WEBVTT

NOTE Sentence-level transcript of https://www.youtube.com/watch?v=pPj_tjlvYjA

NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/pPj_tjlvYjA.html. A cue ends where the next begins, or 2 s after its last word.

s1
00:00:01.309 --> 00:00:03.309
[music]

s2
00:00:12.920 --> 00:00:13.760
Hi everyone.

s3
00:00:13.760 --> 00:00:15.440
I think we can start.

s4
00:00:15.440 --> 00:00:19.520
Uh but before we start, I want to ask you a couple of questions.

s5
00:00:19.520 --> 00:00:24.680
So, first of all, how many of you took a photo or video during this conference?

s6
00:00:24.680 --> 00:00:27.320
Please raise your hands.

s7
00:00:27.320 --> 00:00:35.320
Okay, and how many of you actually posted any video content from it online?

s8
00:00:36.880 --> 00:00:38.360
Not that many.

s9
00:00:38.360 --> 00:00:41.040
And um to be honest, that was me.

s10
00:00:41.040 --> 00:00:53.200
I I was recording a lot of content during conferences, events, trips, meetups, uh and I never posted them online because video editing is hard.

s11
00:00:53.200 --> 00:00:56.480
Uh it sounds and it is a lot of work.

s12
00:00:56.480 --> 00:01:01.000
Uh it's tedious and it's largely still manual.

s13
00:01:01.000 --> 00:01:07.120
So, and and it also feels like an art and not really automated.

s14
00:01:07.120 --> 00:01:18.760
And that's why we're building RealFull and we're trying to tackle agentic video editing problem from uh video editing problem from the agentic standpoint.

s15
00:01:18.760 --> 00:01:19.480
I'm Kate.

s16
00:01:19.480 --> 00:01:22.960
I'm founder and CEO at RealFull.

s17
00:01:22.960 --> 00:01:27.880
But let's first talk about what's agentic video editing is.

s18
00:01:27.880 --> 00:01:34.160
So, as a user, you just drop in your media, photos and videos, and provide some context.

s19
00:01:34.160 --> 00:01:39.960
It can be uh the context what happened in these media files, or it can be some directions.

s20
00:01:39.960 --> 00:01:46.600
For example, like add captions, add music, add voiceover, and something like that.

s21
00:01:46.600 --> 00:02:02.760
And then, the agent will go, understand your media, uh find the right moments, assemble everything together, generate captions, music, voiceover, b-rolls, uh and give you a ready-to-share clip.

s22
00:02:02.760 --> 00:02:09.479
Uh and um yeah, so basically, uh that's uh video agentic video editing.

s23
00:02:09.479 --> 00:02:12.400
So, agent does everything by itself.

s24
00:02:12.400 --> 00:02:30.600
Or another example, you recorded a speak-to-camera video and you have a lot of pauses, unsuccessful shots, and you expect an agent to figure it out to remove unsuccessful shots, remove pauses, and give you a ready-to-share

s25
00:02:30.600 --> 00:02:31.880
clip.

s26
00:02:31.880 --> 00:02:42.040
Um and uh the the interesting thing is that a lot of actually a lot of the things inside this pipeline can be automated.

s27
00:02:42.040 --> 00:02:45.520
And this is exactly what we're doing at Real Fall.

s28
00:02:45.520 --> 00:02:52.280
Oh, this is um the example of uh video edited.

s29
00:02:52.680 --> 00:03:00.080
And um since we're at AI Engineering Conference, I wanted to talk a little bit about infrastructure.

s30
00:03:00.080 --> 00:03:08.519
And from the infrastructure standpoint, agentic uh video editor is very similar to agentic app builder.

s31
00:03:08.519 --> 00:03:10.480
Uh sorry, there is a typo on the slide.

s32
00:03:10.480 --> 00:03:14.200
So, the uh second column is agentic video editor.

s33
00:03:14.200 --> 00:03:23.280
So, both of them have a prompt uh a UI for prompt uh and in the video editor case, it's a media plus prompt.

s34
00:03:23.280 --> 00:03:26.280
And usually on back end, what's happening?

s35
00:03:26.280 --> 00:03:38.000
There is a remote machine which is called sandbox, uh which is spinning up, and inside this machine, there is an agent with tools and skills, which is working on

s36
00:03:38.000 --> 00:03:40.720
uh what you're uh you're you're asking it to do.

s37
00:03:40.720 --> 00:03:44.880
Uh in the case of the agentic uh app builder, it's a code base.

s38
00:03:44.880 --> 00:03:50.560
Uh in the case of the agentic video editor, it's a video video composition.

s39
00:03:50.560 --> 00:04:03.160
And as a result, in the Argentic builder user get an app preview and for the video editor users user get rendered a rendered video.

s40
00:04:03.400 --> 00:04:11.760
And but yes, the the infrastruc- from infrastructural standpoint, it's pretty similar, but there are a couple of differences.

s41
00:04:11.760 --> 00:04:17.840
Uh and this is actually the most interesting to me, generating versus editing.

s42
00:04:17.840 --> 00:04:22.320
At RealFull, we're focusing on editing real footage.

s43
00:04:22.320 --> 00:04:24.760
So, we do not generate a lot of content.

s44
00:04:24.760 --> 00:04:33.040
We are expecting you to provide your real life, your personal content, and we will edit it for you.

s45
00:04:33.040 --> 00:04:43.160
And actually, this is a more complex problem because if the agent has a blank blank canvas, it can do whatever they can.

s46
00:04:43.160 --> 00:04:50.120
But in the editing case, the agent has to figure out which moments are the best.

s47
00:04:50.120 --> 00:04:55.080
Uh what to omit, what to use, how to organize everything together.

s48
00:04:55.080 --> 00:05:11.600
And also, um sometimes footage can be messy or incomplete, and agent still has to deliver a very polished result, professionally made, so that ideally the viewers of this content

s49
00:05:11.600 --> 00:05:17.600
don't get if it is like AI or human edited.

s50
00:05:17.680 --> 00:05:22.440
So, let's actually have a look how we do it at RealFull.

s51
00:05:22.440 --> 00:05:34.120
So, we start, as I already mentioned, with your media plus a prompt, some directions like how you want it to be edited, and we need to get a polished

s52
00:05:34.120 --> 00:05:35.040
clip.

s53
00:05:35.040 --> 00:05:37.480
So, let's go through it step by step.

s54
00:05:37.480 --> 00:05:41.080
So, we are doing first media understanding.

s55
00:05:41.080 --> 00:05:47.240
We need to understand what's what's actually happening uh, on those clips and photos.

s56
00:05:47.240 --> 00:05:54.080
And we also need to transcribe transcribe speech, for example, in the case if you have speak-to-camera videos.

s57
00:05:54.080 --> 00:06:06.560
Then, we are providing a creative plan for the user so that they can approve if they like it or not, what they want to change or maybe regenerate, uh, and we

s58
00:06:06.560 --> 00:06:10.560
create this plan before actually starting editing.

s59
00:06:10.560 --> 00:06:22.000
Once the user approved this plan, uh, we spin up a sandbox, the remote uh, remote machine that we already discussed, and this is an environment for the agent to,

s60
00:06:22.000 --> 00:06:24.160
uh, execute everything.

s61
00:06:24.160 --> 00:06:35.280
So, the agent comes with the skills, and in our case, in in the case of, uh, video editing, our skills are, for example, cut rules, how, for example, how to select

s62
00:06:35.280 --> 00:06:48.240
the best moments, uh, also font pairs, which fonts are, uh, more suitable for this use case, which are not, for example, how to generate B-rolls, and this is where our taste and craft,

s63
00:06:48.240 --> 00:06:50.400
uh, live, actually.

s64
00:06:50.400 --> 00:07:05.720
Um, and then also agent, uh, can, um, can initiate some other sub-processes, for example, generating music that will fit this exact composition, generating voice-over, adding sounds, animating images.

s65
00:07:05.720 --> 00:07:15.440
Yes, this is actually what we do, uh, if you provide photos, we can animate your photos to make them more, uh, dynamic and engaging.

s66
00:07:15.440 --> 00:07:18.240
And then comes Remotion composition.

s67
00:07:18.240 --> 00:07:19.960
So, here a little bit of background.

s68
00:07:19.960 --> 00:07:20.919
What's Remotion?

s69
00:07:20.919 --> 00:07:30.160
Remotion is a framework, open-source open-source framework, uh, to create videos as code, as React code.

s70
00:07:30.160 --> 00:07:39.720
Uh, so basically, it's just like a a file with the order with all your assets and tracks and how they're following each other.

s71
00:07:39.720 --> 00:07:42.480
And, why it is important?

s72
00:07:42.480 --> 00:07:51.160
Because, uh, agents are really good at writing code and therefore we can use them to create videos with this remotion framework.

s73
00:07:51.160 --> 00:07:55.280
And then the last thing is the verification layer.

s74
00:07:55.280 --> 00:08:08.640
Of course agent can make mistakes and that's why we develop this verification layer to make sure that all the the composition is clean, is well defined, everything will be rendered

s75
00:08:08.640 --> 00:08:15.160
and if there are there are some problems then the agent will reiterate on the composition.

s76
00:08:15.160 --> 00:08:19.115
And this is how we got to a polished clip.

s77
00:08:19.115 --> 00:08:19.680
[sighs and gasps]

s78
00:08:19.680 --> 00:08:21.640
So it's a lot, right?

s79
00:08:21.640 --> 00:08:31.080
It's like very complex workflow and ideally we don't want our users to even know anything about it.

s80
00:08:31.080 --> 00:08:40.240
And this is even maybe a bigger problem how to deliver this complex agentic workflow to mass consumer.

s81
00:08:40.240 --> 00:08:43.159
And this is how we're tackling that at Real Flow.

s82
00:08:43.159 --> 00:08:53.600
So we decided to go mobile first so that users can edit videos videos while driving, walking or maybe lifting weights.

s83
00:08:53.600 --> 00:08:59.920
Also, I know that prompting videos can sometimes be also challenging.

s84
00:08:59.920 --> 00:09:03.240
That's why we create directional templates.

s85
00:09:03.240 --> 00:09:15.520
For example, like speak to camera in videos or maybe you want to add B-rolls or voiceover so that users can just select these directional templates, drop their media and that's it.

s86
00:09:15.520 --> 00:09:19.480
Even without any prompt it will it will work.

s87
00:09:19.480 --> 00:09:23.040
And the third thing is a building editor.

s88
00:09:23.040 --> 00:09:23.400
Why?

s89
00:09:23.400 --> 00:09:28.720
Because we want to make this experience convenient and familiar for users.

s90
00:09:28.720 --> 00:09:38.600
So a lot of people are already sort of using regular video editors and that's why we want to provide this experience as well.

s91
00:09:38.600 --> 00:09:39.720
So, how it works?

s92
00:09:39.720 --> 00:09:53.440
User first generates a video agentically, but if they want to tweak it, for example, remove a second or maybe correct some word in the captions, they can go into building editor and

s93
00:09:53.440 --> 00:09:55.640
edit it a little bit.

s94
00:09:55.640 --> 00:10:03.320
Um Yeah, and actually I have a couple of examples here that I recently created with Real Fill.

s95
00:10:03.320 --> 00:10:05.920
I will play them just maybe one of that.

s96
00:10:05.920 --> 00:10:08.000
Oh, sorry.

s97
00:10:08.200 --> 00:10:09.720
Last week I was invited

s98
00:10:09.720 --> 00:10:12.040
Do do you hear?

s99
00:10:12.040 --> 00:10:12.560
A little bit.

s100
00:10:12.560 --> 00:10:15.120
Okay, you just can enjoy the video.

s101
00:10:15.120 --> 00:10:20.320
because of course our creators dinner and honestly it was one of the best event experiences I've had.

s102
00:10:20.320 --> 00:10:22.520
First of all, the venue was stunning.

s103
00:10:22.520 --> 00:10:26.400
It was a custom event space transformed into this tropical sunset feast.

s104
00:10:26.400 --> 00:10:29.120
The whole atmosphere felt so warm and cinematic.

s105
00:10:29.120 --> 00:10:31.720
Second, the food was way beyond my expectations.

s106
00:10:31.720 --> 00:10:38.920
They brought in private chefs from LA and we had lobster, blue crab, and this incredible ice cream that I'm still thinking about.

s107
00:10:38.920 --> 00:10:42.120
But most importantly, the conversations were so much fun.

s108
00:10:42.120 --> 00:10:43.480
It was kind of atmosphere.

s109
00:10:43.480 --> 00:10:48.920
It felt really easy to connect with people, talk about what they're working on, and just enjoy the community.

s110
00:10:48.920 --> 00:10:50.760
And lastly, we got gifts.

s111
00:10:50.760 --> 00:10:52.520
One of the highlights

s112
00:10:52.520 --> 00:11:04.800
So, yeah, basically all these videos they were assembled only using agent, no regular video editor, and I already posting them on social media.

s113
00:11:04.800 --> 00:11:08.760
And yeah, it I I have a lot of fun with that.

s114
00:11:08.760 --> 00:11:11.080
And oh, sorry.

s115
00:11:11.080 --> 00:11:22.000
And exclusively for this conference, we are giving our better, which is our new new second version.

s116
00:11:22.000 --> 00:11:28.440
Please give it a try and let me know if you have any feedback.

s117
00:11:28.440 --> 00:11:30.400
Here is my email.

s118
00:11:30.400 --> 00:11:33.200
Please Uh free to reach out.

s119
00:11:33.200 --> 00:11:34.560
Uh we're still early.

s120
00:11:34.560 --> 00:11:37.120
We're actively working on it.

s121
00:11:37.120 --> 00:11:44.000
Uh so we will have uh we will be happy to hear any feedback and also curious how you use it.

s122
00:11:44.000 --> 00:11:46.920
And also um some exciting news.

s123
00:11:46.920 --> 00:11:50.600
We recently got funded by A16Z speed run.

s124
00:11:50.600 --> 00:11:55.000
Uh so I'm very excited to uh continue working.

s125
00:11:55.000 --> 00:11:55.923
Um

s126
00:11:55.923 --> 00:11:57.120
[applause]

s127
00:11:57.120 --> 00:11:58.200
Yeah, that's it.

s128
00:11:58.200 --> 00:11:59.160
Thank you so much.

s129
00:11:59.160 --> 00:12:09.400
And because it's a presentation about content and how to edit videos, I have to uh film a video with you all.

s130
00:12:12.720 --> 00:12:14.720
Second.

s131
00:12:25.080 --> 00:12:27.080
Yay!

s132
00:12:42.428 --> 00:12:44.428
[music]