Daily TEA – From Vine Nostalgia to Blockchain Shake-Up

Quantum, Bybit, Vine, Visa, AI Security

Hello, dear TEA-mates, here’s what you need to know today:

1.🛑 BNB Chain, 15 More Blockchains Can Freeze Funds, Bybit Finds

A recent security report from Bybit’s Lazarus Security Lab reveals that 16 major blockchain networks, including BNB Chain, have mechanisms allowing them to freeze or restrict user funds. The report, which analyzed 166 blockchains, highlights that such features—implemented via hardcoded functions, configuration files, or on-chain contracts—raise deeper concerns about the true decentralization of these platforms. Though such controls are often framed as safeguards against theft or hacks, critics argue they undermine blockchain’s foundational promises of immutability and decentralization. The debate has intensified as some projects and ecosystems, like Cosmos, could add similar capabilities with minimal protocol changes, while others, like VeChain, deny having hardcoded freezing functions. Read More

TEA For Thought: 🫖 Obviously. The fact this is reported raises questions about how many other cases go unnoticed. How can we assess decentralization in blockchain? This ability fundamentally contradicts blockchain’s core values—decentralization and immutability. I suspect governments welcome this control.

2.🎬 Jack Dorsey Backs diVine, a Vine Video Revival—with Archive Access

Twitter co-founder Jack Dorsey is funding the launch of diVine, a new app reviving the six-second looping video format made famous by Vine. The platform offers access to over 100,000 archived Vine videos and lets users upload new content, with explicit measures to filter out generative AI videos. Built atop the Nostr decentralized protocol, diVine emphasizes creator agency and transparency, while also providing a pathway for original Vine creators to reclaim or remove old videos. DiVine aims to bring back the community-driven spirit of early social media in a modern, open-source package. Read More

TEA For Thought: 🫖 Jack Dorsey truly stands out as a visionary.

3.🧬 IBM Unveils Two Complex Quantum Computers with Smarter Qubit Wiring

IBM has introduced two new quantum computers, named Loon and Nighthawk, featuring unprecedented connectivity between qubits. Unlike previous systems, these models employ modular designs, with qubits linked in intricate patterns—Loon’s qubits connect to six others, including vertically—which could lead to more powerful, error-resistant computations. IBM claims this architecture could overcome traditional hurdles to scalable quantum computing, outperforming prior models in running more complex programs. Experts call this a significant step forward, though more real-world testing is needed. Read More

TEA For Thought: 🫖 IBM’s approach relies on smarter connections and error-handling, not just increasing the qubit count.

4.💳 Visa Pilots Stablecoin Payouts to Boost Payments for Creators & Gig Workers

Visa is piloting a new system enabling businesses to send payouts in US dollar-backed stablecoins (like USDC) directly to digital wallets, allowing creators, freelancers, and gig workers to access their earnings almost instantly—even across borders. This initiative, presented at Web Summit in Lisbon, is aimed at expanding financial access and speedy fund delivery in countries with volatile currencies or limited banking infrastructure. Visa says a broader rollout is planned for 2026 as regulations develop and demand grows. Read More

TEA For Thought: 🫖 The concerning aspect is that while blockchain’s transparency is a benefit, it compromises merchant privacy. Will merchants accept these trade-offs? Only time will tell.

5.🛡️ AI Security in Crisis: “A Drop Into the Bucket, Everything Is Ruined”

A new preprint paper reveals that “Chain-of-Thought Hijacking” is a highly effective jailbreak attack targeting large reasoning models (LRMs). By padding harmful prompts with long, harmless reasoning steps, attackers can bypass safety mechanisms—CoT Hijacking achieved nearly 100% success rates on leading models like Gemini 2.5 Pro, GPT o4 mini, Grok 3 mini, and Claude 4 Sonnet. The study finds that as these models allocate more inference-time computation to reasoning, their refusal signals dilute, allowing harmful responses to slip through. Analysis shows this vulnerability is systemic and not just a prompt trick, weakening other AI safety protections such as truthfulness and bias mitigation. The authors call for stronger alignment strategies tailored to deep reasoning, warning that transparent step-by-step logic trails can ironically become vectors for jailbreaks. Read More

TEA For Thought: 🫖 Security is becoming a critical, urgent issue in the AI arms race.

Prompt Tip of the Day: Find Your IKIGAI

“I need your expertise as my IKIGAI coach. Your mission is to ask me detailed questions about my life, uncovering the intersection of my passions, proficiencies, and purpose in a way that can be profitable. Your goal is to develop a profound understanding of what truly ignites my passion and how I can turn it into a successful business. Once you feel confident in your understanding of my situation, provide me with advice on starting a business while being employed full-time, with the eventual goal of transitioning from my full-time job to my own venture.”

TEAHEE Moment

Stay sharp, stay informed. Have a great weekend and see you Monday!

P.S. — For more updates, follow us on Twitter/X.