Name: 競爭扼殺AI突破？LLion Jones TED演講揭曉！ (How Competition Is Stifling AI Breakthroughs | Llion Jones | TED)
Uploaded: 2026-03-02T21:37:53.000Z
Duration: 16 min 1 s

情態副詞

So as mentioned, I'm probably most well known as one of the Transformers authors.

所以就像我剛才說的，我大概最為人熟知的身份，就是其中一位 Transformers 的作者。

Transformers are of course the T in ChatGPT and are the architectures that run most of the state of the art artificial intelligence.

Transformers 當然就是 ChatGPT 中的 T，也是目前大部分最先進的人工智慧所採用的架構。

程度副詞

If I think back to that time when we were working on the Transformers, I remember it as a very organic, bottom up kind of project where the idea came from talking over lunch or scribbling randomly, whiteboards in the office.

如果我回想當時我們在做 Transformers 的時候，我記得那是一個非常有機、由下而上的專案，點子都是來自午餐時的哈拉，或是隨意在辦公室的白板上塗鴉。

And importantly, when we felt like we did actually have a good idea, we had the freedom to actually spend the time and go and work on it.

而且重要的是，當我們覺得我們真的有個好點子的時候，我們就有自由去花時間好好地把它做出來。

And even more importantly, we didn't have any pressure that was coming down from management.

更重要的是，我們沒有來自管理階層的任何壓力。

No pressure to work on any particular project, publish a number of papers, to push a certain number up.

沒有壓力要我們做任何特定的專案，發表一定數量的論文，或是達成某個特定的數字。

So that's the image I want you to have in your mind, right?

所以這就是我希望你們腦中浮現的畫面，對吧？

That is the kind of environment that allowed the Transformer to come into existence.

這就是讓 Transformer 得以誕生的那種環境。

An organic, open ended and with a lot of freedom to pursue the ideas that we thought were interesting and important.

有機、開放、而且有很大的自由去追求我們認為有趣且重要的想法。

And my deep concern is that right now in the AI industry we do not have this kind of environment.

而我非常擔心的是，現在的人工智慧產業，並沒有這樣的環境。

And I want to talk about why not and what can we do about it.

我想來聊聊為什麼會這樣，以及我們可以怎麼做。

So the main paradox that I see in artificial intelligence research or the industry in general right now is that despite the fact that there's never been so much interest and resources and money and talent, this has somehow caused a narrowing of the research that we're doing.

所以我現在看到的人工智慧研究或整個產業的主要悖論是，儘管前所未有的關注、資源、資金和人才湧入，卻反而導致我們正在做的研究範圍縮小了。

And to me, I think the reason is fairly obvious.

It's because the immense amount of pressure that comes with that, right?

就是因為隨之而來的巨大壓力，對吧？

Pressure from investors that are going to ask for a return on their investment, and

來自投資者的壓力，他們會要求投資報酬，還有

Because this is such an overcrowded industry right now, where it's very difficult to stand out.

因為現在這個產業實在太過擁擠了，要脫穎而出非常困難。

And the researchers, they're really feeling this pressure, right?

研究人員真的能感受到這種壓力，對吧？

If you're doing, let's say, standard AI research right now, you kind of have to assume that there's maybe three or four other groups doing something very similar, or maybe exactly the same.

如果你現在做的是一般的人工智慧研究，你大概得假設有三四個其他團隊也在做非常類似，甚至完全相同的事情。

So you have to spend the time checking to see if you've been scooped, to see if someone else has put your idea out there.

所以你必須花時間去確認自己有沒有被搶先，有沒有其他人已經發表了你的想法。

And even in academia, where you would hope you would have more freedom, there's pressure to publish and to have your papers published.

而且即使在學術界，你原本以為會有更多自由的地方，也承受著發表論文、讓論文被發表的壓力。

So if you have an interesting idea that could create something very interesting, or you have kind of a mediocre idea that no, it'll probably get a paper and probably get accepted, the temptation is to go for the low hanging fruit.

所以如果你有一個有趣的點子，可以創造出很棒的東西，或者你只有一個普通的點子，但覺得「嗯，這大概能寫成一篇論文，而且很可能會被接受」，那麼誘惑就會讓你選擇那些容易達成的目標。

So unfortunately, this pressure damages the science because people are Rushing out papers and it's reducing the amount of creativity that we have.

所以很可惜，這種壓力損害了科學，因為大家都在趕著發表論文，這也減少了我們的創意。

So I want to take an analogy from AI itself.

所以我想從人工智慧本身來做個類比。

So when we're designing AI search algorithms, we have to trade off something called the exploration exploitation.

當我們在設計人工智慧搜尋演算法時，我們必須在所謂的「探索」和「利用」之間做權衡。

When you're searching for a solution you can either spend your time exploring or exploiting.

當你在尋找解決方案時，你可以花時間去探索，也可以花時間去利用。

If you spend all your time exploring, then you will probably only find a large number of medical solutions.

如果你花所有時間去探索，那麼你很可能只會找到大量的普通解決方案。

If you spend your time just exploiting, then you might lose out on finding other alternative solutions that you might be able to exploit better and improve better.

如果你只花時間去利用，那麼你可能會錯過其他更好的替代解決方案，而這些方案你本來可以更好地利用和改進。

And we are almost certainly in that situation right now in the AI industry.

而我們現在在人工智慧產業裡，幾乎可以肯定就處於這種情況。

So all I really want to ask you today is to consider just changing that balance a little bit, just turning up the dial and exploring more.

所以今天我真正想請大家做的，就是稍微調整一下這個平衡，多轉動一點旋鈕，多去探索。

So I actually remember what it was like just before the Transformers, and

我還記得在 Transformers 出現之前是什麼樣子的，

I want to paint that picture for you as well.

我也想為大家描繪一下當時的景象。

Back then, my main memory was there were a lot of papers coming out and they were always permutating the current architecture, but

當時我主要的記憶是，有很多論文發表，它們總是在現有的架構上做排列組合，

which was recurrent neural networks at the time, just endlessly trying different things, different gates, different layers, mostly for incremental gains.

那時候是循環神經網路（recurrent neural networks），不斷嘗試不同的東西、不同的閘（gates）、不同的層，但大多隻是為了漸進式的進步。

And then after the transformer came out, all of that work that was spent on improving the recurrent neural network kind of felt a bit pointless.

然後 Transformer 出現之後，那些花在改進循環神經網路上的工夫，感覺都變得有點沒意義了。

How much time do you think those researchers would have spent trying to improve the recurrent neural network if they knew something like Transformers was around the corner?

你覺得那些研究人員如果知道 Transformer 就在不遠處，他們會花多少時間去改進循環神經網路呢？

It turned out we needed a longer conceptual leap.

結果我們需要一個更長遠的概念性飛躍。

We needed to throw away recurrence completely.

And again, I am worried that we're in that situation right now where we're just concentrating on one architecture and just commuting it and trying different things, where there might be a breakthrough just around the corner.

再次強調，我擔心我們現在就處於這種情況，我們只專注於一種架構，不斷地調整它，嘗試不同的方法，而突破可能就在不遠處。

And if there is, then we should be acting like it.

如果真是如此，那我們就該有所行動。

The next breakthrough, almost by definition, has to come from this sort of open ended, much more speculative research.

下一個突破，幾乎可以肯定，必須來自這種開放式的、更具推測性的研究。

And the only way to really hedge your bets against missing out on the next big thing is to invest in this kind of research.

而真正能避免錯過下一個重大突破的方法，就是投資這類研究。

So if I came up here and did nothing but just moan about the current situation, I don't think it would be a great talk.

所以，如果我今天上臺只是抱怨現況，我覺得這不會是一場很棒的演講。

So I want to give you a couple of suggestions.

First of all, in my company we champion having nature inspired.

首先，在我的公司，我們大力提倡從大自然中汲取靈感。

So there are still things, plenty of things that the human brain can do that current state of the art AI can't do.

所以，人類大腦還有很多事情是目前最先進的人工智慧做不到的。

So maybe if we take some inspiration from nature, we can get some of those properties, but

所以，也許我們從大自然汲取靈感，就能獲得其中一些特性，但

You should follow what's interesting to you.

There's actually a quote I heard two weeks ago, and I thought, that's perfect.

其實兩週前我聽到一句話，我當時覺得，這太完美了。

I'm having that for my talk, and I think I'm stealing it from a guy called Brian Chung.

我演講就要用這句話，我想我應該是從一個叫 Brian Chung 的人那裡「借」來的。

You should only do research that wouldn't happen if you weren't working on it.

你只應該做那些如果你不參與，就不會發生的研究。

我認為這句話完美地捕捉了精髓。

And if we all did that, we wouldn't be stepping on each other's toes and we'd be exploring much more efficiently.

如果我們都這麼做，就不會互相踩到腳，而且我們會更有效率地探索。

So I want to give you a concrete example.

所以我想給大家一個具體的例子。

There's a piece of research that we put out recently called the Continuous Thought Machine.

我們最近發表了一篇研究，叫做 Continuous Thought Machine。

And all we did is we just took a little bit of inspiration from nature.

我們只是從大自然中獲得了一點靈感。

So in the human brain, synchronization is very important, and

we try to add this kind of synchronization into artificial neural networks.

我們試圖將這種同步性加入人工神經網路。

I remember the employee coming to me with the idea, and

我記得那位員工帶著這個想法來找我，

I said, okay, work on it for a week and we'll see what happens.

我說，好，你先做一個禮拜看看結果如何。

That employee later confided in me that in his previous employment, or even in his academic position before that, that he probably would have gotten skepticism and told not to waste his time.

那位員工後來告訴我，在他之前的公司，甚至是他之前的學術職位，他可能都會受到懷疑，被告知不要浪費時間。

But after that week, he started to find much more interesting properties of this model, and

但在那一個禮拜後，他開始發現這個模型有更多有趣的特性，

We actually announced that we got the spotlight at NeurIPS this year.

我們今年其實有宣佈在 NeurIPS 獲得了大家的關注。

And I think there's a couple of reasons for that.

I think there's a hunger for this kind of new and differentiated research.

我覺得大家渴望這種嶄新且具差異化的研究。

And more interestingly, at no point when we were working on this project did we have to worry about being scooped so we could take our time to do the science properly and run the benchmarks that we wanted to run.

更有趣的是，我們在進行這個專案時，從來不用擔心被別人搶先發表，所以我們可以花時間好好做研究，跑我們想跑的基準測試。

And I think that's the kind of research we should be doing.

我認為這才是我們應該做的研究。

So hopefully from that, you can tell that I'm not just up here trying to make a talk that sounds good.

所以希望從這裡，你們可以看得出來，我不是在這裡講一些聽起來很棒的演講。

我是真心相信這件事的，對吧？

I am putting my money where my mouth is, and

I am creating this kind of environment, the kind of environment that allow transformers to come into existence at my company.

我正在創造這種環境，一種讓 Transformers 能在我的公司誕生的環境。

I'm not sure if I should tell you this, because it's a bit of an advantage that the company has right now, but

我不太確定我該不該告訴你們，因為這算是公司目前的一個優勢，但是

it's a really, really good way of getting talent.

這是招募人才的一種非常非常好的方式。

Talented, intelligent people, ambitious people will naturally seek out this kind of environment with high autonomy.

有才華、聰明、有企圖心的人，自然會尋找這種高自主性的環境。

And some of our best hires recently have been explicitly because of this reason.

而且我們最近一些最優秀的員工，就是因為這個原因被聘用的。

Oh, by the way, it works better than just money.

喔對了，這比單純給錢還有用。

These superstars that are apparently being snapped up for literally a million dollars a year in some cases.

那些據說年薪百萬美元的超級巨星。

Do you think that when they start their new position, they feel empowered to try their mad ideas, their more speculative ideas?

你覺得他們開始新工作時，會覺得有空間去嘗試他們瘋狂的點子、更具投機性的想法嗎？

Or do they feel immense pressure to prove their worth and will once again go for the low hanging fruit?

還是他們會感受到巨大的壓力，必須證明自己的價值，然後再次去摘取容易的成果？

I think that maybe we're not exploring quite as efficiently as we should be.

我想我們可能沒有像應該的那樣有效率地探索。

And that's because Transformers are too good.

這是因為 Transformers 太強大了。

I know, modesty, but seriously, I mean, what can I mean by that?

我知道，這聽起來有點自誇，但說真的，我的意思是，我到底想表達什麼？

What I mean is I think the punchline is going to be that when we look back at this point in AI history, the fact that the current technology is so powerful and flexible that it stopped us from looking for better, it makes sense, right?

我的意思是，我想重點是，當我們回顧人工智慧歷史的這個時刻，現有的技術如此強大且靈活，反而阻止我們尋找更好的技術，這聽起來很合理吧？

If the current technology was worse, more people would be looking for better.

如果現有的技術比較差，就會有更多人去尋找更好的。

So there's two points I would like to clarify.

First of all, I'm not saying that there isn't already plenty of very interesting research happening.

首先，我不是說現在已經沒有很多很有趣的研究了。

I'm just saying that given the amount of talent and resources that we have currently, we can afford to do a lot more.

我只是說，考量到我們目前擁有的人才和資源，我們可以做得更多。

I and several other many other researchers believe we're not done and we should be looking for better.

我和許多其他研究人員都相信，我們還沒到終點，應該要尋找更好的。

But I'm also not saying that we should throw away the current technology.

但我也不是說我們應該拋棄現有的技術。