youtube .com/watch?v=KbZfbqHeJNU
Fresh off my scathing critique of Mark Gurman (which I assure you doesn't even begin to encompass the issues his work has), I will now move onto critiquing the next false narrative constantly pushed by him and others recently: "Apple has no clue about AI."
First, "AI" is a complete made up term as it's being used today. What is it actually?
People are often using it to describe machine learning, specifically neural networks, specifically transformer models. These are often used as chatbots to give the appearance of intelligence (the language fallacy: if it speaks it therefore must be intelligent).
Transformer models are a new, highly specialized form of neural network, and while this isn't the focus of what I'm writing, I needed to clarify that before I continue
Apple may not yet excel at transformer models (but are catching up, according to their latest releases), but what they do excel at is machine learning, which is arguably far more important to technology and its role in our lives than only transformer models. That is because it encompasses a broad range of "intelligent" models, not only transformers.
This brings me to the very underrated and understated and oft-undiscussed (up to now) brand new UI: spatial. The hallmark of it being blending digital content into your real world. Not only is this UI only possible through machine learning, the experiences it enables are simply impossible on any other interface (GUI, etc).
One of the hallmark first features of spatial is the ability to do holographic video chats.
Simply put, you are no longer limited by distance when you want to be in the same room as your friends and family. Your friends and family can appear in the same room as you, even though you're thousands of miles away
By using the array of sensors Apple custom-designed and engineered, Apple 3D scans your face. This scanning process already uses machine learning to coordinate sensors and tell users what they need to do (by voice direction).
It goes further. Unlike what people normally think, Apple does not create an "avatar" of you necessarily -- meaning, they do not scan your face and then determine a cartoon likeness of you.
Instead, in a world's first, they scan your face and upper torso using LiDAR, high definition cameras, and Face ID on device; then, using a sophisticated machine learning technique of over 12 highly advanced neural networks unified and working in tandem to fuel a photo rendering technique Gaussian Splatting , it digitally recreates you in 3 dimensions as if you were really there. A real hologram.
It goes further. Unlike Star Wars and virtually every other TV/movie concept where holographic video calls are possible, this does not have any artifacts. There is no "line scanning," no visual glitches, no odd coloring. It is you, simply represented as if you were really in front of your friends and family
Not only is this possible, but it is completely created on device, never sent to any servers for processing, completely private, and accomplishes this feat in mere seconds.
Lastly, a wide array of cameras, lasers, and sensors on the outside and inside constantly track every single movement you do. If you smile, it can see that. If you raise your eyebrows, it can see that. If you laugh with your teeth showing, it can see that. If you stick your tongue out, it can see that.
If you stand up, it can sense that. If you sit down, it can sense that. If you walk around, it can sense that. If you jump up and down, it can sense that. Everything you do it sees and senses accurately.
Finally, Using even more machine learning, it translates this computer vision and sensor fusion into real actions performed with your hologram. So if you smile, frown, laugh, smirk, jump, and move around, it replicates this exactly as if you were doing this in the real world
Sensor fusion and Computer vision is done and interpreted within milliseconds, transmitted over the internet, and shown to other people in what people have described as an 'instant, lag free experience; it feels like you're really in the room with them.'
The result is something that has never been accomplished before for consumers (or even enterprise): the ability to conduct holographic video calls, to be in the same room as your friends and family even if you aren't.
All of this is accomplished with a product that's easy to use as tapping, available in the Apple Store today, and is at $3,500.
This feature has never been available before to anyone: consumers, enterprise, governments, military, etc. and yet it's so simple and easy to use, you simply just start a FaceTime call.
Apple is not behind in machine learning (or "AI"). It merely chooses to focus on and make products that enhance and enrich people's lives, not destroys the earth and creativity
No one else has done this before. If anything has come close, it's always been restricted to labs, and never shipped to consumers. It never was done on device, nor as simple and easy to use like this. A world first.
A chatbot that tells you whatever you want to hear, or the ability to connect with your friends and family in ways literally never before possible.
What do you choose?
The first video above has the world's first holographic interview with Apple executives (in depth). I recommend turning the video quality up if not set automatically.
The second video below is from another YouTuber showing off what it's like to use by himself and with his friends and coworkers. Turn on subtitles too.
A reviewer uses it with his friends:
youtube .com/watch?v=zChhRT-y08Q&t=760s
Set up:
twitter .com/M1Astra/status/1724539857752519028
I welcome all discussion: technical discussions around Apple machine learning related to this, discussion about the feature, utility, etc.
Fresh off my scathing critique of Mark Gurman (which I assure you doesn't even begin to encompass the issues his work has), I will now move onto critiquing the next false narrative constantly pushed by him and others recently: "Apple has no clue about AI."
First, "AI" is a complete made up term as it's being used today. What is it actually?
People are often using it to describe machine learning, specifically neural networks, specifically transformer models. These are often used as chatbots to give the appearance of intelligence (the language fallacy: if it speaks it therefore must be intelligent).
Transformer models are a new, highly specialized form of neural network, and while this isn't the focus of what I'm writing, I needed to clarify that before I continue
Apple may not yet excel at transformer models (but are catching up, according to their latest releases), but what they do excel at is machine learning, which is arguably far more important to technology and its role in our lives than only transformer models. That is because it encompasses a broad range of "intelligent" models, not only transformers.
This brings me to the very underrated and understated and oft-undiscussed (up to now) brand new UI: spatial. The hallmark of it being blending digital content into your real world. Not only is this UI only possible through machine learning, the experiences it enables are simply impossible on any other interface (GUI, etc).
One of the hallmark first features of spatial is the ability to do holographic video chats.
Simply put, you are no longer limited by distance when you want to be in the same room as your friends and family. Your friends and family can appear in the same room as you, even though you're thousands of miles away
By using the array of sensors Apple custom-designed and engineered, Apple 3D scans your face. This scanning process already uses machine learning to coordinate sensors and tell users what they need to do (by voice direction).
It goes further. Unlike what people normally think, Apple does not create an "avatar" of you necessarily -- meaning, they do not scan your face and then determine a cartoon likeness of you.
Instead, in a world's first, they scan your face and upper torso using LiDAR, high definition cameras, and Face ID on device; then, using a sophisticated machine learning technique of over 12 highly advanced neural networks unified and working in tandem to fuel a photo rendering technique Gaussian Splatting , it digitally recreates you in 3 dimensions as if you were really there. A real hologram.
It goes further. Unlike Star Wars and virtually every other TV/movie concept where holographic video calls are possible, this does not have any artifacts. There is no "line scanning," no visual glitches, no odd coloring. It is you, simply represented as if you were really in front of your friends and family
Not only is this possible, but it is completely created on device, never sent to any servers for processing, completely private, and accomplishes this feat in mere seconds.
Lastly, a wide array of cameras, lasers, and sensors on the outside and inside constantly track every single movement you do. If you smile, it can see that. If you raise your eyebrows, it can see that. If you laugh with your teeth showing, it can see that. If you stick your tongue out, it can see that.
If you stand up, it can sense that. If you sit down, it can sense that. If you walk around, it can sense that. If you jump up and down, it can sense that. Everything you do it sees and senses accurately.
Finally, Using even more machine learning, it translates this computer vision and sensor fusion into real actions performed with your hologram. So if you smile, frown, laugh, smirk, jump, and move around, it replicates this exactly as if you were doing this in the real world
Sensor fusion and Computer vision is done and interpreted within milliseconds, transmitted over the internet, and shown to other people in what people have described as an 'instant, lag free experience; it feels like you're really in the room with them.'
The result is something that has never been accomplished before for consumers (or even enterprise): the ability to conduct holographic video calls, to be in the same room as your friends and family even if you aren't.
All of this is accomplished with a product that's easy to use as tapping, available in the Apple Store today, and is at $3,500.
This feature has never been available before to anyone: consumers, enterprise, governments, military, etc. and yet it's so simple and easy to use, you simply just start a FaceTime call.
Apple is not behind in machine learning (or "AI"). It merely chooses to focus on and make products that enhance and enrich people's lives, not destroys the earth and creativity
No one else has done this before. If anything has come close, it's always been restricted to labs, and never shipped to consumers. It never was done on device, nor as simple and easy to use like this. A world first.
A chatbot that tells you whatever you want to hear, or the ability to connect with your friends and family in ways literally never before possible.
What do you choose?
The first video above has the world's first holographic interview with Apple executives (in depth). I recommend turning the video quality up if not set automatically.
The second video below is from another YouTuber showing off what it's like to use by himself and with his friends and coworkers. Turn on subtitles too.
A reviewer uses it with his friends:
youtube .com/watch?v=zChhRT-y08Q&t=760s
Set up:
twitter .com/M1Astra/status/1724539857752519028
I welcome all discussion: technical discussions around Apple machine learning related to this, discussion about the feature, utility, etc.
Last edited:
