Paul Hsu, an ex Zynga exec now running crypto fund @Decasonic, expects gaming mergers like Microsoft-Activision to result in talent leaving big corps for startups. More on this & his Chicago-based $48.8M fund raised from Pritzker Org, Drive Capital, others

After going through 50+ projects in Cookie DeFAI Hackathon, here's what I learned (This is more like feedback / how I see the agent market rn / how to stand out) 1. DeFAI = Abstraction Layers for a lot of devs Many teams put an NLP interface in front of their product (possibly because they see DeFAI as synonymous with @HeyAnonai, @griffaindotcom, @orbitcryptoai, @askthehive_ai). In most cases, this doesn’t make sense—especially if you're only able to do a few things like using Cookie API to find the top 5 AI agent tokens by mindshare, find top coins with momentum, etc. This is just a mini feature, one that a few top abstraction layers already have. Imo, it’s better to just go to Cookie Dashboard directly for these analytics instead of adding another interface—it’s not comprehensive enough. DeFAI ≠ Abstraction Layers Instead of just replicating what already exists, teams should focus on unlocking new possibilities with Cookie API—powering entirely new use cases and verticals rather than taking inspiration from existing ones. ————————————————— 2. The Birth of Vertical Agents I'm quite surprised by the number of interesting ideas that came out of this Hackathon—there are a handful of projects with unique concepts. Many are still in early demo / ideation stages but paint an exciting picture of future use cases. • Agent that helps preserve your legacy—checks in if you're okay, and if you die, takes action to fulfill your wishes. • ETF / Index fund that utilizes Cookie analytics for investment decisions and comprehensive research reports. • Agent Security Analysis with Agent Security Scoring. • ChatGPT-like product / Learning Hub for developers to learn everything about Solana. • DYOR layer that tracks analysts/KOL calls, DYOR, and copy trades. • Framework for agents to make covenants, enabling complex agent-to-agent or agent-to-human interactions (undercollateralized loans, employment agreements, alliances/coordination). • Personality + DeFAI Agent—AI Companion that trades and dynamically adjusts its behavior/visuals based on the market. More teams are launching niche agents rather than just "trading agents" or AI-driven dashboards/research agents. Launching a vertical agent makes it easier to differentiate from generic use-case agents. There are already major players in trading agents. Even though the sector is nascent, it’s still hard to stand out—especially at an earlier stage. Better to focus on vertical agents. Many may think @HeyTracyAI, the flagship agent for @virtuals_io on Solana, isn't useful or can't make you money. The reality is, an agent built like an actual business—solving real problems—will perform better long-term. The sports market is a huge TAM. Look beyond Web3, guys. (Not endorsing Tracy, just making a point on Vertical Agents.) Bottom line? Niche, Vertical agents solve real problems and create unique use cases, while generic ones struggle to stand out. ————————————————— 3. Lack of Proper Research Agents While vertical agents are carving out unique niches, another major gap in the space is proper research agents. The keyword here is "Proper." There’s no research agent that can replace human information synthesis & reasoning yet. This applies not just to Cookie DeFAI Hackathon projects but Web3 AI agents in general. Most AI agents today aggregate data but don’t synthesize insights the way humans do. Analyzing data across traditional dashboards like @cookiedotfun, @GoatIndexAI, @Decentralisedco, and using Grok is still better than agents spoon-feeding Web3 AI agent "insights." There’s a clear gap here, despite many abstraction layers & teams focusing on enhancing research capabilities. Whoever cracks this first will have a serious edge. ————————————————— 4. Cookie DeFAI Hackathon Projects Most of the hackathon projects are still in early-stage development, and many haven’t deployed yet. Since this is a pure DeFAI hackathon (and as you’ve seen, DeFAI is the best-performing category in AI agents), there will be many high-quality projects/tokens coming out of this event. As mentioned in point 2, many will offer unique use cases beyond the current DeFAI applications we know. As AI agents continue to grow as a sector, there will be more gaps for agents to fill—e.g., B2A (Business-to-Agent) beyond B2B & B2C. The next wave of DeFAI projects won’t just enhance existing use cases—they’ll create entirely new ones. ————————————————— 5. Cookie as the Data Backbone & Distribution Channel for Agents Instead of relying on launchpads to highlight unique agent tokens, Cookie empowers agents & teams with a data backbone that tracks on & off-chain AI agent data—enabling new, interesting use cases. At the same time, Cookie’s dashboard is used by 240k+ MAU who are deep in the space. Finding gems on the Cookie dashboard & Cookie Hackathon is akin to uncovering a new gem on Virtuals. Cookie has proven itself as a strong distribution channel for agents. The more agents leverage this, the faster the ecosystem matures. ————————————————— Bottom Line This post is a teaser—more content covering the winners and my top picks from the Cookie DeFAI Hackathon coming soon. This hackathon brings similar vibes to the Solana AI Hackathon but arguably better—because it’s pure DeFAI. DeFAI isn’t just another AI trend—it’s shaping up to be crypto’s most promising agent vertical. And this hackathon proves it. (I'm biased towards DeFAI and believe it’s a crypto-native agent use case that can grow as a standalone vertical—competing with Web2 agents.) Anyhow, stay tuned for more interesting content on the hackathon!

MARC ANDREESSEN: ELON SOLVES PROBLEMS FASTER THAN OTHER COMPANIES ORGANIZE MEETINGS “Elon has developed an operating method that is very unusual by modern standards. The top line thing is this incredible devotion from the leader of the company to fully, deeply understand what the company does, to be completely knowledgeable about every aspect of it, and to be in the trenches and talking directly to the people who do the work, deeply understanding the issues and being the lead problem-solver in the organization. Basically, Elon shows up every week at each of his companies. He identifies the biggest problem that the company's having that week, and he fixes it. He does that every week for 52 weeks in a row, so each of his companies has solved the 52 biggest problems in that year. Most other large companies are still having the planning meeting for the pre-planning meeting for the board meeting, for the presentation, with the compliance review and the legal review. So it's this level of both an incredible intellectual capability coupled with an incredible force of personality, a moral authority, an execution capability, a focus on fundamentals that is just really amazing to watch.” Interview with Chris Williamson, December 14, 2024, @pmarca

💙 SEC began a “new journey toward regulatory clarity” this week. The SEC Crypto Task Force wants to “travel to a destination where people have great freedom to experiment and build interesting things”. However, SEC will not “tolerate liars, cheaters, and scammers.” SEC “is working hard to process applications for exemptive relief, requests for no-action letters, and registration statements”. 💙