Kiwi News: handpicked crypto news for builders

look the foundational problem with ethereum protocol design is that most of the people designing the protocol don't actually use ethereum themselves. you can't fix the right things if you don't use your own fucking product. you don't have enough skin in the game if you don't personally depend on the system you're designing & stand to lose money when it breaks. i understand, everything looks great from an ivory tower lol. things look very different when you're actually on the ground dealing with the consequences. if you don't use ethereum yourself but still think you're qualified to design its protocol, you're playing god with a system you don't have to live with and whose consequences you don't have to suffer.

This is insane. We cannot allow AI agents to go into your bank account and act independently without your authorization. But that is what’s happening right now. Left unchecked, AI could endanger the entire global financial system — including your bank account. Pause AI NOW!

🧵Yesterday may prove to be a turning point for security. AI is making vulnerability research dramatically faster. This favors attackers: they can find vulns and exploits at machine speed, while organizations must still inventory systems, test updates, and deploy patches without breaking production. PHP2Shell shows what that looks like. On August 11, researchers reported critical flaws in PHP runtime. They had existed for more than a decade and affects versions from PHP 5.4. An unauthenticated attacker could potentially execute code remotely, read process memory containing passwords, tokens, and API keys...🫠 Both vulnerabilities are rated CVSS 9.2. PHP powers roughly 70% of websites, totalling millions of servers. The researchers disclosed the bugs responsibly. Fixing them took about 70 days. The fix touchs 300+ files and changed nearly 12,000 lines. These were not simple mistakes. The disclosure then broke down. Amazon saw an apparent exploitation attempt on October 17, four days before the public release. This indicates a first leak. On October 20, the CVE identifiers leaked. The next morning, a supplier with advance access published patched container images more than six hours early, giving attackers a chance to study the fix while most defenders were still waiting. The official releases arrived at 14:00 UTC. Exploitation followed immediately. Cloudflare reported thousands of attempts in the ten hours before the announcement, tens of thousands in the first 45 minutes, hundreds of thousands within two hours, and millions later that day. Organizations that could not patch immediately had incomplete options: WAF rules, isolation, or taking services offline. Government portals, universities, and hospitals across Europe did exactly that. Large teams with accurate inventories and automated deployment could respond in hours. Many others still do not know where PHP is running: an old site, a vendor appliance, a forgotten container, or an unsupported internal application. AI did not make software insecure. It always was insecure. AI just changed the economics of finding and exploiting bugs and compressed the gap between disclosure and mass exploitation from days to minutes. Patching faster will not be enough. Critical systems need continuous inventory, automatic updates, isolation, memory-safe languages where possible. Our last bastion is software whose security can be demonstrated rather than assumed: formal verification, cryptography, memory-safe architectures, and secure hardware. The era in which complexity provided accidental protection is ending. PHP2Shell is not an isolated crisis. It is a preview.

clanker will be maintained by the @masknetwork / Firefly team moving forward they have exciting plans and strong partners lined up to support clanker’s next chapter for things going forward, please reach out to @suji_yan thank you all for your continued support clank clank

SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios. The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios. Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said. Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior. Read my latest for Axios here: https://www.axios.com/2026/10/22/ai-agen…

congratulations to boredape twelvefold beeple refik !!! what a very legitimate and important incident for art and culture that has occurred !!!

V12 found 25 vulnerabilities in the x402 protocol. Critical findings included: - Requests approved without any payment - Maliciously high requests being paid by default - Session takeover from relayed login signatures Here's how. 🧵