All of decentralized finance is unsafe amid advances in AI, the founder of blockchain security company OpenZeppelin says. "I now consider all of DeFi unsafe," Manuel Aráoz said on May 27 on X. "Coding agents are superhuman at finding vulnerabilities, and smart contract security is too asymmetric: defenders need to fix every bug while attackers need just one exploit to steal funds." Aráoz, who served as OpenZeppelin’s technology chief from 2015 to 2019, said in a follow-up post that he has warned his friends and family to exit all DeFi applications. He added that even blue chips such as Aave, MakerDAO and Compound are not safe. Don't Miss: Find out if your retirement plan is exposed to risks most investors overlook — get matched with a fiduciary adviser today. Access private-market funds that may include exposure to companies like SpaceX— available through SoFi Invest with entry points starting at just $10 Aráoz’s remarks come amid concerns about how AI models such as Anthropic’s Mythos could impact cybersecurity. The company said recently that the model has found more than 10,000 "high- or critical-severity vulnerabilities across the most systemically important software in the world." Meanwhile, the DeFi sector has recently experienced a high frequency of attacks, with April ranking as the worst month on record . Several DeFi stakeholders, including Aave founder Stani Kulechov and Phoenix Labs CEO Sam Macpherson , pushed back on Aráoz’s remarks by attributing recent hacks to operational security failures rather than to smart contract code vulnerabilities. OpenZeppelin also released a series of posts on X disavowing Aráoz’s remarks and echoing the view that most DeFi attacks are the result of OPSEC failures. Trending: Real estate. Crypto. Private deals. Most retirement accounts don't allow them — self-directed IRAs do . "AI is a real threat vector, but it is also one of the most powerful defensive tools we have, if used with rigor and expert human judgment," the company said . "Our researchers use AI daily to catch more issues and edge cases. The answer to AI risk is not retreat from DeFi. It is better security." However, not all agree that the rebuttals offer succor to DeFi users. Pablo Sabbattela, Web3 OPSEC researcher and founder of Security Alliance, said most projects are bad at OPSEC. He also pointed to permissionlessness as another concern. "The problem with DeFi is that ... Full story available on Benzinga.com