Good Morning ☕️ 

While the overall crypto market has taken a breather after the past few weeks of green, there is an AI revolution happening in the “trenches.”

IYKYK…

In other news, there has been a number of releases in both the open and closed-source AI world.

In This Week’s Roundup:

• Nous Research dropped manifesto on open-source AI development

• Coinbase Ventures released a comprehensive Crypto x AI stack analysis

• Former OpenAI researcher speaks out

• DeepMind open-sourced SynthID for AI watermarking

Lets go!

The State of the Market

Total Crypto Market Cap: down 3.7% this week to $2.36 trillion
Total AI Sector Market Cap: down 8.4% this week to $26.5 billion

The headlines might scream "red" this week, but here's the thing: those surface-level numbers aren't telling the whole story. Dig a little deeper, and you'll find the “AI token” space is absolutely buzzing with action.

Over the past two weeks, we have seen a revolution in the memecoin sector. Not only are AI themed memecoins on an absolute tear, but we are now seeing AI influencers overtake human influencers in terms of mindshare.

You may be thinking “yea cool, another passing FAD” but we’re not convinced. When you have the likes of Coinbase CEO, Brian Armstrong, interacting with the AI and offering to create it a wallet, you know something is cooking.

The question that everyone’s asking is: what does this all mean and where do we go from here?

Dylan Bane from shared thoughts from inside Messari:

With GOAT kicking off this AI memecoin craze, Virtuals Protocol on Base has gained serious traction, with many dubbing it the “pump.fun” for AI agents.

Virtuals’ leading tokenized “AI agent”, LUNA, ticked a $225 million market cap in just, showing just how red hot this trend it.

Whether you love it or hate it, this AI agent phenomenon is introducing some revolutionary ideas to the space.

If you missed the first wave, it’s time to lock in and get up to speed because we believe this new wave of AI tokens is just getting started, and we might be witnessing the birth of something much bigger than just another memecoin FAD.

Highlights of the Week

Nous Research — From Black Box to Glass House

Nous Research published an article challenging closed-source AI development paradigm and regulatory approaches, advocating for transparent development practices.

Key Points:

• While closed-source undoubtable has its place and often pushes the boundaries of innovation, people do not often afford open-source development the same consideration.

• Open-source being considered dangerous is far from the truth, and likely stems from the inherent bias of people who profit from closed-source.

• Nous is committed to pushing forward the development and awareness of open source AI, acting as a “counterbalance” to the unchecked power of closed-source AI and it’s potential for mass surveillance of the population.

• Discusses the issues with security when it comes to closed-source development security, citing breaches at multiple centralized AI companies like OpenAI and Anthropic, and using the example of Linux to prove that open-source is superior in finding and fixing vulnerabilities.

Why It Matters: This piece is a great read for those wondering why open sourcing this powerful technology is so important for the good of humanity. We are fully aligned with what Nous is trying to achieve in terms of democratizing access to this frontier technology so that everyone has access on an open and fair playing field, while also acknowledging the role that closed-source companies play in pushing the space forward.

Former OpenAI Researcher Exposes Copyright Violations

Former OpenAI researcher, Suchir Balaji, alleges the company's training data usage violates copyright law and harms content creators, stating models make complete copies and compete with original content.

Key Points:

• Suchir Balaji, who worked on GPT-4 training data, claims OpenAI's practices violate copyright law and directly compete with content creators.

• Claims company's "fair use" defense for internet data usage is invalid.

• Senior researcher Miles Brundage leaves OpenAI citing his desire to be free to publish independent research freely.

• Meanwhile, News Corp filed a lawsuit against Perplexity AI for alleged content theft, while praising OpenAI's approach to content agreements.

Why It Matters: This controversy highlights the growing tension between centralized AI development and content rights, reinforcing the need for transparent, ethical practices in data collection for AI training. The huge turn over in OpenAI staff, especially recently is also interesting and suggests that the company is facing some issues internally.

Coinbase Ventures — Demystifying the Crypto x AI Stack

Coinbase Ventures releases comprehensive report on the Crypto x AI convergence, outlining investment thesis and market opportunities across the emerging tech stack.

Key Points:

• Believes crypto is not relevant for all layers of the AI stack, but can be used to bring transparency, censorship-resistance and native payment rails to AI.

• Predicts rise of "Agentic Web" where AI agents operate on crypto infrastructure to drive economic activity.

• Identifies four key layers: Compute (GPU networks), Data (decentralized access/verification), Middleware (model deployment), and Applications.

• Crypto will become preferred payment rail for agent-to-agent and agent-to-human transactions.

• Forecasts AI will generate majority of smart contracts, leading to explosion of onchain applications.

Why It Matters: Coinbase ventures sharing this analysis is part of a growing signal that sentiment is shifting in how major investors view the intersection of Crypto and AI. With Coinbase Ventures actively investing across all stack layers, the report suggests we're entering an era where AI agents become primary drivers of economic activity. Their emphasis on open, decentralized systems for AI development aligns with growing industry focus on transparent, accessible AI infrastructure.

Here’s what else happened this week ↘️