The AI-Powered Tools Supercharging Your Imagination | Bilawal Sidhu | TED

翻译人员: Yip Yan Yeung

All right. Good afternoon, y’all.

大家下午好。

Let's talk about blending reality and imagination.

我们来聊一聊现实与想象的融合。

But first, let's take a step back in time to 2001.

但首先，让我们倒退到 2001 年。

As an 11-year-old in India,

我这个 11 岁的印度少年

I became obsessed with computer graphics and visual effects.

爱上了计算机图形学和视觉效果。

Of course, at that age,

当然，在那个年代，

it meant making cheesy videos kind of like this.

等同于做做这种中二视频。

But therein started a foundational theme in my life,

但也从此铺就了我的人生底色：

the quest to blend reality and imagination.

追求现实与想象的融合。

And that quest has stayed with me

这个追求伴随着我，

and permeated across my decade-long career in tech,

贯穿了我在科技领域 十年的职业生涯，

working as a product manager at companies like Google

在谷歌这样的公司 担任产品经理，

and as a content creator on platforms like YouTube and TikTok.

也在 YouTube 和 TikTok 这样的平台担任内容创作者。

So today, let's deconstruct this quest to blend reality and imagination

今天，我们来解构一下 对融合现实与想象的追求，

and explore how it’s getting supercharged -- buzzword alert --

探索如何为它添砖加瓦， 流行词预警——

by artificial intelligence.

借助人工智能的力量。

Let's start with the reality bit.

从现实开始吧。

You probably heard about photogrammetry.

你可能听说过摄影测量学。

It's the art and science of measuring stuff in the real world

它是通过照片和其他传感器 测量现实事物的艺术和科学。

using photos and other sensors.

What required massive data centers and teams of experts in the 2000s

在 21 世纪初需要 大量数据中心和专家的工作

became increasingly democratized by the 2010s.

到了 2010 年以后 就变得越来越触手可及。

Then, of course, machine learning came along

当然，之后机器学习的发展

and took things to a whole new level

将一切提升到了全新的高度，

with techniques like neural radiance fields, or NeRFs.

利用了神经辐射场 （NeRF）这样的技术。

What you're seeing here is an AI model

你眼前的是一个 AI 模型，

creating a ground-up volumetric 3D representation

仅凭 2D 图像 生成立体 3D 呈现。

using 2D images alone.

But unlike older techniques for reality capture,

与传统用于现实捕捉的技术不同的是，

NeRFs do a really good job of encapsulating the sheer complexity

NeRF 相当擅长捕捉现实的 微小细节和细微差别。

and nuance of reality.

The vibe, if you will.

也可以说是那个“味儿”。

Twelve months later, you can do all of this stuff

12 个月以后，你就能用 你口袋里的 iPhone 做到这一切，

using the iPhone in your pocket,

using apps like Luma.

使用 Luma 这样的 app。

It's like 3D screenshots for the real world.

就像是现实世界的 3D 截屏。

Capture anything once and reframe it infinitely in postproduction,

捕捉一次，然后 在后期制作中改造无数次，

so you can start building that collection of spaces, places and objects

你就能打造出你真正在乎的 空间、地点和物体的集合，

that you truly care about

and conjure them up in your future creations.

在你未来的创作中召唤它们。

So that's the reality bit.

说完了现实。

As NeRFs were popping off last year,

NeRF 从去年风靡起来，

the AI summer was also in full effect,

AI 之夏也如火如荼，

with Midjourney, DALL-E 2, Stable Diffusion

Midjourney、DALL-E 2、 Stable Diffusion，

all hitting the market around the same time.

几乎在同一时间涌现在市场之中。

But what I fell in love with was inpainting.

但我心仪的是图像修复。

This technique allows you to take existing imagery

这项技术能让你使用现有图像，

and augment it with whatever you like,

增强成你想要的任何样子，

and the results are photorealistically fantastic.

结果逼真得要命。

It blew my mind

我觉得惊为天人，

because stuff that would have taken me like three hours in classical workflows

因为在传统工作流程中 需要花上我三个小时的工作

I could pull off in just three minutes.

只需要三分钟就搞定了。

But I wanted more.

但我想要的还不止这些。

Enter ControlNet,

来看 ControlNet，

a game-changing technique by Stanford researchers

由斯坦福大学研究者 研发的变革性技术，

that allows you to use various input conditions

你可以使用各种输入条件

to guide and control the AI image generation process.

引导、控制 AI 生成图像的过程。

So in my case, I could take the depth information

以我自己的情况为例， 我可以使用 3D 扫描的

and the texture detail from my 3D scans

深度和材质信息

and use it to literally reskin reality.

为现实世界重新涂装。

Now, this isn't just cool video.

不仅仅是炫酷的视频。

There’s a lot of useful use cases, too.

还有很多实用的用途。

For example, in this case I'm taking a 3D scan of my parents' drawing room,

比如，我用我父母的 “画室”的 3D 扫描，

as my mother likes to call it,

我母亲喜欢这么称呼这间房间，

and reskinning it to different styles of Indian decor

给它重新布置上 各式各样的印度装饰，

and doing so while respecting the spatial context and the layout

同时考虑到了这个室内空间的 空间环境和布局。

of the interior space.

If you squint, I'm sure you can see

如果你细看，我相信你会发现

how this is going to transform architecture

这会如何永久地改变 建筑和室内设计。

and interior design forever.

You could take that 2016 scan of a Buddha statue

你可以将一尊佛像的 2016 年扫描件

and reskin it to be gloriously golden

涂成金灿灿的，

while pulling off these impossible camera moves

没有别的方式可以让你 这么调整摄影角度。

you just couldn't do any other way.

Or you could take that vacation footage from your trip to Tokyo

你也可以用上 东京之行的度假录影，

and bring these cherry blossoms to life in a whole new way.

用全新的方式重现樱花。

And let me tell you, cherry blossoms look really good during the day,

我想告诉大家， 白天看樱花非常漂亮，

but they look even better at night.

但晚上看更漂亮。

Oh, my God. They sure are glowing.

天哪，它们在发光。

It's almost like this dreamlike quality

你可以得到这种梦幻般的高质量，

where you can use AI to accentuate the best aspects of the real world.

借助于 AI 突出现实世界的美好一面。

Natural landscapes look just as beautiful.

自然景观风光绮丽。

Like this waterfall that could be on another planet.

就像这道可以出现在 另一个星球上的瀑布。

But of course, you could go over the hills and far away

当然，你也可以 翻山越岭，长途跋涉，

to the French Alps from another dimension.

从另一个维度 前往法国阿尔卑斯山。

But it's not just static scenes.

但这不仅限于静态的场景。

You can do this stuff with video, too.

你可以将其用于视频。

I can't wait till this technology is running at 30 frames per second

我迫不及待想看到 这项技术以每秒 30 帧地速度运行，

because it's going to transform augmented reality and 3D rendering.

因为它将改变 增强现实和 3D 渲染。

I mean, how soon until we're channel-surfing realities

我们再过不久就能在现实的基础上 跨越不同现实的“频道”了吧？

layered on top of the real world?

Of course, just like reality capture got democratized,

当然，正如现实捕捉的普及，

all these tools from last year are getting even easier.

这些去年出现的工具 也越来越容易使用。

So instead of me spending hours

不再由我花上日夜 把各种工具组合到一起，

weaving together a bunch of different tools,

tools like Runway and Kaiber let you do exactly the same stuff

而像 Runway 和 Kaiber 这样的工具已经能让你

with just a couple clicks.

在点击之间达到同样的效果。

Want to go from day to night? No problemo.

把白天换成夜晚？没问题。

Want to get that retro 90s aesthetic from "Full House"?

想渲染一个 90 年代 《欢乐满屋》复古风？

You can do that too.

当然可以做到。

But it goes beyond reality capture.

还能超越现实捕捉的范畴。

Companies like Wonder Dynamics

像 Wonder Dynamics 这样的公司

are turning video into this immaculate form of performance capture

将视频转化为这种 精美的动作捕捉形式，

so you can embody fantastical creatures using the phone in your pocket.

让你可以用口袋里的手机 展现神奇生物。

This is stuff that James Cameron only dreamt about in the 2000s.

这是詹姆斯·卡梅隆 （James Cameron）

在 21 世纪初 只敢想一想的东西。

And now you could do it with your iPhone? That’s absolutely wild to me.

你现在都可以用 iPhone 实现了？我觉得太牛了。

So when I look back at the past two decades

我在回望过去的 20 年时，

and this ill-tailored tapestry of tools that I've had to learn,

看到我得学习的工具大杂烩，

I feel a sense of optimism for what lies ahead

我对下一代创作者眼前的未来 感到了一丝乐观。

for the next generation of creators.

The 11-year-olds of today

现在的 11 岁孩子 不用担心这堆烂事。

don't have to worry about all of that crap.

All they need to do is have a creative vision

他们只需要拥有 发挥创造力的畅想、

and a knack for working in concert with these AI models,

与这些 AI 模型协同工作的技能，

these AI models that are truly a distillation

这些 AI 模型真的是 人类智慧和创造力的精华。

of human knowledge and creativity.

And that's a future I'm excited about,

这是我期待的未来，

a future where you can blend reality and imagination

一个可以用你值得信赖的 AI 助手 融合现实和想象的未来。

with your trusty AI copilot.

Thank you very much.

谢谢。

(Applause)

（掌声）