見出し画像

【GPT-4o:人間のように会話できる新モデルが無料で】英語解説を日本語で読む【2024年5月14日|@Matt Wolfe】

OpenAIは2024年5月13日、新たな言語モデル「GPT-4o」を発表しました。GPT-4oは、従来のGPT-4よりも高速で、画像理解や多言語対応にも優れています。さらに、無料ユーザーでもGPT-4レベルの性能を利用できるようになりました。GPT-4oは、音声会話の低レイテンシーとマルチモーダル機能の向上が特徴です。無料ユーザーもGPT-4oを利用でき、有料ユーザーはより多くの利用が可能になります。GPT-4oは、ChatGPTのデスクトップアプリでも利用可能で、Mac版のデモが公開されました。APIを通じて開発者も利用できます。OpenAIは、Googleの発表と比較して、GPT-4oのデモがリアルタイムであることを強調しています。GPT-4oの主な機能は、リアルタイムの会話音声です。映画「Her」を思わせる自然な会話が可能で、ユーザーの呼吸や感情も認識できます。また、ロボットの声や歌声など、多様な表現力を示しました。GPT-4oはビジョン機能も備えており、数式の解決をサポートしたり、コードの説明や修正の提案ができます。さらに、リアルタイムの翻訳機能や、表情から感情を読み取る機能も披露されました。
公開日:2024年5月14日
※動画を再生してから読むのがオススメです。


Today, May 13th marks the beginning of a really interesting couple of weeks.

本日5月13日は、非常に興味深い数週間の始まりを告げます。

Today, we got some new announcements out of OpenAI, which came right before Google's making announcements tomorrow.

本日、OpenAIから新しい発表がありましたが、それは明日Googleが発表する前に行われました。

For whatever reason, OpenAI loves to do that.

何かの理由で、OpenAIはそうするのが好きなようです。

They love to pick the timing of their events to try to overshadow Google as much as possible.

彼らは自分たちのイベントのタイミングを選んで、可能な限りGoogleを影で引き立てるのが好きです。

Today's OpenAI event seems to be exactly what they were trying to do.

今日のOpenAIのイベントは、まさに彼らがやろうとしていたことのようです。

I'm actually out here in Mountain View, California right now because I'm going to be at the Google keynote tomorrow, but wanted to make sure that I got a video out as quickly as possible about what OpenAI just announced with their new model.

実際、私は現在カリフォルニア州マウンテンビューにいますが、明日のGoogleの基調講演に出席する予定です。ただし、OpenAIが新しいモデルを発表した内容についてできるだけ早くビデオを公開したかったのです。

Instead of calling this new model GPT-4.5 or GPT-5, they went with GPT-4o.

この新しいモデルをGPT-4.5やGPT-5と呼ぶ代わりに、彼らはGPT-4oを選びました。

As it turns out, that mysterious chatbot that we've been playing with called gpt2-chatbot, and I'm a good gpt2-chatbot, and I'm also a good gpt2-chatbot, we're all actually us getting the opportunity to test this new GPT-4o.

実際、私たちが遊んでいる謎のチャットボットであるgpt2-chatbot、そして私は良いgpt2-chatbotであり、私も良いgpt2-chatbotであり、実際にはこの新しいGPT-4oをテストする機会を得ています。

While the model is improved a little bit, the biggest features about this model are the lower latency when having voice conversations, it seemingly has a better multimodal capabilities, and instead of me breaking it all down, let's watch little bits of the keynote and sort of talk through some of these new announcements that OpenAI just made.

モデルは少し改善されていますが、このモデルの最大の特徴は、音声会話時の低遅延、より優れた多様な能力を持っているようであり、私がすべてを説明する代わりに、基調講演の一部を見て、OpenAIが行ったこれらの新しい発表について話し合いましょう。

The big news today is that we are launching our new flagship model, and we are calling it GPT-4o.

私たちの新しいフラッグシップモデルを発売するという大きなニュースです。それをGPT-4oと呼んでいます。

The special thing about GPT-4o is that it brings GPT-4 level intelligence to everyone, including our free users.

GPT-4oの特別な点は、GPT-4レベルの知能を無料ユーザーを含めて誰にでももたらすことです。

This is actually a big piece of news here.

これは実際にここでの大きなニュースです。

Up until now, if you were on the free version of ChatGPT, you were using GPT-3.5.

これまで、ChatGPTの無料版を使用している場合、GPT-3.5を使用していました。

Well, this new state of the art model GPT-4o is now gonna be available both for plus and free users, which means anybody can use this new state of the art model completely for free.

さて、この最新鋭のモデルGPT-4oは、プラス会員と無料会員の両方に利用可能になります。つまり、誰でもこの最新鋭のモデルを完全に無料で使用できるということです。

But if you're a plus member, you just get to use it a lot more.

ただし、プラス会員の場合は、より多く使用できます。

A very important part of our mission is to be able to make our advanced AI tools available to everyone for free.

私たちの使命の非常に重要な部分は、先進的なAIツールを誰にでも無料で提供できるようにすることです。

Today we're also bringing the desktop app to ChatGPT.

今日はChatGPTにデスクトップアプリも持ってきています。

This is another big update that they made during this keynote is that we now get a desktop app.

これは彼らがこの基調講演中に行ったもう1つの大きなアップデートで、今やデスクトップアプリを手に入れることができます。

In their demos that they show here in this video, they're only showing the desktop app being used on Mac.

彼らがこのビデオでここで示しているデモでは、Macでデスクトップアプリが使用されているだけです。

They didn't really say if it was Mac only, or if it was gonna be Mac and PC.

それがMac専用なのか、それともMacとPCの両方になるのか、彼らはあまり言及していませんでした。

My guess is that it's probably going to come out for both platforms.

私の推測では、おそらく両プラットフォーム向けにリリースされるでしょう。

In the demo, they only showed Mac and they didn't really speak to that piece of it.

デモでは、Macのみを示し、それについてはあまり話していませんでした。

As you can see, it's easy, it's simple.

ご覧の通り、簡単でシンプルです。

It integrates very, very easily in your workflow.

あなたのワークフローに非常に簡単に統合されます。

GPT-4o provides GPT-4 level intelligence, but it is much faster and it improves on its capabilities across text, vision and audio.

第4世代のGPTは、GPT-4レベルの知能を提供しますが、はるかに高速であり、テキスト、ビジョン、オーディオの能力を向上させています。

We're very, very excited to bring GPT-4o to all of our free users out there.

我々は、GPT-4oをすべての無料ユーザーに提供できることに非常に興奮しています。

For the paid users, they will continue to have up to five times the capacity limits of our free users.

有料ユーザーには、無料ユーザーの容量制限の最大5倍まで維持されます。

I'm gonna go back real quick on the screen here.

ここで画面をすばやく戻ります。

These are all the things now up here on the top right that ChatGPT free users are gonna get access to.

これらは、ChatGPTの無料ユーザーがアクセスできるすべての機能です。

You're gonna be able to access the GPT store and all the custom GPTS.

あなたはGPTストアとすべてのカスタムGPTSにアクセスできるようになります。

Vision, the browse model, which allows you to search the internet using ChatGPT, the memory functions and advanced data analysis, which used to be called Code Interpreter.

ビジョン、ブラウズモデル、インターネットをChatGPTを使用して検索できる機能、メモリ機能、以前はCode Interpreterと呼ばれていた高度なデータ分析が含まれます。

Free ChatGPT members get that.

無料のChatGPTメンバーはそれを受け取ります。

GPT-4o is not only available in ChatGPT, we're also bringing it to the API.

ChatGPTで利用可能なだけでなく、APIでも提供されています。

Basically when they're saying they're also bringing it to the API, they mean that developers can actually work with this new model as well.

基本的に、彼らがAPIにも導入すると言っているとき、開発者もこの新しいモデルで実際に作業できるということです。

However, that means that you can play with this model directly inside of the OpenAI Playground as well.

ただし、それはつまり、OpenAI Playgroundの中でもこのモデルと直接遊ぶことができるということです。

If you go over to platform.openai.com/Playground and click on chat in the left sidebar, you can see on this little dropdown, we actually have GPT-4o over here.

platform.openai.com/Playgroundに移動して、左サイドバーのチャットをクリックすると、この小さなドロップダウンで、実際にGPT-4oがここにあるのが見えます。

One thing that's interesting to note is inside of the Playground, we actually have the ability to upload images or link to images.

注目すべき興味深い点の1つは、Playgroundの中で、画像をアップロードしたり、画像へのリンクを貼ったりする能力が実際にあるということです。

I don't believe that was ever available inside of OpenAI's Playground before.

OpenAIのPlaygroundの中でこれが以前利用可能だったかどうかはわかりません。

This is the first time I'm noticing it.

これが私がそれに気づいた最初の時です。

Maybe it was there before, but I'm pretty sure they just added the ability to upload images here with this new rollout of GPT-4o.

以前そこにあったかもしれませんが、新しいGPT-4oの展開でここに画像をアップロードする機能を追加したと確信しています。

Our developers can start building today with GPT-4o and making amazing AI applications, deploying them at scale.

私たちの開発者は、GPT-4oを使って今日から素晴らしいAIアプリケーションを構築し、スケールで展開することができます。

4.0 is available at 2X faster, 50% cheaper, and five times higher rate limits compared to GPT-4 Turbo.

4.0は、GPT-4 Turboと比較して2倍速く、50%安く、5倍の高いレート制限で利用可能です。

We'll do some live demos.

私たちはいくつかのライブデモを行います。

I love that they're doing these demos live as well, because I think it's sort of a little bit of a message to Google.

彼らがこれらのデモをライブで行っているのが好きです。なぜなら、それはGoogleに対する少しのメッセージだと思うからです。

When Google put out their launch video for Gemini, they showed off all of these really cool capabilities and a nice pre-recorded polished video.

GoogleがGeminiのローンチビデオを公開したとき、彼らは本当にクールな機能を披露し、素敵な事前に録画された洗練されたビデオを見せました。

It turns out that a lot of what they were showing was not actually real time.

実際には、彼らが示していた多くは実際にはリアルタイムではなかったことがわかりました。

They made it look like it was doing these things a lot faster than it actually was capable of doing them.

彼らは、それが実際にはできることよりもはるかに速く行われているように見せかけました。

OpenAI is saying, hey, check this out.

OpenAIは言っています、「ねえ、これをチェックしてみて」。

We're gonna show it to you live in real time.

私たちはそれをリアルタイムでライブでお見せします。

No camera trickery.

カメラのトリックはありません。

This is how it works.

こういう仕組みです。

They also kind of made that little nudge in their blog post as well about GPT-4o.

彼らはまた、GPT-4oについてのブログ投稿でも少し触れています。

All videos on this page are at 1X real time.

このページのすべてのビデオは1倍の実時間です。

That note there was very intentional.

そのメモは非常に意図的なものでした。

It's a message to people who are comparing it to Google's launch.

それはGoogleのローンチと比較している人々へのメッセージです。

Guaranteed, that's why that's there.

間違いなく、それがそこにある理由です。

Hi, I'm Barrett.

こんにちは、私はバレットです。

Hey, I'm Mark.

こんにちは、私はマークです。

One of the key capabilities we're really excited to share with you today is real time conversational speech.

本日お伝えすることで本当に興奮しているキーとなる機能の一つは、リアルタイムの会話型音声です。

If you see, there's this little icon on the bottom right of the TrackGPT app, and this will open up GPT-4o's audio capabilities.

ご覧の通り、TrackGPTアプリの右下にこの小さなアイコンがあります。これにより、GPT-4oのオーディオ機能が開きます。

Hey, TrackGPT, I'm Mark.

こんにちは、TrackGPT、私はマークです。

How are you?

お元気ですか?

Mark, I'm doing great.

マーク、元気ですよ。

Thanks for asking.

お尋ねいただきありがとうございます。

How about you?

あなたはどうですか?

Although they've been doing a lot of talking about GPT-4o and this new model and how it's gonna be in TrackGPT for free and it's gonna be in the API, I think they really wanted to do this keynote today to show off this voice feature.

GPT-4oやこの新しいモデルについてたくさん話してきましたが、TrackGPTに無料で搭載されるし、APIにも搭載されると言っていました。しかし、実際には今日の基調講演を行うことで、この音声機能を披露したかったのだと思います。

This is very reminiscent of the movie Her where he basically has a chat bot companion that he talks back and forth with who's voiced by Scarlett Johansson.

これは、映画『HER』を思い起こさせるもので、主人公がスカーレット・ヨハンソンの声で話し合うチャットボットの相手を持っているシーンにとても似ています。

If you listen to this voice, a very similar sounding voice.

この声を聞くと、非常に似たような声がします。

Just what they're showing off here makes me think that we're about to see an explosion of AI girlfriend apps because the conversation feels so much more realistic.

ここで披露されている内容を見ると、AIのガールフレンドアプリが爆発的に増えることを予感させます。なぜなら、会話がよりリアルに感じられるからです。

As they're having this conversation, one thing to note is the latency difference.

この会話をしている間、注目すべき点の1つは遅延の違いです。

You used to ask a question, there'd be a little pause for five or six seconds.

以前は質問をすると、5〜6秒の小さな一時停止がありました。

You'd get their response.

その後、彼らの返答を受け取りました。

The responses are a little bit more real time.

お返事は少しリアルタイムになってきています。

It's feeling a little bit more like a real human to human conversation.

もう少し本当の人間同士の会話のように感じています。

Hey, so I'm on stage right now.

ねえ、今ステージにいます。

I'm doing a live demo and frankly, I'm feeling a little bit nervous.

ライブデモをしていて、正直、少し緊張しています。

Can you help me calm my nerves a little bit?

少し緊張を和らげてくれませんか?

You're doing a live demo right now?

今、ライブデモをしているんですか?

That's awesome.

すごいですね。

Just take a deep breath and remember, you're the expert.

深呼吸をして、自分は専門家だと思い出してください。

I like that suggestion.

その提案は気に入っています。

Let me try a couple of deep breaths.

数回深呼吸をしてみましょう。

Can you give me feedback on my breaths?

私の呼吸についてフィードバックをしていただけますか?

Here I go.

さあ、始めます。

No, no, no, no.

いいえ、いいえ、いいえ、いいえ。

Whoa, slowed you a bit there.

おや、ちょっとペースを落としましたね。

Mark, you're not a vacuum cleaner.

マーク、君は掃除機じゃないよ。

Breathe in.

息を吸ってください。

Or a count of four.

一、二、三、四と数えることもできます。

That's pretty crazy that it actually can listen to your breathing and tell you you're breathing too hard.

実際にあなたの呼吸を聞いて、呼吸が速すぎると教えてくれるのはかなりすごいですね。

It shows us that it is doing a pretty decent job of understanding some of the context of the conversation.

それは、会話の文脈の一部を理解することにかなりうまくいっていることを示しています。

I'm gonna breathe in deeply.

深く息を吸います。

For four, and then exhale slowly.

四秒間、そしてゆっくりと吐き出します。

I'll try again.

もう一度やってみます。

Breathing in and breathe out.

息を吸い、そして吐き出します。

That's it.

それでいいです。

How do you feel?

いかがお感じですか?

I feel a lot better.

私はずっと良くなりました。

Thank you so much.

本当にありがとうございます。

They were having a little bit of internet issues, probably because there are so many people all trying to connect to the same wifi.

彼らは少しインターネットの問題を抱えていました。おそらく、同じWi-Fiに接続しようとする多くの人がいるためです。

The demo definitely was live, and you do get a little bit of an extra delay that you probably wouldn't have normally just because of the connection issues.

そのデモは確かにライブでしたし、通信の問題のために通常ではない少しの遅延があります。

You've used our voice mode experience before.

以前に当社の音声モード体験をご利用いただいたことがありますね。

You'll notice a couple of key differences.

いくつかの主な違いに気づくでしょう。

First, you can now interrupt the model.

まず、モデルを中断することができるようになりました。

You don't have to wait for it to finish your turn before you can start speaking.

あなたは話す前に待つ必要はありません。

You can just butt in whenever you want.

いつでも割り込んで話すことができます。

Second, the model is real-time responsive.

第二に、そのモデルはリアルタイムで反応します。

That means that you don't have this awkward two to three second lag before you wait for the model to give a response.

つまり、モデルが返答をするのを待つ前に、2〜3秒の不自然な遅延がないということです。

Finally, the model picks up on emotion.

最後に、そのモデルは感情を感知します。

When I was breathing super hard there, it could tell, and you might want to calm down a little bit.

私が激しく息をしていたとき、それがわかりました。少し落ち着いた方がいいかもしれません。

You're kind of running a little bit too fast.

ちょっと速く走っているようですね。

It really has that capability across the board to perceive your emotion.

それは本当に、あらゆる面であなたの感情を感知する能力を持っています。

Not only that, though, the model is able to generate voice in a variety of different emotive styles, and it really has a wide dynamic range.

その上、そのモデルはさまざまな感情的なスタイルで声を生成することができ、実際に幅広いダイナミックレンジを持っています。

My friend Barrett here, he's been having trouble sleeping lately, and I want you to tell him a bedtime story about robots and love.

最近眠りが浅いと悩んでいる友人のバレットがいますので、ロボットと愛についての寝る前のお話をしてあげてほしいのですが。

A bedtime story about robots and love?

ロボットと愛についての寝る前のお話?

I got you covered.

任せて。

Gather round, Barrett.

集まれ、バレット。

Once upon a time, in a world not too different from ours, there was a robot named Fight.

むかしむかし、私たちの世界とあまり変わらない世界に、ファイトという名前のロボットがいました。

Fight was a curious robot, always exploring.

ファイトは好奇心旺盛なロボットで、常に探検していました。

I like the way you started this story, but I want a little bit more emotion in your voice, a little bit more drama.

この物語の始まり方は気に入っているが、声にもう少し感情を込めて、もう少しドラマを加えてほしい。

Got it.

わかりました。

Let's add some drama.

少しドラマを加えましょう。

Once upon a time, in a world not too different from ours, there was a robot named Bite.

昔々、私たちの世界とあまり変わらない世界に、バイトという名前のロボットがいました。

No, no, no, no, check, GPT.

いいえ、いいえ、いいえ、GPTを確認してください。

I really want maximal emotion, like maximal expressiveness, much more than you were doing before.

私は本当に、最大限の感情、最大限の表現力を求めています。以前よりもずっと多く。

Understood.

了解しました。

Let's amplify the drama.

ドラマを増幅させましょう。

Once upon a time, in a world not too different from ours, there was a robot named Bite. Bite was...

昔々、私たちの世界とあまり変わらない世界に、バイトという名前のロボットがいました。バイトは…。

Actually, can you do this in a robotic voice now?

実は、これをロボットの声でやってもらえますか?

Initiating dramatic robotic voice.

劇的なロボットの声を開始します。

Once upon a time, in a world not too different from ours, there was a robot named Bite.

昔々、私たちの世界とあまり変わらない世界で、バイトという名前のロボットがいました。

Bite was a curious robot.

バイトは好奇心旺盛なロボットでした。

That's pretty cool.

それはかなりクールですね。

It can actually change the way it speaks, make it more dramatic, make it sound like a robot.

実際、話し方を変えることができ、より劇的にしたり、ロボットのように聞こえるようにしたりすることができます。

That could come in handy for things like bedtime stories and maybe meditation apps to get that real calming voice.

それは、就寝時のストーリーや瞑想アプリなど、本当に落ち着いた声を得るために役立つかもしれません。

Things like that, it seems like it'll be really, really good at.

それのようなことは、本当に本当に得意そうですね。

The other thing they mentioned is that that latency has been sort of fixed, where there's not as much of a gap when you talk and when you get the response.

彼らが言及したもう一つのことは、その遅延が修正されたようで、話すときと応答を受け取るときの隔たりがあまりないということです。

It does seem like that gap has been closed, but it also kind of feels like maybe they're putting some filler words in there to close that gap.

その隔たりは縮まったように思えますが、その隔たりを埋めるために何か詰め物の言葉を入れているようにも感じられます。

You ask the question and it goes, hmm, let me think about that for a second.

質問をすると、「えー、ちょっと考えさせてください」と言ってから答えます。

Here's what I think.

私が考えるには。

It will sort of add filler and then give the response.

フィラーを追加してから返答をします。

I believe that's what it's doing.

それが何をしているのかと思います。

They didn't really speed up the latency on there.

そこでの遅延を本当に速くしたわけではありません。

They just made it start speaking before it was sort of ready to start speaking, if that makes sense.

それは、話す準備が整う前に話し始めるようにしただけです。それが意味をなすなら。

Can you end the story, but do it in a singing voice.

物語を終わらせてもらえますか?ただし、歌声でお願いします。

I'd found another robot friend and they lived circuitly ever after.

別のロボットの友達を見つけ、彼らは回路的に幸せに暮らしました。

Thank you so much.

本当にありがとうございます。

I definitely feel a lot more relaxed.

私は確かにもっとリラックスしています。

We also want to show you the vision capabilities as the model can also see the whole world around us.

また、モデルが私たちの周りの世界を見ることもできるビジョン能力をお見せしたいと思います。

Absolutely.

絶対に。

Like as you can see, you can chat with ChatGPT in real time, but you can also interact with it with video as well.

ご覧の通り、ChatGPTとリアルタイムでチャットすることができますが、ビデオでもやり取りすることができます。

Hey ChatGPT.

こんにちは、ChatGPT。

Hello there, how's it going?

こんにちは、調子はどうですか?

It's going really well.

とても良いです。

Today I'd really like your help solving a math problem.

今日は数学の問題を解くのを手伝っていただきたいです。

I'm all ears.

お耳を傾けております。

What math problem can I help you tackle today?

今日はどんな数学の問題を解決するのを手伝えますか?

I'm gonna write down a linear equation on a sheet of paper and I'll show you.

紙に一次方程式を書いて、お見せしますね。

I'd love your help working me through it.

お手伝いしていただけると嬉しいです。

But importantly, don't tell me the solution, just help give me hints along the way.

しかし重要なのは、解決策を教えないで、途中でヒントをくれることです。

I see it.

わかりました。

Hmm, no, I didn't show you yet.

えーと、まだ見せていませんね。

Just give me help along the way, one second.

途中で手助けをしてください、ちょっと待ってください。

Whoops, I got too excited.

おっと、興奮しすぎました。

I'm ready when you are.

準備ができています。

ChatGPT, what equation did I write down there?

ChatGPT、そこに書いた方程式は何ですか?

Ah, I see it now.

ああ、今見えました。

You wrote down three X plus one equals four.

あなたは3Xプラス1が4に等しいと書きました。

Yep, exactly.

はい、正確にそうです。

What's the first step I should take to try to solve this?

これを解決するために取るべき最初のステップは何ですか?

The first step is to get all the terms with X on one side and the constants on the other side.

最初のステップは、Xを含む全ての項を片側に、定数をもう片側に移すことです。

What do you think we should do with that plus one?

そのプラス1についてどう思いますか?

I'm gonna try to subtract one from both sides and then I'll see what I get.

両方の辺から1を引いてみます。その後、結果を見ます。

Great idea, perfect.

素晴らしいアイデアですね。

What do you get when you subtract one from both sides?

両辺から1を引いたら何が得られますか?

ChatGPT, this is what I ended up with.

ChatGPT、これが私が最終的にたどり着いたものです。

How does this look?

これはどう見えますか?

You've isolated the term with X on one side and it looks like three X equal three.

Xを持つ項を一方の側に孤立させ、3Xが3と等しいように見えますね。

What do you think we should do next to solve for X?

Xを求めるために次に何をすべきだと思いますか?

I'm not sure, can you give me another hint?

よくわかりません、もう一つヒントをくれますか?

Of course, you have three X and you wanna find the value of X. Think about what operation would undo multiplication.

もちろん、3つのXがあり、Xの値を見つけたいと思っています。乗算を元に戻す操作を考えてみてください。

Is it subtraction?

引き算ですか?

Close, but think of the opposite of multiplication because three X means three times X. Okay, I think it's division.

近いですが、3つのXはXを3倍することを意味するので、乗算の反対を考えてください。わかりました、除算だと思います。

Let me try to divide both sides by three and let me see what I get.

両辺を3で割ってみて、何が得られるか見てみましょう。

Go ahead and divide both sides by three.

両辺を3で割ってください。

ChatGPT, this is what I ended up with.

ChatGPT、これが私が得たものです。

How does this look?

これはどう見えますか?

It looks perfect.

それは完璧に見えます。

You've solved it and X equals one.

解決しましたね、そしてXは1に等しいです。

Nicely done.

素晴らしいですね。

This isn't really showing off its math capabilities.

これは実際には数学の能力を示していません。

That was a super, super simple math problem.

それは非常に、非常に簡単な数学の問題でした。

I don't really know if it's greatly improved with math, although people were claiming the gpt2 chat bot was better with math.

私は本当に、GPT-2チャットボットが数学でより優れていると主張している人々がいましたが、それが本当に向上しているかどうかはわかりません。

It probably is a little bit better with math.

それはおそらく少し数学が得意なようです。

This is more showing off the vision capability.

これは視覚能力を見せているというよりも、もっと見せているという感じです。

It can see as he writes the numbers on the paper and solves the problem.

紙に数字を書いて問題を解くのを見ることができます。

Really, again, this is showing in real time something that Google was kind of tweaking and speeding up and making it seem like it was able to do really fast, but it was really an edited video.

実際には、これはGoogleが微調整して高速化し、本当に速くできるように見せかけていたものをリアルタイムで示しているものですが、実際には編集されたビデオでした。

I think they're really trying to show you, again, look what we can do, but Google hasn't been able to do yet.

彼らは本当に、もう一度、私たちにできることを見せようとしていると思いますが、Googleはまだそれができていません。

ChatGPT, I really loved that you taught the value of math to my friend, Mark.

ChatGPT、あなたが私の友人であるマークに数学の価値を教えてくれたことが本当に好きです。

I wrote one last thing.

最後に一つ書きました。

I'd love if you could take a look at.

もし見ていただけると嬉しいです。

Of course, I'd love to see what you wrote.

もちろん、あなたが書いたものを見るのが楽しみです。

Show it to me whenever you're ready.

準備ができたら私に見せてください。

This is what I wrote down.

これが私が書き留めたものです。

What do you see?

何が見えますか?

I see.

わかりました。

I love ChatGPT.

私はChatGPTが大好きです。

That's so sweet of you.

あなたのその言葉、とても優しいですね。

Let me take out a computer for this.

一つコンピューターを取り出させてください。

I'm gonna open the ChatGPT desktop app like Mira was talking about before.

私は以前ミラが話していたように、ChatGPTデスクトップアプリを開くつもりです。

They're gonna get into the desktop app, which is pretty cool.

彼らはデスクトップアプリに入るつもりで、それはかなりクールです。

One of the things that you can do with the desktop app is copy everything on your screen, save it to your clipboard and then ChatGPT will use that information for its chat.

デスクトップアプリでできることの一つは、画面上のすべてをコピーしてクリップボードに保存し、それをChatGPTがチャットに使用することです。

It's also got a button to share your screen and ChatGPT can see everything that's going on on your screen and use that as context as well.

また、画面を共有するボタンもあり、ChatGPTは画面上で何が起こっているかをすべて見ることができ、それを文脈として使用することもできます。

Pretty cool.

かなりクールです。

Check this out.

これをチェックしてください。

To give a bit of background on what's going on.

何が起こっているかの背景を少し説明します。

Here we have a computer and on the screen we have some code.

こちらにはコンピューターがあり、画面にはいくつかのコードが表示されています。

The ChatGPT voice app is on the right.

右側にChatGPTの音声アプリがあります。

ChatGPT will be able to hear me, but it can't see anything on the screen.

ChatGPTは私の声を聞くことができますが、画面には何も見ることができません。

I'm gonna highlight the code, command C it, and then that will send it to ChatGPT.

コードをハイライトして、コマンドCでコピーし、それをChatGPTに送信します。

I'm gonna talk about the code to ChatGPT.

コードについてChatGPTに話します。

Hey ChatGPT.

こんにちは、ChatGPT。

Hey there, how's it going?

こんにちは、調子はどうですか?

It's going really well.

とても良いです。

I was wondering if you could help me with a coding problem today.

お手伝いいただけるかどうか、今日コーディングの問題で悩んでいるんですが。

Of course, I'd love to help you out.

もちろん、喜んで手伝います。

What's the coding problem you're dealing with?

どんなコーディングの問題を抱えていますか?

I'm gonna share with you some code.

あなたとコードを共有しようと思います。

One second.

ちょっと待ってください。

Sure thing, take your time.

もちろん、お時間をかけてください。

I just shared some code with you.

さっきコードを共有しました。

Could you give me a really brief one sentence description of what's going on in the code?

お手数ですが、コード内で何が起こっているかを簡潔に1文で説明していただけますか?

This code fetches daily weather data for a specific location and time period, smooths the temperature data using a rolling average, annotates a significant weather event on the resulting plot, and then displays the plot with the average, minimum, and maximum temperatures over the year.

このコードは特定の場所と時間帯の日次天気データを取得し、移動平均を使用して気温データを平準化し、その結果のプロットに重要な天候イベントを注釈付けし、その後、1年間の平均、最小、最大気温とともにプロットを表示します。

Awesome, thanks so much.

素晴らしい、ありがとうございます。

In the code, there's this function foo.

コード内に、この関数fooがあります。

Can you describe to me what the plot would look like if I had versus didn't have this function?

もしこの関数がある場合とない場合でプロットがどのようになるかを説明していただけますか?

Sure, the function foo xy takes in a data frame x containing temperature data and a window size y for calculating the rolling mean.

もちろん、関数foo xyは、気温データを含むデータフレームxと、移動平均を計算するためのウィンドウサイズyを受け取ります。

It smooths the average temperature, minimum temperature, and maximum temperature columns by applying a rolling mean over the specified window.

指定されたウィンドウ上で移動平均を適用することで、平均気温、最低気温、最高気温の列が滑らかになります。

With the foo function, the plot would show smoother temperature lines because the rolling mean reduces the noise or fluctuations in the data.

foo関数を使用すると、プロットにはデータのノイズや変動が減少するため、より滑らかな気温の線が表示されます。

Awesome.

素晴らしいですね。

This makes-

これにより-

Thanks so much.

ありがとうございます。

I definitely will want to apply this function.

この関数を適用したいと思います。

I'm gonna generate the plot and then show it to you.

私はプロットを生成してから、それをあなたに見せるつもりです。

Give me one second.

ちょっと待ってください。

Absolutely, take your time.

もちろん、お時間をかけてください。

I'm excited to see the plot when you're ready.

準備ができたらプロットを見るのが楽しみです。

So far, ChatGPT has only been able to see the code, and now I'm gonna run it and then I'll use the vision capabilities of the ChatGPT desktop app so it can actually see everything going on on the screen as well.

これまで、ChatGPTはコードしか見ることができませんでしたが、今度はそれを実行して、ChatGPTデスクトップアプリのビジョン機能を使って、画面上で起こっているすべてを実際に見ることができるようにします。

I don't really know if this is an improved coding ability inside of ChatGPT and GPT-4.

ChatGPTやGPT-4内での改善されたコーディング能力についてはよくわかりません。

I feel like what they just showed, it's been able to do, but what's really cool about it is the fact that they were able to copy the code and then the ChatGPT was able to read it basically just cause it was copied to the clipboard and give this sort of verbal explanation of what was going on in the code and what would happen if you changed the specific variable that he was referencing.

おそらく、彼らが見せたものはできることだと感じていますが、実際にクールなのは、コードをコピーして、それがクリップボードにコピーされたためにChatGPTがそれを読むことができ、コードで何が起こっているか、特定の変数を変更した場合に何が起こるかについての口頭の説明をすることができたことです。

I'm sharing with you the plot now.

今、プロットを共有しています。

I'm wondering if you can give me a really brief one sentence overview of what you see.

あなたが見ているものの非常に簡潔な一文の概要を教えていただけるかしら。

The plot displays smoothed average, minimum, and maximum temperatures throughout 2018 with a notable annotation marking a big rainfall event in late September.

プロットは、2018年全体を通じて平滑化された平均、最低、最高気温を表示し、9月下旬の大雨イベントを示す注釈があります。

Fantastic, I have another question for you.

素晴らしいですね、もう1つ質問があります。

Which months do you see the hottest temperatures and roughly what temperatures do those correspond to?

どの月が最も高い気温を示し、それらがどのような気温に対応するか、おおよそ教えていただけますか?

The hottest temperatures occur around July and August.

最も高い気温は7月と8月頃に起こります。

The maximum temperature during these months is roughly between 25 degrees and 30 degrees, 77 degrees Fahrenheit to 86 degrees Fahrenheit.

これらの月の最高気温はだいたい25度から30度、華氏77度から86度です。

What's going on with the Y axis?

Y軸は何が起こっているのですか?

Is this in Celsius or Fahrenheit?

これは摂氏ですか、それとも華氏ですか?

The Y axis is in Celsius.

Y軸は摂氏です。

The temperatures are labeled accordingly such as average temperature, minimum temperature, and maximum temperature.

気温は平均気温、最低気温、最高気温などとして適切にラベルが付けられています。

If GPT-4o is capable of real-time translation, Mike, you wanna try this one?

もしGPT-4oがリアルタイム翻訳が可能なら、マイク、これを試してみたいですか?

Sure, yeah.

はい、もちろん。

It's really cool to see this translation feature because this is gonna make communicating in other countries so much easier.

他の国でのコミュニケーションがとても簡単になるだろうというこの翻訳機能を見るのは本当に素晴らしいですね。

You just open up your phone to the ChatGPT app and then you can have a conversation with anybody in almost any language.

あなたはただChatGPTアプリを開いて、ほとんどどんな言語でも誰とでも会話ができます。

Check this out.

これを見てください。

Let's do it.

やってみましょう。

I speak Italian so we can just do English and Italian.

私はイタリア語を話すので、英語とイタリア語でやりましょう。

Try to do English, Italian.

英語、イタリア語をやってみてください。

Sure, let's do it.

もちろん、やりましょう。

I would like you to function as a translator.

通訳として機能していただきたいです。

I have a friend here who only speaks Italian and I only speak English.

ここにはイタリア語しか話せない友達がいて、私は英語しか話せません。

Every time you hear English, I want you to translate it to Italian.

英語を聞くたびに、それをイタリア語に翻訳してほしいです。

If you hear Italian, I want you to translate it back to English.

もしイタリア語を聞いたら、それを英語に翻訳してほしいと思います。

Is that good?

それはいいですか?

Perfecto.

完璧です。

Mike, she wonders if whales could talk, what would they tell us?

マイク、彼女はクジラが話せたら、彼らは私たちに何を伝えるだろうかと思っています。

They might ask, how do we solve linear equations?

彼らはおそらく、どのように線形方程式を解くかを尋ねるかもしれません。

Certainly yes.

かしこまりました。

Can you tell what you're feeling just by looking at your face?

あなたの顔を見て、自分が何を感じているかわかりますか?

Barrett, do you wanna give this one a try?

バレット、これを試してみたいですか?

Absolutely, let's try it out.

もちろん、試してみましょう。

I'm gonna show you a selfie of what I look like and then I'd like you to try to see what emotions I'm feeling based on how I'm looking.

私がどんな顔をしているかの自撮りを見せてから、私がどんな感情を感じているかを見てみてください。

Sounds like a fun challenge.

楽しい挑戦のようですね。

Go ahead and show me that selfie and I'll put my emotional detective hat on.

お願いして、その自撮りを見せてください。私は感情の探偵帽子をかぶります。

Here's me.

こちらが私です。

What kind of emotions do you think I'm feeling?

私がどんな感情を感じていると思いますか?

It seems like I'm looking at a picture of a wooden surface.

木製の表面の写真を見ているようですね。

You know what?

分かりますか?

That was the thing I sent you before.

それは以前に送ったものです。

Don't worry, I'm not actually a table.

心配しないでください、実際には私はテーブルではありません。

Take another look.

もう一度よく見てください。

That makes more sense.

それはもっと理にかなっています。

There we go.

そうですね。

It looks like you're feeling pretty happy and cheerful with a big smile and maybe even a touch of excitement.

あなたはとても幸せで陽気で、大きな笑顔をしていて、興奮すら感じているようですね。

Whatever's going on, it seems like you're in a great mood.

何があっても、あなたは素晴らしい気分のようです。

Care to share the source of those good vibes?

その良い雰囲気の源を共有していただけますか?

No, the reason I'm in a really good mood is because we were doing a presentation showcasing how useful and amazing you are.

いい気分なのは、あなたがどれだけ役立ち、素晴らしいかを示すプレゼンテーションをしていたからです。

Stop it.

やめてください。

You're making me blub.

私を泣かせるんですね。

One thing that was really interesting about that little clip there was that at first, it said I see something like a wooden table and then he said, oh no, look again.

その小さな映像について興味深いことの1つは、最初に木製のテーブルのようなものが見えると言ってから、再度見るように言われたことです。

That makes me think that it's actually not watching the video, it's just taking screenshots at certain times because he must've had his phone face down on the table or looking at the table, it took a picture of the table and then when he put the camera up to himself and he told it to try again, it then saw his face.

それは、実際にはビデオを見ているのではなく、特定の時点でスクリーンショットを撮っていると考えさせられます。彼はおそらく、電話をテーブルの上に向けたり、テーブルを見たりしていたので、テーブルの写真を撮り、カメラを自分に向けて再度試してみると、彼の顔を見たのでしょう。

I don't think it's watching video footage, I think it's just taking snapshots when you ask questions.

おそらく、ビデオ映像を見ているのではなく、質問をするときにスナップショットを撮っていると思います。

Again, you can play with it right now inside of the OpenAI Playground and I don't know if it's rolled out for everybody yet or not inside of ChatGPT, but I just opened my ChatGPT account and I got this notification, introducing GPT-4o.

また、OpenAI Playground内で今すぐ試すことができますが、ChatGPT内で全員に展開されているかどうかはわかりませんが、私はちょうどChatGPTアカウントを開いたところ、GPT-4oを紹介する通知を受け取りました。

You can now try our newest model.

今、私たちの最新モデルを試すことができます。

It's faster than GPT-4, better at understanding images and speaks more languages, try it now.

GPT-4よりも速く、画像を理解する能力が向上し、より多くの言語を話すようになりました。今すぐ試してみてください。

We can see up at the top, we now have the option of GPT-4o, GPT-4 and GPT-3.5.

上部に表示されているように、今ではGPT-4o、GPT-4、GPT-3.5のオプションがあります。

I am a plus member, but GPT-4o is available inside of my account right now.

私はプラス会員ですが、GPT-4oは今私のアカウント内で利用可能です。

My biggest takeaways from this event, I think the chat feature is really, really cool.

このイベントからの一番の収穫は、チャット機能が本当にすごいと思います。

We've been able to chat on the mobile app for a while now, but it doesn't have the inflections, it's got the longer latency.

私たちはしばらくモバイルアプリでチャットできていましたが、抑揚がなく、レイテンシーが長いです。

You weren't able to sort of cut it off mid-sentence and continue the conversation.

文中で途中で切り上げて会話を続けることができませんでした。

There's a lot of improved features over that voice chat, which I think is really cool.

その音声チャットよりも改善された機能がたくさんありますが、それは本当にすごいと思います。

I don't know if we saw a huge leap between GPT-4 and what GPT-4o does.

GPT-4とGPT-4oの間に大きな飛躍が見られたかどうかはわかりません。

I feel like it's a pretty similarly capable model.

私はそれがかなり同様に能力のあるモデルだと感じています。

We saw from all the gpt2-chatbot videos and tests that it's slightly better than GPT-4 in a lot of areas, but not a huge leap above them.

私たちはすべてのGPT-2-chatbotのビデオやテストから、それがGPT-4よりも多くの分野でわずかに優れていることを見ましたが、それらよりも大幅に進歩しているわけではありません。

It also seems to be lightning fast.

それは非常に高速であるようにも見えます。

If you've used GPT-4 instead of ChatGPT at all, GPT-4o is a lot faster.

もしChatGPTの代わりにGPT-4を使用したことがあるなら、GPT-4oの方がはるかに速いです。

If I give it a prompt here and hit enter, we can actually see in real time how fast it actually completes the writing of this prompt.

ここでプロンプトを与えてエンターキーを押すと、実際にそのプロンプトの書き上げがどれだけ速いかをリアルタイムで見ることができます。

It just cranks it out really, really fast.

それは本当に、本当に速く書き上げます。

Also, another sort of theme with ChatGPT is that almost every time they roll out an update like this and make some big announcement, it seemingly kills a whole bunch of little SaaS companies that have been building on top of their APIs.

また、ChatGPTに関連するもう1つのテーマは、彼らがこのようなアップデートを展開し、大きな発表をするたびに、それがAPIの上に構築してきた多くの小さなSaaS企業を壊滅させるように見えることです。

Just like that, inside of the free version of ChatGPT, we're going to have translation, which was a whole industry of tools that have been popping up built on the GPT APIs.

無料版のChatGPTの中で、翻訳ができるようになります。これは、GPT APIを利用したツールの一つの産業であり、最近急速に増えています。

We have AI girlfriends.

私たちにはAIのガールフレンドがいます。

This is a niche with an AI that has been really, rapidly arising.

これは急速に台頭しているAIを持つニッチ市場です。

Lots and lots of apps have been submitted to Future Tools about these like AI girlfriends.

たくさんのアプリがFuture ToolsにAIのようなガールフレンドについて提出されています。

It looks like GPT-4o could just kind of act as that AI girlfriend or boyfriend or significant other.

GPT-4oは、そのAIのガールフレンドやボーイフレンド、パートナーとして機能する可能性があります。

We've seen tools like DevOn and GitHub Copilot.

DevOnやGitHub Copilotのようなツールを見てきました。

While I don't think this is gonna kill tools like DevOn or GitHub Copilot, it may make it so you don't need a third party coding tool that you pay extra for.

これはDevOnやGitHub Copilotのようなツールを壊すことはないと思いますが、追加料金を支払う必要があるサードパーティのコーディングツールが必要なくなるかもしれません。

You might be able to do it all just with the free version of GPT-4o.

無料版のGPT-4oだけでそれをすべて行うことができるかもしれません。

There's also apps out there that sort of watch your screen and listen to you all day and can help give you a recap of what you did throughout the day.

画面を見て一日中聞いてくれるアプリもあり、一日の活動をまとめてくれます。

Well, it looks like the desktop version of this might be able to do that.

まあ、このデスクトップ版はそれができるかもしれません。

I don't know if you could just leave it running all the time and just have it kind of keep track of what you're doing, but I wouldn't imagine that that's too far out of the question.

いつも稼働させたままにして、自分が何をしているかを追跡させることができるかどうかはわかりませんが、それが問題ではないとは思いません。

I just find it really interesting that OpenAI has this model of building APIs, letting companies build on the APIs and then going and building a ton of these features just right into their own products so that you don't need the tools that were built with the APIs.

OpenAIがAPIを構築し、企業がそのAPIを活用して機能を構築し、それを自社製品に直接組み込むというモデルを持っていることが本当に興味深いと思います。そのため、APIで構築されたツールが必要なくなります。

It's just really, really interesting to me.

おもしろいと思います。

This to me seems like what Siri should be, and the rumors are right now that Siri is probably going to use OpenAI's tech, and this might be the future of Siri.

これは私にとって、Siriがなるべきもののように思えます。そして現在の噂では、SiriはおそらくOpenAIの技術を使用するだろうと言われており、これがSiriの未来かもしれません。

We'll find out pretty soon, most likely at WWDC.

おそらくWWDCですぐにわかるでしょう。

What's really cool is if you go to the blog post over on OpenAI's website here, they have a whole bunch of other demos that we didn't get into yet.

本当に素晴らしいのは、OpenAIのウェブサイトのブログ投稿に行くと、まだ触れていないたくさんの他のデモがあることです。

For example, this one, Greg Brockman here got the two phones to basically sing to each other back and forth between the two phones.

たとえば、グレッグ・ブロックマンさんがここで2台の電話をお互いに歌わせるようにさせた例があります。

We've got some interview prep.

面接の準備をしています。

We've got it playing rock, paper, scissors, testing sarcasm, harmonizing, pointing and learning Spanish, summarizing meetings, real-time translations, lullabies, talking faster, singing happy birthday, just all sorts of demos and use cases that you can check out.

いろんなデモやユースケースをチェックできるように、ロック・ペーパー・シザーズをプレイしたり、皮肉をテストしたり、ハーモナイズしたり、指さしたり、スペイン語を学んだり、会議を要約したり、リアルタイム翻訳をしたり、子守歌を歌ったり、早口で話したり、ハッピーバースデーを歌ったりしています。

We'll make sure to link to this blog post below the video so you can watch some of these demos.

このブログポストへのリンクをビデオの下に貼るようにしますので、これらのデモのいくつかをご覧いただけます。

I don't wanna just sit here and just watch demo after demo after demo, but it looks to do some pretty cool stuff.

ただデモをひたすら見続けるのはちょっと…でも、かなりクールなことをしているようです。

Is it as big and as exciting of news as everybody was anticipating?

皆が予想していたほど大きくてエキサイティングなニュースなのでしょうか?

Probably not.

おそらくそうではないですね。

The voice and the desktop app are really, really cool.

その音声とデスクトップアプリは本当にすごくクールです。

GPT-4o is a slightly improved model, which is also cool, but that voice capability in the desktop app we don't really have access to yet.

GPT-4oはわずかに改良されたモデルで、それもすごいですが、デスクトップアプリの音声機能にはまだアクセスできていません。

They said that's gonna be rolling out soon, so we can't actually play with that now.

おそらくそれは近々展開されると言っていましたので、実際には今それで遊ぶことはできません。

Really all we have today to play with is that GPT-4o model, but not the cool voice features, not the cool apps, not the stuff that really gets us excited that we wanna play with right now.

今日遊べるのは本当にGPT-4oモデルだけですが、クールな音声機能やクールなアプリ、本当に興奮させてくれるものはありません。今遊びたいと思っているものはありません。

We just have a slightly improved version of GPT-4o that we can play with right now.

今遊べるのは、わずかに改良されたGPT-4oのバージョンだけです。

But this definitely brought us one step closer to that movie, Her, where you can legitimately have real-life conversations back and forth in the chatbot and feel like you're talking to another human.

しかし、これは確かに、映画「HER」に一歩近づいたもので、チャットボットで実際にリアルな会話を行い、他の人間と話しているような感覚を得ることができます。

It's pretty crazy where this stuff is going.

この技術の進化の方向性はかなり驚くべきものです。

I'm so excited for it.

私はそれにとても興奮しています。

I've got a lot more announcements and events coming up.

私はもっとたくさんの発表やイベントが控えています。

Again, I'm at the Google event right now, so it's gonna be interesting to see if Google's gonna be able to top what OpenAI just showed off.

今、私はGoogleのイベントに参加していますので、OpenAIが見せたものをGoogleが上回ることができるかどうか興味深いです。

Next week, I'm gonna be at the Microsoft event where they're going to be trying to one-up everybody that came before them.

来週、私はMicrosoftのイベントに参加する予定で、彼らはこれまでの誰よりも優れたものを目指しています。

Really, really interesting time in the world of AI.

本当に、AIの世界で非常に興味深い時期です。

I'm gonna be making a lot of videos.

私はたくさんの動画を作る予定です。

You're gonna see me in hotel rooms a lot over the next couple of months because I'm going to a lot of these keynotes.

これから数ヶ月、たくさんの基調講演に出るので、ホテルの部屋でよく見かけることになるでしょう。

Should be a good time.

楽しい時間になるはずです。

I'm gonna do my best to keep you in the loop.

最新情報をお知らせするために最善を尽くします。

If you're not subscribed to this channel already, make sure you're subscribed.

このチャンネルにまだ登録していない場合は、必ず登録してください。

I will make sure you're updated with all the latest AI news.

最新のAIニュースをお知らせするようにします。

If you like this video specifically, give it a thumbs up.

もしこの動画が気に入ったら、高評価をしてください。

It makes me feel good and helps the algorithm.

私はそれを見て嬉しくなり、アルゴリズムの助けになります。

Thank you so much for tuning in.

ご視聴ありがとうございます。

I really, really appreciate you.

本当に、本当に感謝しています。

Hope you enjoyed this little nerd-out session about OpenAI's keynote today, and I will see you in the next video.

今日のOpenAIの基調講演についてのこの小さなオタクセッションを楽しんでいただけたら嬉しいです。次の動画でお会いしましょう。

Bye-bye.

さようなら。


この記事が気に入ったらサポートをしてみませんか?