» Tag
llm
473 postsMCP Security: Trust Boundaries and the Attacks the Spec Won't Stop
A breakdown of MCP's four trust boundaries, stdio server risks, OAuth 2.1/PKCE/CIMD updates, and the confused-deputy attack class engineers must handle.
TutorMoments: Testing If AI Tutors Know When to Help or Hold Back
Allen AI's TutorMoments benchmark tests whether LLM tutors know when to scaffold and when to push students toward harder reasoning.
Persistent State Machines Recast LLM Attention as Hardware FSMs
A PSM framework recasts LLM attention as deterministic hardware state machines, synthesized in Vivado with an estimated 0.0267 pJ/op dynamic energy.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comWhy AI Models Perform Worse Outside English
Why do LLMs underperform outside English? Training data, tokenizer inefficiency, and instruction-tuning gaps explain the cost and accuracy divide.
Turo: Aggressive Token-Saving Proxy for CLI AI Agents
Turo is an open-source CLI proxy that trims AI agent prompts, cutting LLM token usage by up to 70% while preserving meaning.
Code Mode Cuts Agent Costs by 99.2%, Measured in Production
A production benchmark shows code-execution-based agent workflows cut token costs by 99.2% versus raw sequential MCP tool calls.
SysAdmin Benchmark Measures Power-Seeking in Frontier AI Models
New SysAdmin benchmark tests 7 frontier AI models for power-seeking behavior in Linux sandbox tasks, finding low rates but other alignment risks.
Self-State Attacks: New Threat Poisons AI Agents via Their Own Memory
New research defines 'self-state attacks'—AI agents compromised via their own memory files—and finds OS-level defenses hit a structural limit.
data2prompt: CLI Tool Fits Data-Heavy Projects into LLM Context Windows
data2prompt CLI profiles and samples CSVs, Excel files, and SQLite databases to fit data-heavy projects into any LLM context window.
Clustering Two Dell Pro Max GB10 Systems for Local AI Compute
Two Dell Pro Max GB10 systems clustered via RDMA/RoCE for 256GB of local AI memory, revealing setup quirks and firmware pitfalls.