The Road to AGI | AI + 3D Interaction + Web3 — Three Converging Waves Igniting a Productivity Revolution

As the era of intelligent civilization accelerates toward us, a new age inevitably demands new productive forces and new modes of production. When AI truly becomes a productivity tool, what is the corresponding tool for production relations?

Not long ago, BlueRun Ventures held a marathon internal discussion, mapping out the landscape of the intelligent civilization era and the possibilities across various sectors. We believe the coming transformation will stem from the convergence of three waves: AI, 3D interaction, and Web3, achieving innovation in the pairing of productive forces and production relations.

In the second installment of our "Path to AGI" column, we've compiled highlights from this discussion. Over an extended time horizon, we want to share where we see these waves originating —

Let's stretch our timeframe and share what major waves we see through a relatively abstract model.

In artificial intelligence, the most anticipated breakthrough is in NLP — natural language processing builds upon the accumulation of language and text, can draw upon vast reserves of existing knowledge, and interacts with humans more frequently and closer to industry applications. Once NLP achieves its breakthrough, humanity may "reach" more advanced data scenarios and train more sophisticated AI.

For years, computer vision has truly demonstrated its value only in autonomous driving; its value enhancement in security and other domains remains limited. But when CV combines with NLP, the mobilization of multimodal data will generate enormous value across industries.

We believe future transformation will come from the convergence of three waves: AI, 3D interaction, and Web3. Their combination — whether two of the three or all three — could spark innovation. From there, they can permeate various industries.

01

3D Data Creates Higher-Order Intelligence

The 3D wave began long ago, but its truly massive opportunity lies as a new interface for interaction.

3D interaction captures spatial data. CV is applied to all intelligent vehicles that assist humans — autonomous driving, robotics — using a pair of eyes (cameras) to observe the world, further exploring, understanding, and interpreting spatial data of this world, thereby generating massive amounts of 3D spatial data and multimodal data.

Take autonomous driving as an example. We previously discussed an important thesis with founders — the internet of the past was two-dimensional; humans processed data in two-dimensional space. But starting with artificial intelligence, we must process data in three-dimensional space, that is,立体空间数据 (stereoscopic spatial data).

What value does spatial data from 3D interaction bring to AI? In the past, AI+X has created considerable value, but the greater significance of AI+X is to create even more powerful AI. When AI can access 3D data, the pace of its advancement accelerates.

Current AI — ChatGPT, for instance — still operates in relatively simple, standardized service modes. But as it gains multimodal data in certain scenarios, it can iteratively develop higher-order intelligence, penetrate vertical domains, and enhance its capabilities. More data enables more advanced services, which in turn yield more advanced data to train AI, forming a positive feedback loop. This possibility has already begun emerging in many industries. This also explains why AI deserves particular attention at this stage: it still possesses potential for higher-order iteration.

For example, using ChatGPT for sales and intelligent customer service in the future would collect more feedback data; through video conferencing and cameras, additional process data — voice, expressions, body language, eye contact, micro-expressions — could be synthesized into comprehensive data to feed back into AI.

Apple's VR headset is a key variable that the market is watching closely. Some argue Apple's VR device is at best another Macintosh — the Mac elevated the PC experience from software to hardware after PCs emerged, but it didn't reinvent the PC. So it's still impossible to judge whether Apple will trigger the 3D inflection point.

Compared to AI, 3D interaction and Web3 remain at earlier stages of their development curves. But this isn't a simple curve; for instance, recent crypto price rises don't necessarily mean genuine ascent or decline — it may still be oscillating within a range before gradually returning to trend.

02

When AI Truly Becomes a Productivity Tool

As AI intelligence strengthens, the possibility of decentralization grows increasingly strong. We currently frame the relationship between AI and humans in terms of empowerment — computers, phones, smart glasses... enhancing individual capabilities. But as intelligence grows increasingly sophisticated, our relationship with intelligence will become more like collaboration, akin to how Iron Man collaborates with J.A.R.V.I.S. We judge that future AI value creation will occur at the edge. Increasingly, data, computing power, and storage may be generated at the edge rather than necessarily solving problems through centralized cloud.

When new productive forces explode onto the scene, they will disrupt traditional business models and introduce previously nonexistent problems. Because older business models can no longer contain the surplus value provided by explosive productivity growth. Before search engines emerged, we couldn't have known that paid search ranking could so dramatically improve advertising efficiency. In the future, when AI capabilities become sufficiently powerful and 3D interaction penetration sufficiently high, similar innovations may emerge.

For example, Google's current AdWords (media buying channel) and AdSense (advertising sales channel) are based on display within search interfaces — a perfectly constructed commercial product built atop search engines, but potentially soon to be disrupted. After ChatGPT partnered with Bing, what users receive is a conversation, a solution. Going forward, we may no longer need to open a search engine page; there may be nowhere to place AdWords and AdSense keywords. Could future business models become answering the user's question first, then incidentally attaching an advertisement?

Broadly speaking, the previous information revolution, the internet, was still about transmitting information. This generation of AI creates information — GPT, for example, generates. Many previous business models were based on connection; upcoming business models will be based on creation, raising certain value distribution questions. Previous human-machine collaboration resembled SaaS services, but going forward, machines will likely介入人类的思考和生产流程中 (intervene in human thinking and production processes) — it becomes reciprocal, no longer merely plug-in or auxiliary.

If AI and 3D interaction interfaces solve productivity problems, then these new, disruptive productive forces require new tools for production relations to contain them. We believe Web3 is the tool for solving production relations.

Intelligence that collaborates with humans produces value that can be distributed. How do we recognize its value? In a field with hundreds or thousands of devices, how much value does each device truly create? What proportion does it represent globally? How do we distribute accordingly? This implicates questions of rights confirmation — and specifically, global, decentralized rights confirmation.

03

Envisioning Web3.0 as a Tool for Production Relations

Why this moment to discuss what changes are occurring as AI becomes a productivity tool? Where will the convergence of AI + Web3 generate value in the future? We offer some conjectures through deductive reasoning, and explain the definitions, relationships, and applicable scope of productivity and production relations tools in this context.

For example, suppose a machine posts an answer on Reddit or Zhihu. That answer gets used in a PowerPoint. Assuming the answer is worth $1,000, should the machine share in the proceeds? If an AI-designed architectural blueprint gets used by an architect, how should that be calculated?

Our past solution to such problems was establishing intellectual property rights. But in the virtual world, traditional intellectual property frameworks don't apply — what is needed is a transformation in production relations. Production relations ultimately concern ownership, value distribution, and organizational methods.

We identify four key dimensions of change in production relations: content rights confirmation for machines, content rights confirmation for humans, global identity verification for machines, and eliminating bias against machines.

These changes involve: first, granting AI equal status. We envision that humans and machines will become equal units in the future; humans and machines, machines and machines, can collaborate in modular combinations.

Second, verifying that machine's identity and distributing value to it. Only through payment will the machine or the organization behind it have greater motivation to work. We imagine that an original artwork when generated might itself be an NFT, carrying its own token. Every view, share, or remake gets counted. Counters are what Web3 is solving on the ecological chain — achievable in the future at very low cost and relatively high efficiency.

Global rights confirmation can be realized through blockchain. What exactly is blockchain? It can be understood as a global Excel spreadsheet, or a database; so each person's ID is globally unique, enabling global rights confirmation. This table requires enormous cost and is controlled by no one — something humanity truly couldn't have imagined before.

A Bonus

Robotics and autonomous driving are important means of capturing 3D spatial data. Through hardware form factors interacting with the environment, they can better feed back into AI algorithms. Currently, the forms carrying AGI capability are predominantly software-based. When AGI combines with Robotics, what revolution will ensue?

Leave a comment to discuss with us, or join the BlueRun AGI Pioneer Club and ask hundreds of AGI pioneers what they think —

Are you an AGI pioneer? Scan the QR code to join "BlueRun AGI Pioneer Club" 👇👇👇

More Content

Originating in Silicon Valley, BlueRun Ventures was established in 2005 as a venture capital firm focused on early-stage startups.

Currently, BlueRun Ventures manages multiple USD and RMB dual-currency funds in China, with assets under management exceeding RMB 15 billion, making it one of the largest early-stage funds domestically. The firm invests primarily at Pre-A and Series A stages, covering hard tech and innovative interaction, enterprise technology, new consumption, and healthcare. It has cumulatively invested in over 150 startups, including Li Auto, Waterdrop, QingCloud, Guazi.com, Qudian, Songguo Mobility, Ganji.com, Energy Monster, Yuntu Semiconductor, Machenike, Clouds Intelligence, Anxin Network Shield, and BioMap.

BlueRun Ventures has been ranked #1 on Zero2IPO's "China's Top 30 Early-Stage Investment Institutions," #1 on ChinaVenture's "China's Best Early-Stage Venture Capital Institutions TOP30," and named among Preqin's Top 10 VC Fund Managers Globally for Sustained High Returns.

Additionally, BlueRun Ventures has received consecutive honors from Forbes China, 36Kr, Cyzone, Caixin Media, CBNweekly, Jiemian, and other media institutions, including "China's Best Early-Stage Institution of the Year," "China's Top Venture Capital Firm," "Most Entrepreneur-Friendly Early-Stage Institution of the Year," and "Most Influential Early-Stage Institution of the Year."