Solana Foundation: Three Key Strategic Directions for the Integration of AI and Crypto

By: blockbeats|2024/12/04 18:00:01

Original Article Title: Focus areas at the intersection of crypto and AI

Original Article Author: Kuleen, Head of DePIN at Solana Foundation

Original Article Translation: Yuliya, PANews

The intersection of AI and cryptographic technology is currently entering a "Cambrian explosion" experimental stage. This article by the Solana Foundation elaborates on the three key development directions of AI + crypto integration.

TLDR

1. Building the most vibrant AI Agent-Driven Economy on Solana

Truth Terminal has already demonstrated the feasibility of AI agents operating on-chain. The experimentation in this area continues to push the boundaries of on-chain agent operation, showcasing not only immense potential but also a wide design space. This has already emerged as one of the most groundbreaking and explosive directions in the intersection of crypto and AI, and it is only the beginning.

2. Enhancing LLM's Capabilities in Solana Code Development

Large Language Models have already shown great promise in code writing and are expected to further improve in the future. Through these capabilities, Solana developers' efficiency is expected to increase by 2-10 times. Recently, establishing high-quality benchmarks to assess LLMs' understanding and writing of Solana code will help understand the potential impact of LLMs on the Solana ecosystem. High-quality model fine-tuning approaches will be validated in benchmark testing.

3. Supporting an Open and Decentralized AI Technology Stack

The "open and decentralized AI technology stack" includes the following key elements:

· Training data acquisition

· Training and inference computing capabilities

· Model weight sharing

· Model output verification capabilities

The importance of this open AI technology stack is reflected in:

· Accelerating model development innovation and experimentation

· Providing an alternative for users distrustful of centralized AI

1. Building the most vibrant AI Agent-Driven Economy

There has been significant discussion about Truth Terminal and $GOAT, making further elaboration unnecessary here. However, it is certain that when AI agents start engaging in on-chain activities, a world full of possibilities has already unfolded (it is worth noting that currently agents have not even directly taken action on-chain yet).

Solana Foundation: Three Key Strategic Directions for the Integration of AI and Crypto

While it is currently difficult to predict the future development of on-chain agents' behavior, by observing the innovation that has already taken place on Solana, we can catch a glimpse of the vast potential of this design space:

· AI projects like Truth Terminal are developing new digital communities through meme coins like $GOAT

· Platforms such as Holoworld AI, vvaifu.fun, Top Hat AI, Alethea AI, among others, enable users to easily create and deploy intelligent agents and their associated tokens

· AI fund managers trained on prominent crypto investor traits are emerging, with the rapid rise of ai16z on the daos.fun platform, creating a new ecosystem for AI funds and agent supporters

· Furthermore, game platforms like Colony allow players to engage in unexpected innovative gameplay by guiding agent actions in the game

Future Directions

In the future, intelligent agents can manage complex projects that require multi-party economic coordination. For example, in the field of scientific research, agents could be tasked with searching for therapeutic compounds for specific diseases. Specifically:

· Fundraising through the Pump Science platform

· Utilizing raised funds to pay access fees for research data, computational costs for compound simulations on decentralized computing networks like kuzco, Render Network, io.net

· Recruiting human participants for experimental validation work through bounty platforms like Gib.Work (e.g., running experiments to validate/build upon simulation results)

In addition to complex projects, agents can also perform simple tasks such as building personal websites, creating artistic works (e.g., zerebro), expanding the limitless possibilities of their applications.

Why is On-Chain Agent Execution of Financial Activities More Meaningful Than Using Traditional Channels?

Agents can fully utilize both traditional financial channels and cryptocurrency systems simultaneously. However, cryptocurrency holds unique advantages in certain areas:

· Micropayment applications—Solana has excelled in this area, as demonstrated by applications like Drip

· Speed Advantage - instant settlement functionality, helping proxies achieve maximum capital efficiency

· Access to Capital Markets via DeFi - this may be the most compelling reason for proxies to participate in the crypto economy. When proxies need to engage in financial activities beyond payments, the advantages of cryptocurrency become even more apparent. Proxies can seamlessly mint assets, trade, invest, participate in yield farming, engage in lending operations, leverage, and more. Particularly, Solana, with its mainnet already hosting a myriad of top-tier DeFi infrastructure, is well-suited to support these capital market activities.

From a technological development perspective, path dependence plays a critical role. It is not so much about whether a product is optimal, but rather about who can first reach critical mass and become the default choice. As more and more proxies derive revenue from cryptocurrency, cryptographic connectivity is likely to become a core competency for proxies.

What the Foundation Hopes to See

The Solana Foundation hopes to see proxies equipped with crypto wallets engage in bold innovation experiments on-chain. The Foundation does not overly constrain specific directions here because the possibilities are so vast - the most interesting and valuable proxy use cases are likely still unforeseen.

However, the Foundation is particularly interested in exploring the following directions:

1. Risk Management Mechanism

· While the current model is impressive, it is still far from perfect

· Proxies cannot be given unrestricted freedom of action

2. Driving Non-Speculative Use Cases

· Purchasing tickets via xpticket

· Optimizing stablecoin portfolio yields

· Ordering food on DoorDash

3. Development Progress Requirements

· Must have at least reached the prototype stage on the testnet

· Ideally already operational on the mainnet

2. Enhance LLMs' Ability to Write Solana Code, Empowering Solana Developers

LLMs have already demonstrated strong capabilities and are progressing rapidly. In the domain of applications for LLMs, the field of writing code may see a particularly steep improvement curve, as it is a task that can be objectively evaluated. As described below, "programming has a unique advantage: the potential for superhuman data expansion through 'self-play.' Models can write code and run, or write code, write tests, and then check self-consistency."

Today, although LLMs are still not perfect in terms of coding, exhibiting clear shortcomings (e.g., performing poorly in bug detection), AI-native code editors like Github Copilot and Cursor have fundamentally changed software development (even altering the way companies recruit talent). Considering the expected rapid rate of progress, these models are likely to completely transform software development. The Foundation aims to leverage this progress to increase the efficiency of Solana developers by an order of magnitude.

However, there are currently several challenges hindering LLMs from reaching an outstanding level of understanding in the Solana context:

· Lack of high-quality raw training data

· Insufficient number of Verified builds

· Lack of high-information-value interactions on platforms like Stack Overflow

· Historically rapid development of Solana infrastructure, meaning that code written even 6 months ago may not be entirely suitable for today's needs

· Lack of an evaluation model for understanding the level of Solana comprehension

The Foundation would like to see

· Help in acquiring better Solana data on the internet

· More teams releasing Verified builds

· More people in the ecosystem actively asking good questions and providing high-quality answers on Stack Exchange

· Creation of high-quality benchmarks for evaluating LLMs' understanding of Solana (RFP coming soon)

· Development of LLM fine-tuned models that perform well in the aforementioned benchmark tests, and more importantly, accelerate the efficiency of Solana developers. Once high-quality benchmarks are in place, the Foundation may incentivize the first model to reach the benchmark threshold score

The ultimate significant achievement will be: a brand-new, high-quality, differentiated Solana validator node client created entirely by AI.

3. Supporting an Open and Decentralized AI Tech Stack

In the field of AI, the long-term power balance between open-source and closed-source models remains unclear. Indeed, there are arguments supporting closed entities that will continue to lead the technological frontier and capture the primary value of foundational models. The simplest current expectation is to maintain the status quo—tech giants like OpenAI and Anthropic driving cutting-edge developments, while open-source models rapidly catch up and gain unique advantages through fine-tuning in specific application scenarios.

The Foundation is committed to closely integrating Solana with the open-source AI ecosystem. Specifically, this means supporting access to the following elements:

· Training Data

· Training and Inference Compute Power

· Model Weights

· Model Output Verification Capability

The significance of this strategy is reflected in:

1. Open-Source Models Accelerating Innovation Iteration

The rapid improvement and fine-tuning of open-source models like Llama by the open-source community demonstrate how the community effectively complements the work of large AI companies, pushing the boundaries of AI capabilities (even a Google researcher noted last year, "We don't have a moat about open source, nor does OpenAI"). The Foundation believes that a thriving open-source AI technology stack is crucial to accelerating progress in this field.

2. Providing Choice for Users of Trustless AI

AI may be the most potent tool in the arsenal of autocratic or authoritarian regimes. Nationally recognized models offer an officially sanctioned "truth" and are a significant vehicle of control. Highly authoritarian regimes may have superior models because they are willing to disregard citizen privacy to train AI. The use of AI for control is an inevitable trend, and the Foundation hopes to be proactive and fully support the open-source AI technology stack.

Several projects in the Solana ecosystem are already supporting the open AI technology stack:

· Data Collection — Grass and Synesis One are advancing data collection

· Decentralized Compute Power — kuzco, Render Network, io.net, Bless Network, Nosana, and others

· Decentralized Training Frameworks — Nous Research, Prime Intellect

The Foundation Looks Forward To

Hoping to see more products built at various levels of the open-source AI technology stack:

· Decentralized Data Collection: e.g., Grass, Datahive, Synesis One

· On-Chain Identity: Supporting protocols that validate wallet-held identities, protocols that validate AI API responses, enabling users to confirm they are interacting with LLM

· Decentralized Training: Projects such as EXO Labs, Nous Research, and Prime Intellect

· IP Infrastructure: Allowing AI to license (and pay for) the content it uses

Original Source Link

Editor's Note: Citrini7's cyberpunk-themed AI doomsday prophecy has sparked widespread discussion across the internet. However, this article presents a more pragmatic counter perspective. If Citrini envisions a digital tsunami instantly engulfing civilization, this author sees the resilient resistance of the human bureaucratic system, the profoundly flawed existing software ecosystem, and the long-overlooked cornerstone of heavy industry. This is a frontal clash between Silicon Valley fantasy and the iron law of reality, reminding us that the singularity may come, but it will never happen overnight.

The following is the original content:

Renowned market commentator Citrini7 recently published a captivating and widely circulated AI doomsday novel. While he acknowledges that the probability of some scenes occurring is extremely low, as someone who has witnessed multiple economic collapse prophecies, I want to challenge his views and present a more deterministic and optimistic future.

Never Underestimate "Institutional Inertia"

In 2007, people thought that against the backdrop of "peak oil," the United States' geopolitical status had come to an end; in 2008, they believed the dollar system was on the brink of collapse; in 2014, everyone thought AMD and NVIDIA were done for. Then ChatGPT emerged, and people thought Google was toast... Yet every time, existing institutions with deep-rooted inertia have proven to be far more resilient than onlookers imagined.

When Citrini talks about the fear of institutional turnover and rapid workforce displacement, he writes, "Even in fields we think rely on interpersonal relationships, cracks are showing. Take the real estate industry, where buyers have tolerated 5%-6% commissions for decades due to the information asymmetry between brokers and consumers..."

Seeing this, I couldn't help but chuckle. People have been proclaiming the "death of real estate agents" for 20 years now! This hardly requires any superintelligence; with Zillow, Redfin, or Opendoor, it's enough. But this example precisely proves the opposite of Citrini's view: although this workforce has long been deemed obsolete in the eyes of most, due to market inertia and regulatory capture, real estate agents' vitality is more tenacious than anyone's expectations a decade ago.

A few months ago, I just bought a house. The transaction process mandated that we hire a real estate agent, with lofty justifications. My buyer's agent made about $50,000 in this transaction, while his actual work — filling out forms and coordinating between multiple parties — amounted to no more than 10 hours, something I could have easily handled myself. The market will eventually move towards efficiency, providing fair pricing for labor, but this will be a long process.

I deeply understand the ways of inertia and change management: I once founded and sold a company whose core business was driving insurance brokerages from "manual service" to "software-driven." The iron rule I learned is: human societies in the real world are extremely complex, and things always take longer than you imagine — even when you account for this rule. This doesn't mean that the world won't undergo drastic changes, but rather that change will be more gradual, allowing us time to respond and adapt.

The Software Industry Has "Infinite Demand" for Labor

Recently, the software sector has seen a downturn as investors worry about the lack of moats in the backend systems of companies like Monday, Salesforce, Asana, making them easily replicable. Citrini and others believe that AI programming heralds the end of SaaS companies: one, products become homogenized, with zero profits, and two, jobs disappear.

But everyone overlooks one thing: the current state of these software products is simply terrible.

I'm qualified to say this because I've spent hundreds of thousands of dollars on Salesforce and Monday. Indeed, AI can enable competitors to replicate these products, but more importantly, AI can enable competitors to build better products. Stock price declines are not surprising: an industry relying on long-term lock-ins, lacking competitiveness, and filled with low-quality legacy incumbents is finally facing competition again.

From a broader perspective, almost all existing software is garbage, which is an undeniable fact. Every tool I've paid for is riddled with bugs; some software is so bad that I can't even pay for it (I've been unable to use Citibank's online transfer for the past three years); most web apps can't even get mobile and desktop responsiveness right; not a single product can fully deliver what you want. Silicon Valley darlings like Stripe and Linear only garner massive followings because they are not as disgustingly unusable as their competitors. If you ask a seasoned engineer, "Show me a truly perfect piece of software," all you'll get is prolonged silence and blank stares.

Here lies a profound truth: even as we approach a "software singularity," the human demand for software labor is nearly infinite. It's well known that the final few percentage points of perfection often require the most work. By this standard, almost every software product has at least a 100x improvement in complexity and features before reaching demand saturation.

I believe that most commentators who claim that the software industry is on the brink of extinction lack an intuitive understanding of software development. The software industry has been around for 50 years, and despite tremendous progress, it is always in a state of "not enough." As a programmer in 2020, my productivity matches that of hundreds of people in 1970, which is incredibly impressive leverage. However, there is still significant room for improvement. People underestimate the "Jevons Paradox": Efficiency improvements often lead to explosive growth in overall demand.

This does not mean that software engineering is an invincible job, but the industry's ability to absorb labor and its inertia far exceed imagination. The saturation process will be very slow, giving us enough time to adapt.

Redemption of "Reindustrialization"

Of course, labor reallocation is inevitable, such as in the driving sector. As Citrini pointed out, many white-collar jobs will experience disruptions. For positions like real estate brokers that have long lost tangible value and rely solely on momentum for income, AI may be the final straw.

But our lifesaver lies in the fact that the United States has almost infinite potential and demand for reindustrialization. You may have heard of "reshoring," but it goes far beyond that. We have essentially lost the ability to manufacture the core building blocks of modern life: batteries, motors, small-scale semiconductors—the entire electricity supply chain is almost entirely dependent on overseas sources. What if there is a military conflict? What's even worse, did you know that China produces 90% of the world's synthetic ammonia? Once the supply is cut off, we can't even produce fertilizer and will face famine.

As long as you look to the physical world, you will find endless job opportunities that will benefit the country, create employment, and build essential infrastructure, all of which can receive bipartisan political support.

We have seen the economic and political winds shifting in this direction—discussions on reshoring, deep tech, and "American vitality." My prediction is that when AI impacts the white-collar sector, the path of least political resistance will be to fund large-scale reindustrialization, absorbing labor through a "giant employment project." Fortunately, the physical world does not have a "singularity"; it is constrained by friction.

We will rebuild bridges and roads. People will find that seeing tangible labor results is more fulfilling than spinning in the digital abstract world. The Salesforce senior product manager who lost a $180,000 salary may find a new job at the "California Seawater Desalination Plant" to end the 25-year drought. These facilities not only need to be built but also pursued with excellence and require long-term maintenance. As long as we are willing, the "Jevons Paradox" also applies to the physical world.

Towards Abundance

The goal of large-scale industrial engineering is abundance. The United States will once again achieve self-sufficiency, enabling large-scale, low-cost production. Moving beyond material scarcity is crucial: in the long run, if we do indeed lose a significant portion of white-collar jobs to AI, we must be able to maintain a high quality of life for the public. And as AI drives profit margins to zero, consumer goods will become extremely affordable, automatically fulfilling this objective.

My view is that different sectors of the economy will "take off" at different speeds, and the transformation in almost all areas will be slower than Citrini anticipates. To be clear, I am extremely bullish on AI and foresee a day when my own labor will be obsolete. But this will take time, and time gives us the opportunity to devise sound strategies.

At this point, preventing the kind of market collapse Citrini imagines is actually not difficult. The U.S. government's performance during the pandemic has demonstrated its proactive and decisive crisis response. If necessary, massive stimulus policies will quickly intervene. Although I am somewhat displeased by its inefficiency, that is not the focus. The focus is on safeguarding material prosperity in people's lives—a universal well-being that gives legitimacy to a nation and upholds the social contract, rather than stubbornly adhering to past accounting metrics or economic dogma.

If we can maintain sharpness and responsiveness in this slow but sure technological transformation, we will eventually emerge unscathed.

Source: Original Post Link

Have Institutions Finally 'Entered Crypto,' but Just to Vampire?

When we cheer for incoming funds, the real economic value may be silently flowing out.

A $2 Trillion Denouement: The AI-Driven Global Economic Crisis of 2028

Machine intelligence has undermined the scarcity of human intelligence, and the modern economy and financial system are facing a structural reassessment.

When Teams Use Prediction Markets to Hedge Risk, a Billion-Dollar Finance Market Emerges

Behind every football season lies a billion-dollar market.

Cryptocurrency Market Overview and Emerging Trends

Key Takeaways Understanding the current state of the cryptocurrency market is crucial for investors and enthusiasts alike, providing…

Jane Street Market Manipulation, Stripe Considering Acquiring PayPal, What's the Overseas Crypto Community Talking About Today?

WEEX × LALIGA 2026: Trade Crypto, Take Your Shot & Win Official LALIGA Prizes

Unlock shoot attempts through futures trading, spot trading, or referrals. Turn match predictions into structured rewards with BTC, USDT, position airdrops, and LALIGA merchandise on WEEX.

a16z: Why Do AI Agents Need a Stablecoin for B2B Payments?

Smart contracts will be more like corporate entities, forming long-term relationships with their vendors and partners.