

43
00:02:27,600 --> 00:02:31,240
Speaker 1: have them try it, give us feedback, make sure it's specific,

44
00:02:31,280 --> 00:02:33,760
Speaker 1: and then as I think the amount of video that's

45
00:02:33,840 --> 00:02:38,760
Speaker 1: being generated is expanding. We will probably have different pricing

46
00:02:38,840 --> 00:02:39,560
Speaker 1: models for that.

47
00:02:41,200 --> 00:02:43,720
Speaker 3: One of the features is to take an existing piece

48
00:02:43,720 --> 00:02:47,800
Speaker 3: of video and use generative AI to extend it, extend

49
00:02:47,840 --> 00:02:51,359
Speaker 3: beyond actual. As someone that trained in their early career

50
00:02:51,400 --> 00:02:55,560
Speaker 3: on premiere, I'm excited to use that. How technically difficult

51
00:02:55,680 --> 00:02:58,880
Speaker 3: was it for Adobe to achieve that versus the work

52
00:02:58,880 --> 00:02:59,760
Speaker 3: you've done in photo?

53
00:03:01,520 --> 00:03:04,639
Speaker 1: Well, every time you show something in photo and people say,

54
00:03:04,680 --> 00:03:08,920
Speaker 1: oh my god, that's magic. To your point, the ability

55
00:03:08,960 --> 00:03:12,480
Speaker 1: to extend that into video across every frame is so

56
00:03:12,600 --> 00:03:16,640
Speaker 1: much more computationally intensive as well as challenging in order

57
00:03:16,680 --> 00:03:19,399
Speaker 1: to do. But the use cases are so dramatic. You've

58
00:03:19,400 --> 00:03:22,880
Speaker 1: had a video for some reason, unfortunately it got clipped

59
00:03:22,880 --> 00:03:25,440
Speaker 1: off and you just want to extend it. You have

60
00:03:25,560 --> 00:03:28,919
Speaker 1: video that you want to align with audio, the song

61
00:03:29,040 --> 00:03:30,960
Speaker 1: is not long enough. How do you do it? So?

62
00:03:31,040 --> 00:03:35,920
Speaker 1: I think this generative extent concept, which is really something

63
00:03:35,960 --> 00:03:38,960
Speaker 1: that's very pervasive in video editing, is something that the

64
00:03:39,000 --> 00:03:41,600
Speaker 1: customers will really like. The Other thing that we've done

65
00:03:42,040 --> 00:03:45,040
Speaker 1: is because we can now analyze every frame. When you

66
00:03:45,080 --> 00:03:48,960
Speaker 1: want to remove something that inadvertently came into the image,

67
00:03:49,080 --> 00:03:51,720
Speaker 1: you don't have to go through the entire process again,

68
00:03:52,000 --> 00:03:54,600
Speaker 1: you can just through post production change it. So some

69
00:03:54,640 --> 00:03:58,560
Speaker 1: pretty incredible stuff and we're really excited to see what

70
00:03:58,600 --> 00:03:59,720
Speaker 1: our customers do with this.

71
00:04:00,280 --> 00:04:06,040
Speaker 2: You talked about the computational feat the intensity, What about GPUs,

72
00:04:06,160 --> 00:04:08,920
Speaker 2: what about your own access to compute?

73
00:04:10,200 --> 00:04:13,440
Speaker 1: You know, well, we've partnered with you know, Nvideo for

74
00:04:13,800 --> 00:04:17,080
Speaker 1: years in making sure all of this stuff is available.

75
00:04:17,720 --> 00:04:19,960
Speaker 1: You know, it was great to hear from Jensen this

76
00:04:20,080 --> 00:04:24,000
Speaker 1: morning when he heard about our new announcements. But you know,

77
00:04:24,040 --> 00:04:27,240
Speaker 1: we're partnering with all of the chip manufacturers as well

78
00:04:27,279 --> 00:04:30,600
Speaker 1: as all of the hyper scalers because our intent is

79
00:04:31,000 --> 00:04:34,560
Speaker 1: not just to make this magic available through you know,

80
00:04:34,600 --> 00:04:38,440
Speaker 1: those chips, but also increasingly people are talking about how

81
00:04:38,480 --> 00:04:41,039
Speaker 1: you can make this available on a mobile device or

82
00:04:41,080 --> 00:04:44,080
Speaker 1: in a hybrid environment. So you know, as long as

83
00:04:44,120 --> 00:04:47,640
Speaker 1: they keep performing in terms of what they are doing

84
00:04:47,880 --> 00:04:51,520
Speaker 1: innovation on the chips, we'll take advantage of every single

85
00:04:52,160 --> 00:04:54,960
Speaker 1: available GPU and CPU for our customers.

86
00:04:55,440 --> 00:04:59,600
Speaker 2: That's expensive and also, look, investors have been so excited

87
00:04:59,640 --> 00:05:02,800
Speaker 2: around just generative AI full stop, but they want to

88
00:05:02,839 --> 00:05:05,760
Speaker 2: see the revenue. They want to see the profitability chantanute.

89
00:05:05,760 --> 00:05:08,760
Speaker 2: You feel that you're able to give more light, more

90
00:05:08,800 --> 00:05:11,440
Speaker 2: clarity to the market as to when this really becomes

91
00:05:12,080 --> 00:05:13,279
Speaker 2: very revenue generating.

92
00:05:14,640 --> 00:05:18,240
Speaker 1: You're absolutely right, Carolin, that all of the focus thus

93
00:05:18,320 --> 00:05:20,960
Speaker 1: far has been on training, and people have been creating

94
00:05:21,000 --> 00:05:24,400
Speaker 1: all the models, and so all of the investment that's

95
00:05:24,440 --> 00:05:27,760
Speaker 1: actually happening is on the training side of the equation. However,

96
00:05:28,160 --> 00:05:32,400
Speaker 1: I think Adobe has delivered more generative AI software than

97
00:05:32,480 --> 00:05:36,839
Speaker 1: virtually any other large company that had a traditional business.

98
00:05:36,880 --> 00:05:39,279
Speaker 1: So if you think about what we've done with Acrobat

99
00:05:39,279 --> 00:05:42,800
Speaker 1: AI Assistant or Photoshop, and clearly the emphasis has to

100
00:05:42,839 --> 00:05:46,239
Speaker 1: go on inference because that's where people are seeing value,

101
00:05:46,279 --> 00:05:49,680
Speaker 1: that's where it's getting embedded in the workflows. You know,

102
00:05:49,760 --> 00:05:52,920
Speaker 1: we will have three or four different ways in which

103
00:05:52,960 --> 00:05:55,359
Speaker 1: we monetize it, and we have a financial analyst Q

104
00:05:55,440 --> 00:05:57,640
Speaker 1: and A where we will talk about this as well.

105
00:05:57,640 --> 00:06:01,160
Speaker 1: But as a you know, insight into that. Certainly our

106
00:06:01,200 --> 00:06:04,240
Speaker 1: subscription tier pricing will have the ability to lose some

107
00:06:04,279 --> 00:06:08,000
Speaker 1: of this AI. We will have add on services. Firefly

108
00:06:08,160 --> 00:06:11,880
Speaker 1: services is being used by every large enterprise. They want

109
00:06:11,880 --> 00:06:14,680
Speaker 1: to understand how they create a custom model where they

110
00:06:14,720 --> 00:06:18,200
Speaker 1: can have, for example, Bloomberg assets in a custom model

111
00:06:18,240 --> 00:06:21,039
Speaker 1: where only Bloomberg is allowed to generate with that content.

112
00:06:21,160 --> 00:06:23,520
Speaker 1: So I think we have multiple ways in which we

113
00:06:23,560 --> 00:06:28,520
Speaker 1: monetize it today. Already, performance marketers will have a new

114
00:06:28,520 --> 00:06:31,520
Speaker 1: product today which is called gen Studio, so you can

115
00:06:31,600 --> 00:06:34,719
Speaker 1: increase the agility by which you place ads on every

116
00:06:34,760 --> 00:06:38,000
Speaker 1: advertising system. So you know, we're both excited about the

117
00:06:38,040 --> 00:06:42,280
Speaker 1: technology and the value, but we're also confident that customers

118
00:06:42,279 --> 00:06:44,840
Speaker 1: will see the value and enable us to monetize it.

119
00:06:46,000 --> 00:06:49,640
Speaker 3: Chantonnay intense interest in your m and A strategy. Karen

120
00:06:49,680 --> 00:06:52,520
Speaker 3: and I were with the Runway CEO last week. Will

121
00:06:52,560 --> 00:06:54,200
Speaker 3: you look at companies like Runway.

122
00:06:55,320 --> 00:06:57,800
Speaker 1: We're partnering with Runway. You know, I think as it

123
00:06:57,839 --> 00:07:01,840
Speaker 1: relates to models, we believe that Adobe for its core

124
00:07:02,480 --> 00:07:07,120
Speaker 1: domains such as video and imaging and design, and we

125
00:07:07,160 --> 00:07:10,440
Speaker 1: will certainly have our own models, but people will want

126
00:07:10,520 --> 00:07:13,960
Speaker 1: us to support third party models in our applications because

127
00:07:14,000 --> 00:07:17,480
Speaker 1: that is the interface that everybody uses. We've demonstrated that

128
00:07:17,600 --> 00:07:21,280
Speaker 1: for video at IBC with Runway, and certainly we will

129
00:07:21,280 --> 00:07:23,760
Speaker 1: support these custom models. As I said, so if you're

130
00:07:23,840 --> 00:07:26,720
Speaker 1: Coke or Nike or Bank of America and you're using

131
00:07:26,760 --> 00:07:29,920
Speaker 1: it within your context, you have access to that, So

132
00:07:30,560 --> 00:07:33,320
Speaker 1: you know this is going to be such an exciting area.

133
00:07:33,800 --> 00:07:36,800
Speaker 1: Each model, I believe ED will have its own personality.

134
00:07:37,160 --> 00:07:40,040
Speaker 1: So rather than you know, try to create one model

135
00:07:40,080 --> 00:07:42,840
Speaker 1: that's the super set, We're going to support all these

136
00:07:42,880 --> 00:07:45,960
Speaker 1: models because they will do some things differently and better.

137
00:07:47,240 --> 00:07:50,280
Speaker 3: Adobe CEO Shansanuin Ryan is great to finally have you

138
00:07:50,360 --> 00:07:53,000
Speaker 3: here on Bloomberg Technology. Thank you for your time