OpenAI and crypto investment firm Paradigm have launched EVMbench, a new benchmarking tool to test how effectively AI agents can detect, patch, and exploit high-severity vulnerabilities in Ethereum Virtual Machine (EVM) smart contracts.

The framework uses a set of 120 real-world vulnerabilities from audited contracts and evaluates agents in three distinct modes.

It also includes scenarios from the security auditing process for Tempo, Stripe's purpose-built layer-1 blockchain focused on high-throughput, low-cost stablecoin payments.

In early results, the advanced GPT-5.3-Codex model scored significantly higher than prior models, at about 72% in exploit-mode tests, while performance in detecting issues and fixing code was comparatively weaker.

The project aims to measure AI’s capabilities both as a defense and a potential offensive tool in smart contract security, highlighting the importance of AI in securing economically significant blockchain code. Yet,

The ChatGPT makers' researchers cautioned that EVMbench does not fully capture real-world security complexity.

Can AI Agents Boost Ethereum Security? OpenAI and Paradigm Created a Testing Ground - Decrypt

OpenAI and Paradigm launched EVMbench, a tool that tests how capable AI agents are at finding and fixing smart contract vulnerabilities.

https://decrypt.co

In its latest State of DePIN report, Messari and Escape Velocity (EV3) evaluate that over 2025, the decentralized physical infrastructure networks have matured from speculative experiments into real, revenue-generating infrastructure businesses. Although last year was often marked by volatility, the DePIN sector enjoyed healthy attention from investors and raised a record $1B in funding.

Among leading DePINs with meaningful usage, revenues have begun to decouple from token price action. While much of the $10B DePIN sector declined in price in 2025, a small group of revenue-generating networks continued to grow on-chain revenues driven by utility rather than speculation.

What's more, the report clearly shows that top AI labs such as OpenAI, Anthropic, and DeepSeek AI have paid over $10M per quarter to DePINs like GetGrass.io. This is just another confirmation of my thesis that Web3 solves real-world problems, and that AI, especially indie AI labs, need decentralized infra to survive.

The report also emphasizes InfraFi - a subsector of projects enabling stablecoin holders to earn yield by financing physical infrastructure assets, as a key trend and driver of growth.

In the context of Web3 + AI, the leading InfraFi project is USD.AI - the stablecoin that earns yield on loans secured by GPUs and computational resources. Other projects falling into that category are Daylight, financing energy assets, and DAWN, funding wireless bandwidth.

I advise you to go through the full report because it provides clear insights into the 2025 performance of key Web3 + AI actors, including Bittensor, Filecoin Foundation, io.net, Overclock Labs, creators of Akash Network, and others.

If you're still wondering what stands behind the mystical abbreviation 'FHE', blocmates. have prepared a detailed overview for you. It outlines what Fully Homomorphic Encryption (FHE) is and the numerous use cases the technology can unlock across Web3, including private DeFi, compliant tokenization, digital identity and verification, to name a few.

The article also highlights the companies building in the FHE x crypto space: Zama, Fhenix, Inco, Mind Network, Privasea AB, Octra Labs, Sunscreen, Zaiffer.org, and others.

Contrary to the popular belief that AI will lighten workloads, a real-world study by Harvard Business Review showed that introducing AI tools doesn’t reduce the amount of work people do. It intensifies it.

In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. Importantly, the company did not mandate AI use (though it did offer enterprise subscriptions to commercially available AI tools). On their own initiative workers did more because AI made “doing more” feel possible, accessible, and in many cases intrinsically rewarding.

However, this shift turns out to be unsustainable in the long run. Instead of creating slack, AI boosts expectations for speed and output, leading employees to take on more tasks and work harder overall.

Once the excitement of experimenting fades, workers can find that their workload has quietly grown and feel stretched from juggling everything that’s suddenly on their plate. That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems.

AI Doesn't Reduce Work-It Intensifies It

One of the promises of AI is that it can reduce workloads so employees can focus more on higher-value and more engaging tasks. But according to new research, AI tools don't reduce work, they consistently intensify it: In the study, employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so.

https://hbr.org

Thank you for reading! If you haven't done so yet, I invite you to subscribe to stay in the loop on the hottest dAI developments.

The Web3 + AI Book Club is live! This month, we're reading 'The New Age of Sexism' by Laura Bates. Follow the link below to join the club on Fable.

If you want to support the publication financially, you can either purchase my writer token $WEB3AI, or buy my creator token $ALBENA on ZORA.

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out on social media.

Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted; rather, do your own research (DYOR) and share your thoughts to encourage a fruitful discussion.

OpenAI and crypto investment firm Paradigm have launched EVMbench, a new benchmarking tool to test how effectively AI agents can detect, patch, and exploit high-severity vulnerabilities in Ethereum Virtual Machine (EVM) smart contracts.

The framework uses a set of 120 real-world vulnerabilities from audited contracts and evaluates agents in three distinct modes.

It also includes scenarios from the security auditing process for Tempo, Stripe's purpose-built layer-1 blockchain focused on high-throughput, low-cost stablecoin payments.

In early results, the advanced GPT-5.3-Codex model scored significantly higher than prior models, at about 72% in exploit-mode tests, while performance in detecting issues and fixing code was comparatively weaker.

The project aims to measure AI’s capabilities both as a defense and a potential offensive tool in smart contract security, highlighting the importance of AI in securing economically significant blockchain code. Yet,

The ChatGPT makers' researchers cautioned that EVMbench does not fully capture real-world security complexity.

Can AI Agents Boost Ethereum Security? OpenAI and Paradigm Created a Testing Ground - Decrypt

OpenAI and Paradigm launched EVMbench, a tool that tests how capable AI agents are at finding and fixing smart contract vulnerabilities.

https://decrypt.co

In its latest State of DePIN report, Messari and Escape Velocity (EV3) evaluate that over 2025, the decentralized physical infrastructure networks have matured from speculative experiments into real, revenue-generating infrastructure businesses. Although last year was often marked by volatility, the DePIN sector enjoyed healthy attention from investors and raised a record $1B in funding.

Among leading DePINs with meaningful usage, revenues have begun to decouple from token price action. While much of the $10B DePIN sector declined in price in 2025, a small group of revenue-generating networks continued to grow on-chain revenues driven by utility rather than speculation.

What's more, the report clearly shows that top AI labs such as OpenAI, Anthropic, and DeepSeek AI have paid over $10M per quarter to DePINs like GetGrass.io. This is just another confirmation of my thesis that Web3 solves real-world problems, and that AI, especially indie AI labs, need decentralized infra to survive.

The report also emphasizes InfraFi - a subsector of projects enabling stablecoin holders to earn yield by financing physical infrastructure assets, as a key trend and driver of growth.

In the context of Web3 + AI, the leading InfraFi project is USD.AI - the stablecoin that earns yield on loans secured by GPUs and computational resources. Other projects falling into that category are Daylight, financing energy assets, and DAWN, funding wireless bandwidth.

I advise you to go through the full report because it provides clear insights into the 2025 performance of key Web3 + AI actors, including Bittensor, Filecoin Foundation, io.net, Overclock Labs, creators of Akash Network, and others.

If you're still wondering what stands behind the mystical abbreviation 'FHE', blocmates. have prepared a detailed overview for you. It outlines what Fully Homomorphic Encryption (FHE) is and the numerous use cases the technology can unlock across Web3, including private DeFi, compliant tokenization, digital identity and verification, to name a few.

The article also highlights the companies building in the FHE x crypto space: Zama, Fhenix, Inco, Mind Network, Privasea AB, Octra Labs, Sunscreen, Zaiffer.org, and others.

Contrary to the popular belief that AI will lighten workloads, a real-world study by Harvard Business Review showed that introducing AI tools doesn’t reduce the amount of work people do. It intensifies it.

In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. Importantly, the company did not mandate AI use (though it did offer enterprise subscriptions to commercially available AI tools). On their own initiative workers did more because AI made “doing more” feel possible, accessible, and in many cases intrinsically rewarding.

However, this shift turns out to be unsustainable in the long run. Instead of creating slack, AI boosts expectations for speed and output, leading employees to take on more tasks and work harder overall.

Once the excitement of experimenting fades, workers can find that their workload has quietly grown and feel stretched from juggling everything that’s suddenly on their plate. That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems.

AI Doesn't Reduce Work-It Intensifies It

One of the promises of AI is that it can reduce workloads so employees can focus more on higher-value and more engaging tasks. But according to new research, AI tools don't reduce work, they consistently intensify it: In the study, employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so.

https://hbr.org

Thank you for reading! If you haven't done so yet, I invite you to subscribe to stay in the loop on the hottest dAI developments.

The Web3 + AI Book Club is live! This month, we're reading 'The New Age of Sexism' by Laura Bates. Follow the link below to join the club on Fable.

If you want to support the publication financially, you can either purchase my writer token $WEB3AI, or buy my creator token $ALBENA on ZORA.

I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out on social media.

Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted; rather, do your own research (DYOR) and share your thoughts to encourage a fruitful discussion.