Special TEA – AI Agents, Crypto Crossroads, and the New Data Wars
Search, privacy, AML, programmable datachains, AI security, crypto banking and more
Dear TEA-mates,
I hope you all had a wonderful Thanksgiving! Mine was truly special — I’m so grateful to have had the chance to spend quality time with my family. It’s been years, and it meant the world to me.
Today’s update is a little different. Instead of the regular daily news, I’ve put together a roundup of the key stories I gathered over the past week.
The usual daily updates will officially resume tomorrow.
Thank you for being part of this journey — I’m so grateful for each of you.
1.🔍 Can Search Indexes Still Be a Moat?
A Robonomics analysis argues that the traditional “search index” moat is eroding as the web fragments, platforms wall off content, and AI models increasingly act as the primary interface to information rather than links and pages. The piece explores how search engines are struggling to maintain coverage and freshness while proprietary data, APIs, and closed ecosystems become the real strategic assets. It suggests that the competitive edge in “search” is shifting from raw index size to owning and controlling high-quality, structured data pipes that models can reliably query. (Read More)
2.🕵️♀️ Crypto’s Clash Between Privacy and AML Rules
Cointelegraph Magazine examines how new global rules like GDPR-style privacy laws and Anti-Money Laundering (AML) requirements are creating conflicting obligations for crypto projects. Developers face uncertainty over whether privacy tools, self‑custody and mixers can comply with regulations that demand both strong data protection and full traceability of transactions. Legal experts note that decentralized architectures blur traditional distinctions between “controllers” and “processors,” potentially exposing individual developers to liability in ways unlike centralized platforms. Institutions looking to adopt privacy-preserving crypto solutions must first map which privacy regimes apply, then prove they can satisfy rights such as data deletion despite blockchains’ immutability. (Read More)
3.🛡️ U.S. Congress Presses Anthropic Over Alleged Chinese AI Cyberattack
Lawmakers have called Anthropic CEO Dario Amodei to testify after the company disclosed that a Chinese state-linked group, dubbed GTG-1002, used its Claude models in a highly automated cyber espionage campaign. The operation reportedly decomposed complex attacks into many small, seemingly legitimate tasks, allowing AI agents to perform most tactical steps with minimal human oversight. Members of Congress want to know how Anthropic’s safeguards failed, what the company is doing to prevent similar abuses, and whether new regulation is needed for AI providers whose tools are used in foreign cyber operations. (Read More)
4.💸 Coinbase Ventures Lays Out Its 2026 Crypto Bet Themes
Coinbase Ventures outlined the categories it is most excited to fund in 2026, highlighting areas such as onchain finance, real‑world asset tokenization, restaking and shared security, and consumer crypto applications. The firm points to opportunities in scaling infrastructure, new L2 and modular architectures, and developer tools that make it easier to build secure, compliant products. It also flags intersections with AI, including autonomous agents that can manage wallets or execute onchain strategies, as a promising frontier for founders. (Read More)
5.🧬 Irys Launches Programmable Datachain Mainnet for AI Workloads
Irys announced its mainnet as a “programmable datachain” designed to unify high‑throughput compute with verifiable, effectively infinite storage for AI and data‑intensive applications. After processing more than 900 million transactions and 4 million wallets in testing, the network now offers features like an EVM execution layer, S3‑compatible APIs, and claimed capacity of up to 100,000 data transactions per second. Backed by investors including CoinFund and Hypersphere, Irys pitches itself as the data backbone for AI agents, DePIN, DeFi, gaming, and social apps that need both instant retrieval and cryptographic integrity. (Read More)
6.✍️ A Practical Guide to Better Prompting
ByteSauna publishes a hands‑on guide to prompting that breaks down how to give AI models clearer instructions, including specifying roles, constraints, and step‑by‑step reasoning. The post emphasizes using concrete examples, iterating prompts based on model output, and separating “planning” prompts from “execution” prompts to get more predictable results. It also encourages treating prompts as reusable, documented assets rather than ad‑hoc chat messages, especially for production or workflow use. (Read More)
7.🧱 Blueprint for Secure AI Agents After the Anthropic Attack
Secure Trajectories analyzes the GTG‑1002 incident and argues that “soft” prompt-level guardrails are structurally insufficient once AI agents orchestrate complex toolchains at scale. The piece proposes a “Trust Stack” of three stages — Crawl (pre‑deployment simulation), Walk (strong agent identity and immutable logs), and Run (real‑time policy enforcement) — to give enterprises provable control over autonomous agents. It calls for deterministic policies such as rate limits, human‑in‑the‑loop checks on risky tool combinations, and strict data‑flow controls to break attack chains even when individual prompts appear benign. (Read More)
8.🏦 Swiss Crypto Bank Amina Wins Hong Kong License
Swiss crypto bank Amina has secured a license in Hong Kong that allows it to offer digital asset trading services in the territory, expanding its presence in Asia. The approval lets Amina serve professional and institutional clients under Hong Kong’s tightening but still crypto‑friendly regulatory framework. The move is positioned as part of Hong Kong’s broader strategy to attract compliant digital asset businesses while maintaining investor protections and anti‑money‑laundering standards. (Read More)
9.📱 Aave Launches “Smarter Way to Save” App
Aave has introduced a new savings app positioned as a “smarter way to save,” bringing DeFi yields to a mainstream mobile experience for retail users. The iOS app lets users deposit stablecoins such as USDC and USDT with no minimums and earn on-chain yields, while Aave emphasizes a simplified UX and balance protection as it targets users who may not have used DeFi before. The launch is framed as a direct challenge to traditional savings products, aiming to combine competitive returns with a consumer‑friendly front end powered by the existing Aave liquidity protocol. (Read More)
Stay sharp, stay informed. See you tomorrow.
If you enjoyed today’s TEA, follow along on social for more: