Mark Zuckerberg, the creator of Facebook and CEO of Meta, was interviewed by YouTuber Kane Sutter (@Kallaway) last Thursday to discuss trends in the AI world, Meta's growth strategy, and how Meta is "materializing" AI.
The whole interview was a relaxed conversation, and even careful netizens found that Zuckerberg became more and more natural in his dressing style and conversation, and became more "like a person".
However, the conversation is easy, but Meta AI and Zuckerberg are not so easy in the big melee to promote AI products.
Meta AI's rollout plan was first blocked in Europe; Later, it was "unilaterally broken up" by Apple, directly blowing up hundreds of millions of iPhone users; A few days ago, he was ridiculed by the Australian Prime Minister for using Australian media news materials to train a large model, "Meta is an arrogant and rude company".
Even so, Xiaozha still carried through his insistence on open-source AI development, and carefully and sincerely shared his views on developers, AI development, and smart devices.
Closed-source AI is an "artificial god"
While major AI companies are scrambling to work on the biggest, fastest, and best large language models, Zuckerberg has "seen through the essence."
In this interview, Xiaozha firmly pointed out that there will not be only one artificial intelligence in this world, and there will be more and more artificial intelligence in the future, because more and more people will be able to create different things.
This is why he believes that AI should be open sourced: "AI technology should not be 'hoarded' and should not be used by a company to build their god-level products. ”
This passage was ridiculed by many netizens, as if they didn't say anything, and as if they said everything. Because looking at it, whether it is OpenAI, the king of AI today, or Anthropic, the most powerful language model released a few days ago, they are all core and single-product supported AI companies (OpenAI's ChatGPT and Anthropic's Claude).
At the same time, major AI companies are scrambling to advertise themselves as "real artificial intelligence", but Xiaozha is smiling and silent. He believes that every AI company believes that what they are doing is "right" and "noble" artificial intelligence, which is no different from "creating God", and it is the users who use artificial intelligence that really determine whether artificial intelligence is good or bad.
Therefore, maximizing the use of users to obtain data and continuously optimize it may be a topic that Meta AI has always been a topic of. He was very eager to dissociate Meta AI from this "god-making" behavior, "This is not what we are doing. ”
Meta, let everyone enjoy development
As for what "open source" means for the AI industry, it actually corresponds to the question about the "future of individual developers" mentioned in this interview.
In this regard, Xiaozha also predicts that in the next 10 years, the number of developers will continue to increase. Here Xiaozha also exploded in fatherly love, mentioning that his children, although they are still playing Lego, can already be said to have a talent for creators. Using his children as an example to encourage everyone to develop their own "creative ability" is unheard of in Zuckerberg's history, so it is also jokingly called "no longer a robot" by netizens
Back to developers, in fact, Zuckerberg's insistence on the open source policy also confirms Meta's greatest efforts to serve the largest developer team, such as Llama 3, which was considered the ceiling of the strength of the open source model when it was first released, has always maintained personal commercial rights.
Zuckerberg also admitted that they have been opening up the most cutting-edge large language models to commercial use, so that every small and medium-sized development company or developer can enjoy the same convenient and fast technology as large technology companies.
This is a career and effort to "teach people to fish".
Smart glasses, your next phone?
In order to make Meta AI more people use, Zuckerberg is still working hard to popularize AI on terminal devices.
When it comes to terminal devices, it has to be reminded of the biggest fish that slipped through the net on the terminal of Meta AI - Apple.
Not long ago, it was reported that Meta also negotiated a cooperation with Apple on AI, and as a member of a multi-party cooperation, which means that if Meta succeeds, then OpenAI will no longer be an AI company that exclusively enjoys billions of iPhone users around the world, but needs a share of the pie with Meta.
Although Meta's idea of multi-party cooperation to enter iPhone devices is in line with Apple's vice president Craig WWDC's post-meeting interview, according to a Bloomberg report, Apple's reason for excluding Meta is that it dislikes Meta's privacy protection is not strict enough.
Regarding this bowl of unstable dishes, Zuckerberg said something ambiguous:
"People need to release and try as many different things as possible, and one group of people can't decide everything for everyone else."
It seems that because of this failure, Zuckerberg seems to be working extra smart terminals other than mobile phones.
In the interview, Zuckerberg mentioned the progress of Meta AI's smart glasses in collaboration with Ray-Ban and is very proud to announce that the glasses are selling very well.
Regarding the development of this type of smart terminal, he boldly said that Meta AI smart glasses will have multiple development directions, such as display-less glasses, which can be used to capture content, listen to music and podcasts, or make and receive calls; There are also glasses with small displays, which can send and receive notifications, answer questions, etc., and he also mentioned that the ultimate form of these glasses may be holographic display glasses, which can be used to achieve interactive functions such as playing cards.
In addition to glasses, Zuckerberg also mentioned the prospect of their other AI-assisted devices, such as smart wristbands with neural interfaces that can receive signals from the user's brain and hands, allowing them to type and communicate without using their hands. At that time, maybe mobile phones were not a necessity.
Selected original article of the interview
What is Meta's AI strategy
Kane Sutter（@Kallaway）：
Can you talk a little bit about Meta's AI strategy, because there will be thousands of creators listening to this, they know AI, they know players, they already have some tools. But I think hear your take on Meta's AI gameplay and how it fits into the market's strategy.
Zuckerberg:
yes, so I would say that the most different thing about our approach is that many of the big companies, like OpenAI's ChatGPT and Google's Gemini, are both a major AI product. Of course, we also have a major AI assistant that exists, but it's not just our AI universe. We think people will want to interact with different people and businesses, and that requires many different AI products that reflect people's different interests. So the way that every creator and every small and medium-sized entrepreneur can create an AI that can help them is to create this underlying technology yourself, and we want you to understand these underlying technologies.
We're making Llama that is world-class, and we're very happy about that, but most importantly, it's a commercially available tool for creators and businesses to help them create the AI and personalized experiences they need. This is also the announcement of AI Studio this week, the vision of the world I want to enter.
The future of creators
Kane Sutter（@Kallaway）：
I think the world is like a network, they can be single-purpose or multi-purpose, and that's where we're going, and I don't think there will be a lot of controversy in 10 years. And I think in the future, there will be more and more creators than now, and there will be more mainstream content, and Meta has become one of the main mainstream media, what do you think the future of creators will become?
Zuckerberg:
I couldn't agree more with you, and the same is true from the perspective of human history. Many people are using their creativity, and instead of working in the traditional sense, they are doing what they really love. Tech Arc also tends to be more productive in this kind of work and to provide all sorts of new tools, and that's what we're doing.
We want more people, even those who don't think of themselves as creators, to be able to create. Because everybody has the ability to create, even my kids, they're not creators, they're just building blocks with Lego, but they're also creative, and part of our job is to create tools, for each creator to create their own creations.
Why open source
Zuckerberg:
..., so I'm convinced that there won't be just one AI in the future. There will be a lot of artificial intelligence in the future, and a lot of different people will be able to create different things. That's part of the reason why I'm so convinced of open source. Right? Because I don't think AI technology should be hoarded and then a company can use it to build whatever core monolithic product they're developing. I think if you believe that, then the best future experience will be a lot of different eyes and a lot of different experiences, and you're going to want it to get out in all those ways.
So, part of that is building tools for creators, giving platform users the ability to create their own tools, almost like the UGC type of AI situation, all commercial content. But, you know, part of it is also open source. This way, other companies can create something different, and people can hack it themselves and fiddle with it.
So I guess it's a very deep view of the world. I know I'm going to be very disgusted when people in the tech industry talk about building this real AI. It's as if they think they're creating God or something. And that's not what we're doing. We use different tools, such as creatives. Ten years from now, we will use different tools than we do today, just as the tools we use today are different from those we used ten years ago. I think this goes back to your previous main point that there will be a lot more creative work in the future than there is now.
AI terminal devices
Kane Sutter（@Kallaway）：
I'm really curious, what else do you think the Ray-Ban Meadows or smart glasses need to have? Regardless of the iteration, what other features are needed to make billions of people prefer glasses over their phones as their primary hardware devices?
Zuckerberg:
If you had asked me this question a few years ago, or even a year and a half ago, I would have said that we need to get to the level of holographic AR. It's going to be the main platform that people use, and we're working on that, right? We use Meta Ray-Ban Glasses to solve this problem from two different directions.
We looked at how much technology we could put into it if you were to just take the best form factor for today's glasses without compromising on form factor, weight, etc. That's what we get with Meta Ray-Ban Glasses. But then we also looked at it from a different angle and said, we still want to create holographic AR.
It still needs to be glasses, right? It's not headphones, it's glasses. However, maybe it will have a thicker frame because it has more technology inside. So I think those two paths will gradually converge over time. I'm guessing I would have thought before that we needed full holograms for a sense of presence, but I've made such a leap forward that I think even simpler products will quickly become more appealing.
The demand for Meta Ray-Ban virtual glasses is much higher than we expected, we don't produce enough and it's a bit frustrating, most of the styles are sold out, you can still buy that basic black style, but there are a lot of other styles that people want, all sold out.
I still think that for those who can afford a more expensive device and don't mind that it will be a little heavier, they might want a holographic device. Ten years from now, I think we're going to have very small holographic devices like this. That would be great.
Of course, we'll continue to make it smaller and smaller as time goes on. So I think it's a very exciting evolution.
I think there will basically be three different products.
First of all, there will be a display-less glasses that only have artificial intelligence that captures content, you can listen to audiobooks and music, answer phone calls, and more.
And then I think there's going to be another thing that might not be a holographic display, maybe get something like a head-up display. I think it's going to create a range of interesting use cases, you're going to be able to get notifications, you're going to be able to message people, you're going to be able to send messages with artificial intelligence, answer any question and not just have it talk to you, but also be able to see which one has higher bandwidth. So it's going to be exciting. A small screen has many uses, even if it's just a small head-up display.
And then I think there's going to be the superlative version, which is full vision, you know, and we're going to talk about that in the future. It's like a hologram sitting on the couch in your living room, or you're with me, it's not just about video calls. It's not just about having a screen and a hologram. I mean, we can interact, right? For example, if you want to play cards, it's like, okay, we have a deck of cards. This is the hologram we command. You know, we can fiddle with the same thing. You know, you want to create art together or create content or, like, you draw on a whiteboard, you can do all of these things.
Source: @Kallaway (YouTube)