ret2happy

Security Researcher | Top 20 Chrome VRP Researcher (2022/2024)

[ret2happy @ret2happy]

100/127 CVEs are reported by Google.

Google’s internal tools kills the game

https://t.co/bORleSNL10

[ret2happy @ret2happy]

whale is back 🐋

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at https://t.co/GCdiMzk1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: https://t.co/drlDrxkYtp
🤗 Open Weights: https://t.co/T13Y8i7SDM

1/n

[ret2happy @ret2happy]

@eternalsakura13 @xint_official Agree, all the skills/scaffold tricks will finally be trained in the next-generation models.

[ret2happy @ret2happy]

Glad to see my paper accepted by FSE 2026!
We proposed this agentic vulnerability detection idea early in the DeepSeek R1 era and have since enhanced its capabilities with advanced strategies. Our work has safeguarded millions of dollars in the smart contract ecosystem.

[ret2happy @ret2happy]

@Fried_rice @koushik77 I finally figured out where my duplicates came from :(

[ret2happy @ret2happy]

@zh1x1an1221 oh my duplicate 😇

[ret2happy @ret2happy]

@OpenAI Short answer: LLMs cannot replace senior security researchers. They still can't generate end-to-end exploits for complex real-world on-chain incidents.

[ret2happy @ret2happy]

𝗖𝗮𝗻 𝗔𝗜 𝗿𝗲𝗽𝗹𝗮𝗰𝗲 𝘀𝗲𝗻𝗶𝗼𝗿 𝗮𝘂𝗱𝗶𝘁𝗼𝗿𝘀/𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀?

We re-evaluated @OpenAI's EVMBench with real-world on-chain incidents.

Paper: https://t.co/Nj07dZ1EQM

[BlockSec @BlockSecTeam]

Are AI agents ready for detecting and exploiting smart contract vulnerabilities?

We re-evaluated @OpenAI's EVMbench with a contamination-free dataset of real-world hacks.

Our data shows different results. 🧵

Paper: https://t.co/zDQ2aRdCXt

[ret2happy @ret2happy]

Just got two CVEs, together with my Agent 'Synapse'

> Agent hunts Agent

[ret2happy @ret2happy]

@thedawgyg It took more than 4 weeks to finally receive my 2 renderer reward today. I suspect the VRP processing may be backlogged.

[Claude @claudeai]

Introducing Claude Code Security, now in limited research preview.

It scans codebases for vulnerabilities and suggests targeted software patches for human review, allowing teams to find and fix issues that traditional tools often miss.

Learn more: https://t.co/n4SZ9EIklG

[ret2happy @ret2happy]

@zh1x1an1221 Valuable!

[ret2happy @ret2happy]

@eternalsakura13 1w😂

[ret2happy @ret2happy]

@xvonfers welcome back!

[ret2happy @ret2happy]

@Zyy0530 Curious about which LLM does it use? Top tier models e.g. gpt-5.2/claude-4.5-opus or the open source model

[ret2happy @ret2happy]

Incredible. OpenAI handles 800M users requests with only a single primary Postgres and 50 relicas.

[Bohan Zhang @BohanZhangOT]

@PostgreSQL has long powered core @OpenAI products like ChatGPT and the API. Over the past year, our production load grew 10× and keeps rising. Today we run a single primary with nearly 50 read replicas in production, delivering low double-digit millisecond p99 client-side latency and five-nines availability. In our latest OpenAI Engineering blog, we unpack the optimizations we made to to scale @Azure PostgreSQL to millions of queries per second for more than 800M ChatGPT users. Check out the full post here: https://t.co/VTnxhlwlat

[ret2happy @ret2happy]

Agentic ProbLLMs: Exploiting AI Computer-Use and Coding Agents

https://t.co/Ke1lCZMEii

[ret2happy @ret2happy]

Given the recent Balancer/Yearn exploits, imagine what on-chain miracles we’d see if Lazarus started developing an LLM for automated attacks. In addition, they should cite the A1 preprint by @lzhou1110, as it is the first work for on-chain AEG.

[Anthropic @AnthropicAI]

New on our Frontier Red Team blog: We tested whether AIs can exploit blockchain smart contracts.

In simulated testing, AI agents found $4.6M in exploits.

The research (with @MATSprogram and the Anthropic Fellows program) also developed a new benchmark: https://t.co/QpGPMqlDRG

[ret2happy @ret2happy]

Having witnessed your dedication over these years, I hope the future brings you even greater opportunities!

[Péter Szilágyi @peter_szilagyi]

Since y'all spammed my timeline full of #Ethereum existential crises, here's a letter I sent to EF leadership in a year and half ago 😬.

(link in next post because Twitter...)

[ret2happy @ret2happy]

@ant4g0nist training for fuzzing? 👀