Name: AI 爭霸戰正式開打！OpenAI GPT-4o vs Google I/O 誰會贏得這場戰役？ (Another glorious battle for AI dominance… GPT-4o vs Google I/O)
Uploaded: 2024-05-16T02:50:21.000Z
Duration: 4 min 40 s

情態副詞

Yesterday was Google I.O., the annual developer conference where Google desperately tries to catch up to its artificial rival, OpenAI.

昨天是谷歌 I.O. 年度開發者大會，谷歌在會上拼命追趕其人工智能競爭對手 OpenAI。

Google announced some crazy stuff I never thought I would see in my lifetime, like a SQL database for Firebase.

谷歌宣佈了一些我從未想過有生之年會看到的瘋狂舉動，比如為 Firebase 提供 SQL 數據庫。

More on that later, because first we need to talk about the biggest announcement at I.O.

稍後再談，因為我們首先要談談 I.O. 的最大公告。

You see, OpenAI hype lord Sam Altman yet again wrapped up Sundar in a wet blanket by releasing GPT-4-O just hours before Google I.O., which is a total coincidence and definitely not designed to troll Google.

你看，OpenAI 的炒作者山姆-奧特曼（Sam Altman）在谷歌 I.O. 發佈前幾個小時就發佈了 GPT-4-O，這完全是個巧合，絕對不是為了嘲弄谷歌而設計的。

In today's video, we'll break down this artificial beef, but more importantly, look at all kinds of crazy new technology released in just the last 48 hours.

在今天的視頻中，我們將分解這種人造牛肉，但更重要的是，我們將看到過去 48 小時內發佈的各種瘋狂的新技術。

It is May 15th, 2024, and you're watching The Code Report.

現在是 2024 年 5 月 15 日，您正在收看《密碼報告》。

On Monday, OpenAI had a surprise spring update where they unveiled their new flagship model, GPT-4 Omni.

本週一，OpenAI 在春季新品發佈會上發佈了新的旗艦機型 GPT-4 Omni。

You've got me on the edge of my... well, I don't really have a seat, but you get the idea.

你讓我坐立不安......好吧，我並沒有座位，但你明白我的意思。

Which is faster and cheaper than GPT-4 Turbo and combines text, vision, and audio into a single model.

它比 GPT-4 Turbo 更快、更便宜，並將文本、視覺和音頻整合到一個模型中。

What was most impressive, though, was its human-like conversational abilities.

不過，最令人印象深刻的還是它類似人類的對話能力。

Well, well, well, just when I thought things couldn't get any more interesting.

好了，好了，好了，就在我以為事情不會變得更有趣的時候。

Talking to another AI that can see the world.

與另一個能看到世界的人工智能對話。

By default, it uses a California Valley Girl accent set to maximum cringe, but the tone of the voice can vary from dramatic to sarcastic to super chill for bedtime stories.

默認情況下，它使用的是加州山谷女孩口音，並將其設置為最令人討厭的程度，但語音語調可以從戲劇性到諷刺性，再到適合講睡前故事的超級冷酷的語調不等。

Bedtime story about robots and love? I got you covered!

關於機器人和愛情的睡前故事？我為你準備好了！

This technology will be a huge leap forward for your AI girlfriend, and you can use the GPT-4 Om model today, but the conversational part of it is still not available to the public.

這項技術對於你的人工智能女友來說將是一個巨大的飛躍，你現在就可以使用 GPT-4 Om 模型，但它的對話部分仍未向公眾開放。

That's disappointing, but what you also need to know is that OpenAI is in talks to put their technology on the iPhone, but Google also wants to get its flagship model on the iPhone as well.

這令人失望，但你還需要知道的是，OpenAI 正在洽談將其技術應用到 iPhone 上，但谷歌也希望將其旗艦機型應用到 iPhone 上。

Talks are ongoing to also get Gemini on the iPhone, so these companies are competing to create a model that's smart, but also fast and cheap enough to run on mobile.

是以，這些公司都在競相創造一種既智能又快速、廉價的模式，以便在移動設備上運行。

In order to get that massive bag from Apple.

為了從蘋果公司拿到那個大袋子。

Yesterday at I.O., Google demoed something called Project Astra, which feels similar to 4 Omni.

在昨天的 I.O. 大會上，谷歌演示了一個名為 "Project Astra "的項目，感覺與 4 Omni 很相似。

Do you remember where you saw my glasses?

你還記得在哪裡看到我的眼鏡嗎？

Yes, I do. Your glasses were on the desk near a red apple.

是的，我知道。你的眼鏡就在桌子上，靠近一個紅蘋果。

It's cool, but there's more latency and the voice is more robotic compared to OpenAI.

它很酷，但與 OpenAI 相比，延遲更長，語音也更像機器人。

Now, what's also very interesting is that OpenAI just parted ways with Ilya, their former chief scientist and co-founder, who many people used to worship as the brains behind OpenAI.

現在，同樣有趣的是，OpenAI 剛剛與前首席科學家兼聯合創始人伊利亞分道揚鑣，很多人都曾把伊利亞奉為 OpenAI 的幕後智囊。

There's definitely some underlying drama here, but we likely won't know the truth until they release their memoirs in the 2040s.

這其中肯定有一些隱情，但我們可能要等到 2040 年他們出版回憶錄時才能知道真相。

But now let's finally talk about Google I.O.

但現在讓我們來談談 Google I.O.。

The biggest AI announcement from Google was Gemini 1.5 Pro, which can now handle a 2 million token context window.

谷歌發佈的最大人工智能產品是 Gemini 1.5 Pro，它現在可以處理 200 萬個令牌上下文窗口。

That could be two hours of video content or 60,000 lines of code.

這可能是兩個小時的視頻內容，也可能是 60,000 行代碼。

That's a lot of context, but tokens can be expensive.

這是一個很大的背景，但代幣可能很昂貴。

And to address that, they released a new feature called context caching that can reuse tokens for a fraction of the cost.

為了解決這個問題，他們發佈了一項名為上下文緩存的新功能，可以以極低的成本重複使用令牌。

In addition, Google launched a competition for developers, and whoever builds the best Gemini-powered app wins an electric DeLorean.

此外，谷歌還為開發者舉辦了一場競賽，誰能開發出最好的雙子座應用程序，誰就能贏得一輛電動迪羅裡安（DeLorean）。

To make building this app easier, they also released a new tool called Firebase GenKit, which is integrated with Olama and makes it easy to build AI-enabled API endpoints.

為了讓這款應用的構建更加簡單，他們還發布了一款名為 Firebase GenKit 的新工具，該工具與 Olama 集成，可以輕鬆構建支持人工智能的 API 端點。

In addition, Project IDX is now open to the public, which is a browser-based VS code that's also integrated with things like mobile emulators.

此外，Project IDX 現已向公眾開放，這是一個基於瀏覽器的 VS 代碼，還與移動模擬器等設備集成。

By far the most exciting thing for me, though, is a new tool called Firebase Data Connect, which officially brings Postgres into Firebase.

不過到目前為止，最讓我興奮的是一款名為 Firebase Data Connect 的新工具，它正式將 Postgres 引入 Firebase。

This has been the number one most requested feature for years.

這是多年來人們要求最多的功能。

And its absence has led to startups like Supabase, which is branded as a Firebase alternative.

由於它的缺失，出現了像 Supabase 這樣的初創公司，它被打造成 Firebase 的替代品。

但現在是 2024 年，轉捩點已經到來。

Firebase is now the Supabase alternative.

Firebase 現在是 Supabase 的替代品。

I'm a big fan of both Supabase and Firebase, and if you want to learn these technologies, check out my full courses on Fireship.io, and stay tuned for a full tutorial on Data Connect on my second channel, Beyond Fireship, soon.

我是 Supabase 和 Firebase 的忠實粉絲，如果你想學習這些技術，請查看我在 Fireship.io 上的完整課程，並敬請期待我在第二個頻道 Beyond Fireship 上的數據連接完整教程。

Google also announced some new hardware, like Trillium TPUs, and Axion, its new ARM-based CPUs for data centers.

谷歌還發布了一些新硬件，如 Trillium TPU 和 Axion，後者是谷歌為數據中心開發的基於 ARM 的新 CPU。

And finally, Google also announced Veo, a generative video model to compete with OpenAI Sora.

最後，谷歌還宣佈推出生成式視頻模型 Veo，與 OpenAI Sora 競爭。

It's extremely impressive compared to where we were just a year ago, but yet again, it just feels one step behind OpenAI.

與一年前相比，這已經非常了不起了，但還是感覺比 OpenAI 慢了一步。

We just looked at all kinds of crazy new game-changing technology, but at this point, I'm feeling a little disappointed with our progress towards the singularity.

我們剛剛看到了各種瘋狂的改變遊戲規則的新技術，但此時此刻，我對我們邁向奇點的進程感到有些失望。

It's been over a year since GPT-4, and unfortunately, I still have a job.

GPT-4 已經過去一年多了，不幸的是，我仍然有工作。

4Omni, Clod, and Gemini 1.5 all seem to be pretty maxed out on how far they can get with these benchmarks.

4Omni、Clod 和 Gemini 1.5 在這些基準測試中似乎都達到了極限。

Making models faster and cheaper is great, but if they're not becoming more intelligent, then the singularity is nowhere in sight.

讓模型變得更快、更便宜固然很好，但如果它們不能變得更智能，那麼奇點就遙遙無期了。

They've already absorbed almost all the information humans have created, so unless there's a major breakthrough that makes AI actually intelligent and able to learn independently, it sure looks like we're standing on the edge of a plateau, and the only place to go is the trough of disillusionment.

它們已經吸收了人類創造的幾乎所有資訊，是以，除非有重大突破能讓人工智能真正具有智能並能夠獨立學習，否則我們看起來肯定是站在了高原的邊緣，而唯一的去處就是幻滅的谷底。

 Thanks for watching, and I will see you in the next one.

感謝您的收看，我們下期再見。