Yeah all LLMs suddenly become *extremely* dumb when you hit that point.
LLMS
-
Full-stack Chat AI library with React LangGraph Gemini
By
–
发现一个有趣的库，一键搭一个全栈 Chat，
前端 @reactjs 后端 @langchain LangGraph gemini-fullstack-langgraph-quickstart – 开源
直接构建 research-augmented conversational AI 亮点 • 前端基于 React，后端基于 LangGraph 和 FastAPI • 利用 Gemini 大模型自动生成搜索词，动态调用 -
Understanding Prompt Injection Attacks in LLMs
By
–
What is a Prompt Injection Attack? https://
wiz.io/academy/prompt
-injection-attack
… #zeroclick #ArtificialIntelligence #AI #security #vulnerability #CyberSec #M365 #Copilot #LLMs #LLM #GenAI #promptinjection #promptengineering #prompt -
Zero-Click AI Vulnerability Exposes Microsoft 365 Copilot Data
By
–
Zero-Click AI Vulnerability Exposes Microsoft 365 Copilot Data Without User Interaction! https://
thehackernews.com/2025/06/zero-c
lick-ai-vulnerability-exposes.html?m=1
… #zeroclick #ArtificialIntelligence #AI #security #vulnerability #CyberSec #M365 #Copilot #LLMs #LLM #GenAI #promptinjection #promptengineering #prompt @fogle_shane -
Google’s Defense Against Indirect Prompt Injection Attacks Gemini
By
–
Google's Multi-Layer Defense Against Indirect Prompt Injection Attacks on Gemini AI https://
search.app/RWTQh #GoogleGemini #Geminiai #LLMs #LLM #AI #ArtificialIntelligence #GenAI #prompt #Injection #Security #Tech #technology @SpirosMargaris @PawlowskiMario @mvollmer1 @gvalan -
Organizations Unprepared for Agentic AI Risks and Challenges
By
–
Organizations Aren’t Ready for the Risks of Agentic AI https://
search.app/SmE1o #RAG #AgenticAI #AIagentInnovation #AIAgents #LLMs #LLM #GenerativeAI #GenAI #technology #TechRevolution #tech #Engineering #ArtificialIntelligence #AutonomousIntelligence #AI @sonu_monika @enilev -
Empirical/agentic solution o3 cheats with noise suffix
By
–
I think it’s exactly that empirical/agentic solution that other models fail to see. E.g. o3 (non-pro) often cheats by appending a meaningless suffix of random noise, which is the only lever through which it brute forces one fixed, plausible count.
-
Synthesizing Generalizable Algorithms with Minimal Data and Compute
By
–
You want the ability to synthesize new algorithms (not just memorize), and you want these new algorithms to generalize strongly (not just in known situations), and you want these algorithms to be acquired using as little data and compute as possible (down to human level) This
-
LLM Analysis of FDA Approval and Portfolio Strategy
By
–
You're welcome to ask an LLM that question but when I immediately followed up with it the LLM softened its language but not its conclusion, justifying this due to contemporaneous knowledge of FDA approval, the trivial cost, and portfolio management strategies.
-
LLM Integration for Smart Email Filtering and Prioritization
By
–
Hmm I should wire up an LLM to EarthClassMail as a proof of concept here. “Computer: Is it worth interrupting Patrick’s day over this communication? If yes, summarize why in a text message and we will deliver it immediately.”