On a recent episode of the Lex Fridman Podcast, Elon Musk highlighted a challenge facing all the big tech companies fighting for supremacy in the AI space. You’re only as good as the data you have access to, and increasingly, ‘you run out pretty quickly’.
Speaking on the topic of his AI initiative xAI and their Grok LLM, Musk shed light on the limitations of the data currently available:
‘It’s actually humbling to see how little data humans have actually been able to accumulate… discounting spam and repetitive stuff, it’s not a huge number.’
Beyond the user-generated data being used to train LLMs, the next big step in AI is the ability to navigate and interact with the real world: spatial data. In this area, the sensors used to capture data dictate the type of data that is available. Compared to user-generated text and still images, sophisticated sensors with continuous real-time high-definition video and LIDAR arrays capture data orders of magnitude more than previously available.
This is where Musk believes that xAI has an advantage:
With Tesla and the real-time video coming from several million cars, ultimately, tens of millions of cars. With Optimus, there might be hundreds of millions of Optimus robots, maybe billions, learning a tremendous amount from the real world.
The map of the world is growing, but the sensors are currently mounted on cars or in public spaces. Beyond our roads and public spaces is where humanity spends most of its time, in our places of business and in our homes. Tens of millions of Teslas are obviously an invaluable source of data, but ultimately, many valuable areas are still terra incognita.
Optimus, as a humanoid robot with improved utility and mobility, will be everywhere. Able to perform a multitude of tasks in homes, offices, factories and public spaces, it is no wonder that Musk thinks that;
Optimus is going to be the biggest source of data...
There is no doubt that the data that a billion of these spatial data gatherers will be able to gather invaluable and almost incalculable amounts of data. If any company is able to deploy this many sensors and have them infiltrate every aspect of daily life, then they would undoubtedly have a massive advantage over every other AI company.
However, this improved data has darker implications: it is only useful if it is used. Musk seems pretty candid about this in this interview: they're planning to spy on you.
Whereas most of us seem to have made peace with most of the content we share online being scraped, collated, and fed to AI for the purpose of training, the idea of these spatial data gatherers being invited into our homes and offices is less palatable.
Until now, the implied and sometimes explicit agreement between the tech giants and their users has been that we get something of utility for free and, in return, provide our data to them to use or sell as they see fit. The implied spying has generally been seen as a palatable trade, whether that be through more targeted ads or data being used to train LLMs or improve algorithms. And the invasion of privacy has often been made more palatable through measures such as anonymizing data and grouping users together.
However useful Optimus and Tesla cars are, the new spatial data dynamic, with real-time feeds of ‘our’ reality being piped into a data center in Memphis to fuel xAI’s drive to become the dominant AI, should give us pause.
A sensor like Optimus with an ability to not only share continuous video, audio and spatial data of our most intimate spaces but also to move freely within them is something entirely new. Facebook never had the ability to open our bedroom drawers. Beyond the privacy implications, there is the question of whether this is a fair trade. Is the old model of trading data for utility an adequate model for this transaction?
Ultimately, reality doesn’t just scale; in this case, reality will be defined by the choices we make. Do we hand over this treasure trove of spatial data in return for utility, as we have for so much of the internet's recent history, essentially handing some big tech company an unassailable lead in the AI race? Or do we try to find an alternative way that balances privacy with utility?
This is where the decentralization movement and DePINs (Decentralized Physical Infrastructure Networks) like The Posemesh offer an alternative: a way for users to own their own data, trade utility with other users on the network, and preserve privacy.
To learn more about The Posemesh, read the protocol’s whitepaper.
Aukiはポーズメッシュという地球上、そしてその先の1000億の人々、デバイス、AIのための分散型機械認識ネットワークを構築しています。ポーズメッシュは、機械やAIが物理的世界を理解するために使用可能な、外部的かつ協調的な空間感覚です。
私たちの使命は、人々の相互認知能力、つまり私たちが互いに、そしてAIとともに考え、経験し、問題を解決する能力を向上させることです。人間の能力を拡大させる最も良い方法は、他者と協力することです。私たちは、意識を拡張するテクノロジーを構築し、コミュニケーションの摩擦を減らし、心の橋渡しをします。
X | LinkedIn | Medium | YouTube | AukiLabs.com
ポーズメッシュは、分散型で、ブロックチェーンベースの空間コンピューティングネットワークを動かすオープンソースのプロトコルです。
ポーズメッシュは、空間コンピューティングが協調的でプライバシーを保護する未来をもたらすよう設計されています。いかなる組織の監視能力も制限し、空間のプライベートな地図の自己所有権を奨励します。
分散化はまた、特に低レイテンシが重要な共同ARセッションにおいて、競争優位性を有します。ポスメッシュは分散化運動の次のステップであり、成長するテック大手のパワーに対抗するものです。
アウキ・ラボはポスメッシュにより、ポーズメッシュのソフトウェア・インフラの開発を託されました。
X | Discord | LinkedIn | Medium | Updates | YouTube | Telegram | Whitepaper | DePIN Network
