a16z 2024 Outlook List: Modularization, AI, Web3 Games..

Author: a16z

Compiled by: 1912212.eth, Foresight News

Based on feedback from partners in fields such as American Dynamism, biotech, consumer technology, crypto, enterprise, fintech, gaming, and infrastructure, we have released a comprehensive list of major ideas that tech builders may explore in the coming year. Here are some exciting trends that crypto partners believe will emerge in 2024.

Entering a New Era of Decentralization

As we have seen time and again, when control of a powerful system or platform is held by a few (let alone a single leader), it becomes too easy to infringe on user freedoms. This is why decentralization is important: it is a tool that can democratize systems by implementing trustworthy, neutral, and composable internet infrastructure; fostering competition and diversity in ecosystems; and providing users with more choices and greater ownership.

However, in practice, achieving decentralization at scale has been challenging, especially compared to the efficiency and stability of centralized systems. Meanwhile, most Web3 governance models involve DAOs, which use simplified yet cumbersome governance models based on direct democracy or corporate governance that do not fit the sociopolitical realities of decentralized governance.

Nonetheless, due to the experimental labs of Web3 activity over the past few years, best practices for decentralization have begun to emerge. These practices include adapting decentralized models for applications with richer functionalities; incorporating Machiavellian principles in DAOs to design more effective decentralized governance that holds leadership accountable. As these models evolve, we should soon see unprecedented levels of decentralized coordination, operational capabilities, and innovation.

---Miles Jennings, General Counsel and Head of Decentralization (@milesjennings on Farcaster | on Twitter)

Reconstructing the Future User Experience

Despite the user experience in the crypto space being criticized since 2016, its fundamental principles have not changed much. It remains overly complex: self-custody of keys; connecting wallets to decentralized applications (dApps); sending signed transactions to an increasing number of network endpoints, etc. This is something we cannot expect users to learn in the first few minutes of using crypto applications.

However, developers are actively testing and deploying new tools that can reset the crypto front-end user experience in the coming year. One such tool includes simplifying the passkeys for logging into applications and websites; unlike passwords that require user manual work, passkeys are automatically generated cryptographic keys. Other innovations include smart accounts, making accounts themselves programmable and thus easier to manage; embedded wallets, built into applications, making onboarding frictionless; multiparty computation, making it easier for third parties to support signatures without holding user keys; and advanced RPC (Remote Procedure Call) endpoints that can identify user needs and fill gaps, among others. All of these not only help Web3 to be more widely adopted but also make the user experience better and more secure than in Web2.

---Eddy Lazzarin, Chief Technology Officer (@eddy on Farcaster | @eddylazzarin on Twitter)

The Rise of Modular Tech Stacks

In the world of networks, one force always dominates others: network effects. Network effects are often so powerful that there are essentially only two types of modularization: one that expands and strengthens network effects; the other that disrupts and weakens network effects. In all but a few cases, only the former makes sense, especially when it comes to open source.

Monolithic architectures have the advantage of allowing deep integration and optimization at what would otherwise be modular boundaries, thereby improving performance… at least initially. But the greatest advantage of open-source, modular tech stacks is that they unlock permissionless innovation; allow participants to focus on specific areas; and incentivize more competition. In this world, we need more of that.

---Ali Yahya, Partner (@alive.eth on Farcaster | @alive_eth on Twitter)

The Convergence of AI and Blockchain

Decentralized blockchains serve as a counterbalance to centralized AI. Currently, AI models (such as ChatGPT) can only be trained and operated by a handful of tech giants, as the required computing power and training data are prohibitively expensive for smaller players. However, through crypto, it is possible to create multi-sided, global, permissionless markets where anyone can contribute computing power or new datasets for a need in the network and be compensated. The long tail of utilizing these resources will lower the cost of AI, making it more accessible.

But as AI changes the way we produce information, transforming society, culture, politics, and economics, it also creates a rich world of AI-generated content, including deepfakes. Cryptographic technology can also play a role here, opening the black box; tracking the origins of what we see online; and so on. We also need to find ways to govern distributed generative AI democratically so that no single participant ultimately decides for everyone else; Web3 is the laboratory for solving this issue. Decentralized, open-source crypto networks will democratize AI innovation (rather than centralizing it), ultimately making it safer for consumers.

---Andy Hall, Professor at Stanford University (@ahallresearch); Daren Matsuoka, Data Scientist (@darenmatsuoka on Farcaster | on Twitter); Ali Yahya, Partner (@alive.eth on Farcaster | @aliveeth on Twitter)

Play-to-Earn Transitions to Play-and-Earn

In play-to-earn games, players can often earn money in the real world (not just virtually) based on the time and effort they spend in the game. This trend is related to broader transformations that are changing gaming and its surrounding fields, from the rise of the creator economy to shifts in the relationship between people and platforms. Web3 enables us to break the conventional practice where all the profits from playing games and conducting transactions flow only to the gaming companies. Users spend significant time on these platforms and create substantial value for them, so they should also be compensated.

However, gaming does not necessarily have to become a workplace (at least for most players). What we really need are games that are fun and allow players to capture more of the value they create. Therefore, play-to-earn is increasingly evolving into play-and-earn, making an important distinction between gaming and the workplace. As play-to-earn games move beyond their initial growth phase, the dynamics of how the gaming economy is managed will continue to change. However, ultimately, this will not be a fragmented trend but rather a part of gaming.

---Arianna Simpson, @AriannaSimpson

When AI Becomes the Game Maker, Cryptocurrency Provides Assurance

As someone who spends a lot of time thinking about Web3 gaming and the future of games, it is clear to me that AI agents in games must provide assurances: they are based on specific models, and those models are not tampered with during execution. Otherwise, the integrity of the game will be compromised.

When legends, terrains, narratives, and logic are all procedurally generated, in other words, when AI becomes the game maker, we will want to know that the game maker is a trustworthy neutral. We will want to know that this world is built on a foundation of guarantees. What crypto provides most importantly are these guarantees—including the ability to understand, diagnose, and penalize when AI goes awry. In this sense, AI alignment is actually an incentive design problem, just as dealing with any human agent is an incentive design problem… and that is precisely what cryptocurrencies focus on.

---Carra Wu, Partner (@carra on Farcaster, @carrawu on Twitter)

Formal Verification Becomes Less Formal

While formal methods are popular for verifying hardware systems, they are less common in software development. For most developers not dealing with hard or safety-critical systems, these methods are too complex and can add significant costs and delays. However, smart contract developers have different needs: the systems they develop handle billions of dollars; vulnerabilities can have catastrophic consequences and often cannot be patched instantly. Therefore, there is a need for more accessible formal verification methods in software development, especially in smart contract development.

In the past year, we have seen a wave of new tools (including our own) emerge that offer a much better development experience than traditional formal systems. These tools leverage the fact that smart contracts are architecturally simpler than conventional software—having atomic and deterministic execution; no concurrency or exceptions; low memory usage and fewer loops. The performance of these tools is also rapidly improving, utilizing the latest breakthroughs in SMT solver performance (SMT solvers use complex algorithms to identify or confirm the existence of errors in software and hardware logic). As developers and security experts widely adopt tools inspired by formal methods, we can expect the next wave of smart contract protocols to be more robust and less susceptible to costly hacks.

---Karma (Daniel Reynaud), Research Engineering Partner (@karma on Farcaster, @0xkarmacoma on Twitter)

NFTs Become Ubiquitous Brand Assets

An increasing number of well-known brands have begun introducing digital assets to mainstream consumers in the form of NFTs. For example, Starbucks has introduced a gamified loyalty program where participants collect digital assets while exploring the company's coffee products (not to mention the AR pumpkin spice maze!). Meanwhile, Nike and Reddit have developed digital collectibles NFTs explicitly marketed to a broad audience. But brands can do much more than that: they can leverage NFTs to represent and reinforce customer identities and community relationships; connect physical goods with their digital representations; and even co-create new products and experiences with their most loyal fans.

Last year, we saw a trend of low-cost NFTs being collected on a large scale as consumer goods—these NFTs are often managed through custodial wallets and/or "Layer 2" blockchains, resulting in correspondingly low transaction costs. As we enter 2024, the conditions for NFTs to become ubiquitous digital brand assets are in place, as Steve Kaczynski and I explain in an upcoming book, applicable to various companies and communities.

---Scott Duke Kominers, Research Partner (@skominers on Farcaster | on Twitter)

SNARKs Become Mainstream

Historically, technical experts have had several strategies for verifying computational workloads:

1) Re-executing computations on trusted machines;

2) Executing computations on machines specifically designed for the task, i.e., Trusted Execution Environments (TEEs); or

3) Executing computations on reliable neutral infrastructures, such as blockchains. Each strategy has limitations in terms of cost or network scalability, but now, SNARKs (Succinct Non-interactive Arguments of Knowledge) have become more accessible. SNARKs allow untrusted "provers" to compute cryptographic receipts for some computational workloads in an unforgeable manner: in the past, the cost of computing such receipts was 10^9 times higher than the original computation; recent advancements are bringing this number down to around 10^6.

Thus, in scenarios where the initial computation provider can afford a 10^6 overhead while the client cannot re-execute or store the initial data, SNARKs become feasible. The resulting use cases are numerous: edge devices in IoT can verify upgrades. Media editing software can embed authenticity and transformation data; and mixed media emojis can pay homage to their original sources. LLM reasoning can include authenticity information. We could have self-verifying tax forms, unforgeable bank audits, and many more consumer-beneficial uses.

---Sam Ragsdale, Investment Engineer (@samrags on Farcaster, @samrags_ on Twitter)