Macro News

1. S&P 500 futures were on a steady decline all week, closing in on 5,800, which hasn’t traded at those levels since early November. The post-election rally is starting to fully dissipate.

2. Key events this upcoming week that could impact central bank monetary policy:

   1. December PPI Inflation data - Tuesday, 1/14/25

      • Year-over-year inflation: 3.3%, down from 3.4% in November.

      • Month-over-month inflation: 0.2%, down from 0.4% in November.

      • Easier monetary conditions are typically supportive so we expect risk-on sentiment to get a bit of a boost and short-term rally coupled with President-Elect Donald Trump’s inauguration next week.

   2. December CPI Inflation data - Wednesday, 1/15/25

      • Year-over-year inflation: 2.9%, up from 2.7% in November.

      • Core CPI: 3.2%, down slightly from 3.3% in November. (Core CPI excludes volatile food and energy prices)

      • The mixed data here may lead the Fed to maintain a more cautious approach. If further declines in core occur, the Fed may justify a more patient or neutral stance, but they will most likely wait-and-see if overall costs continue to increase. We’ll expect to see more volatility occurring across crypto markets in the short-term but a bullish case still persists for the longer-term investment landscape.

   3. December Retail Sales data - Thursday, 1/16/25

   4. Philadelphia Fed Manufacturing Index - Thursday, 1/16/25

   5. December Housing Starts data - Friday, 1/17/25

3. The catastrophe in California combined with the state’s insurance policy failures is simply a tragedy. Thoughts and prayers to the victims of the wildfires. The state is burning and the destruction thus far is monumental. Regulators force insurers to sell way below cost. Translation: California forces the biggest gap between rates and risk in the nation. A recent Kobeissi Letter post highlights the estimated damages from the fires have tripled to $150bn and estimated losses are up to $20bn for insurers.

4. 30-year treasury yields hit 5% following strong employment data. With a strong labor market, there is less chance of rate cuts and tighter liquidity for risk assets.

Industry Headlines

1. Bitcoin spiked close to $96,000 a week ago, remained down >5% over the past week as it retreated down to $90,000, before rebounding past $100,000 with the latest macro headlines. Thus far, the Fed has highlighted a more neutral environment in this week’s data releases, which has historically fared well for crypto. It didn’t stay above $100,000 for long as there continues to be significant resistance around that point.

2. VanEck files for an OnChain Economy ETF (the “Fund”) according to SEC filings. The ETF is designed to invest primarily in companies that support the development and adoption of blockchains and crypto - this could be anything from hardware to infrastructure to payment applications but the primary investment strategy will be equities, not tokens.

3. Usual, a decentralized stablecoin issuer, announced updates to its USD0++ protocol by introducing dual exit mechanisms. The protocol update led to a significant depeg of USD0++ to $0.92. For several weeks, the Usual team claimed that USD0++ was redeemable for 1:1 USD0. With the updates, the 1:1 redeem function halts without any prior announcements to trap users farming so that the issuer can keep their TVL - pretty much trying to save their token price at the expense of users.

4. For the first time, over 20% of all spot crypto volumes occur on DEX exchanges, indicating a significant shift towards decentralized trading platforms. This trend reflects growing trust and interest in DeFi.

5. Aiccelerate DAO was launched this past weekend with the promise of accelerating innovation at the intersection of crypto and AI. Aiccelerate introduced its native token, AICC, which was meant to incentivize participation and facilitate governance within the DAO, however, faced significant backlash when it was revealed the entire supply was concentrated in the wallets of early investors/insiders. Poor tokenomics and grift accusations aside, I thought it was incredibly bizarre that a group aiming to “accelerate crypto and AI” would not have a single traditional AI subject matter expert be included as an advisor. At a time when the intersection between the two industries is heavily lopsided and agents in crypto predominately use a web2 stack, having advisors with the proper technical experience is paramount.

   Subscribe

GitHub Spotlight: Eliza

Carl and I spend a significant amount of time on GitHub: we’re either perusing repositories as part of investment diligence for a project, using it as a tool to source founders and deals, or just trying to learn more about how a particular project or idea is being implemented. Marketing paints grand visions, but the real story of any project lives in its code. As former engineers turned investors, we've developed a keen eye for separating substance from spectacular claims. Each week, we'll decode a notable project's repository, translating technical realities into clear insights for both technical and non-technical readers alike. No marketing fluff, no buzzword bingo – just honest analysis of what projects can and cannot do based on their actual implementation.

Consider this your technical bullsh*t detector. We thought it would be fun to peek behind the curtain of hyped projects, examine the code that powers their promises, and share an unvarnished perspective on what we find. We’re here to help you see what we're seeing – for better or worse.

Earlier this week, we dove head first into Eliza. Eliza is an open-source agent development framework that promises to make building AI agents simple, powerful, and extensible. But does it really live up to the hype? In our first GitHub Spotlight feature, we break down what Eliza is great at, where it falls short, and what you need to know if you're considering using an agent built using the Eliza framework.

You can read our full post here.

The bottom line on Eliza is that it’s a great tool for creating consistent, cross-platform personas, but falls short of being a true autonomous agent framework in its current state. Eliza has a market cap just shy of $2bn so the market has shown its hand and highly values cross-platform personas and social agents. That said, Eliza in its current state should not be mistaken for a comprehensive agent framework like its web2 counterparts and only scratches the surface of unlocking genuine autonomous agent development.

The X post generated a lot of buzz in the community.

What we’re reading.

From the labs.

Technical research.

1. AI Creation and the Cosmic Host by Nick Bostrom

2. Towards Secure and Private AI: A Framework for Decentralized Inference by Nesa Research

3. Self-Supervised Inference of Agents in Trustless Environments by FortyTwo Network

4. SciAgents: Automating Scientific Discovery Through Multi-Agent Intelligent Graph Reasoning by Alireza Ghafarollahi and Markus Buehler

From the community.

Threads and blog posts.

1. ai16z, ELIZA, and the Bazaar of Agents by Chain of Thought

2. Stop Making Me Bridge by Cumberland Labs

3. Everything is fine with DeFAI by 0xJeff

4. DeFAI, Chain Abstraction, and Multi-Agent Interactions by Rishabh Khurana

   DeFAI, short for decentralized finance + artificial intelligence, is the merger of two rather complementary areas. DeFAI protocols harness the capabilities of AI to enhance the user experience, automation, and decision-making in DeFi. This could cover self-governing portfolio management solutions, personalized user interfaces, or fraud detection solutions. I think this X article does a really good job of surfacing a more pragmatic perspective on the state of DeFAI, where the majority of DeFAI solutions are underwhelming in the present form. One of the barriers to a more impactful landscape of DeFAI offerings is bridging and recommends having a separate chain abstraction agent for unifying balances across networks to complement the agent actually executing actions on-chain.

From the founders.

Product.

1. HyperLiquid 2024 in Review by Hyper Foundation

2. Hyperstructures by Jacob

   A must-read crypto classic. Dusted this one off the archives because the core principles of an autonomous agent have similar properties to hyperstructures in crypto. Jacob defines hyperstructures as crypto protocols that can run for free and forever, without maintenance, interruption, or intermediaries. At Reforge, we define autonomous agents as AI that are self-governing systems that can perform tasks, make decisions, and adapt their actions in pursuit of specific goals with minimal or no human intervention. Today, the self-governing property only exists in a distant future but eventually they will evolve into sovereignty where they have full agency over their internal state and execution. In essence, we believe sovereign agents fit squarely into the vision of hyperstructures. They share the qualities of being unstoppable, credibly neutral, open, and designed for positive-sum outcomes. But not only do sovereign agents embody many hyperstructure traits, but they also inject new dimensions: self-governance, dynamic adaptation, and agency. I’m excited to see the shift from purely infrastructure to agentic protocols. More thoughts to come on this later!

3. Solana MEV Report by Helius

4. The Promise and Challenges of Crypto + AI Applications by Vitalik Buterin

   Another dated piece (from Jan 2024) where I spent some time reviewing some of Vitalik’s ideas and thoughts surrounding the intersection of crypto and AI. It’s been nearly a year since he wrote this and it was an insightful exercise reflecting on all the crypto + AI developments that have happened since this was written.

5. An Open Letter to David Sacks by Dr. Jasper Zhang

6. Ethena 2025: Convergence by Guy