Denoise · Twitter

The conversation is shifting from single-model jailbreaks (@AnthropicAI) to vulnerabilities in agent orchestration and cross-tool exploits (@GoogleDeepMind, @MalwareTechBlog).

The focus is on red-teaming autonomous agents, with major labs releasing frameworks and researchers testing them against live systems.

• Anthropic@AnthropicAIrising

  Responsible disclosure on a Claude jailbreak chain we patched last week. Full write-up including our red team timeline.

• Google DeepMind@GoogleDeepMindrising

  New red team framework for prompt injection in autonomous agents. Covers cross-tool leakage, scanner evasion, and sandbox escape patterns.

• MalwareTech@MalwareTechBlogrepeated

  Autonomous agent running pentest flows against a real SaaS. First real-world run: fewer false positives than I expected on the vulnerability surface.

Anthropic's Claude Code release and @karpathy's commentary signal a convergence on the terminal as the primary AI developer interface, moving beyond IDE plugins.

Major releases indicate a significant shift towards terminal-native coding agents, with benchmarks and early user testimonials showing momentum.

• Anthropic@AnthropicAIrising

  Claude Code 1.5 is live. Terminal-native coding agent with full Claude Opus reasoning, file-ops sandbox, and session replay.

• Andrej Karpathy@karpathyrising

  The developer-experience shift from IDE to terminal agent is underrated. Coding workflows are about to look nothing like 2024.

• swyx@swyxrising

  Codex vs Claude Code terminal agent benchmarks. Pass@1 diverges more than I expected on the long-context editor tasks.

• DSPy@dspy_airising

  DSPy 3.0: prompt optimization via compile-time search over system prompt variations. Benchmarks inside.

• @levelsio@levelsiorising

  Switched my whole editor setup to Claude Code this week. Shipping faster than when I used Cursor + Copilot.

OpenAI's new agent SDK and LangChain's integration thread (@LangChainAI) point to a convergence on standardized agent communication and deployment protocols.

Infrastructure for deploying and orchestrating agents is a central theme, with SDKs and runtimes released by OpenAI, Vercel, and Replit.

• OpenAI@OpenAIrising

  New agent SDK: protocol-level tool calling, deployment harness, and multi-worker orchestration primitives. Docs live.

• LangChain@LangChainAIrising

  MCP protocol integration thread. How to wire existing LangGraph agents into the Anthropic Model Context Protocol server spec.

• Vercel@vercelrising

  Edge runtime for agent workers is live. Spawn durable background agents from any serverless deployment.

• Alex Albert@AlexAlbert__rising

  When your security scanner finds nothing scary on an agent deploy, check the orchestration layer again. That's usually where the jailbreak sneaks through.

• Replit@replitrising

  New agent deployment harness. One command to go from local orchestration to hosted agent worker.

MistralAI's release of a massive OCR dataset indicates continued foundational investment in vision capabilities, even as text agents dominate headlines.

The day's activity is marked by a single but significant release of a large-scale dataset for training web OCR models.

• Mistral AI@MistralAIrising

  Open dataset release: 100M-row web OCR dataset. Cleaned, licensed, ready to train.

Actors like @GregKamradt and @reach_vb are fragmenting the simple RAG pattern into specialized techniques for caching, memory hierarchies, and knowledge graph traversal.

The discourse is advancing from basic RAG to 'context engineering,' exploring failure modes in large contexts and more complex memory structures.

• Vaibhav Srivastav@reach_vbrising

  Tested the new 10M context memory window end to end. Surprising failure modes around rag retrieval cache invalidation, thread below.

• Greg Kamradt@GregKamradtrising

  RAG is dead, long live context engineering. My framework for when to cache, when to retrieve, and when to just dump memory into the prompt.

• mem0@mem0airising

  Memory layer for agents: differentiating working memory from the subconscious store. Vector index isn't enough anymore.

• LlamaIndex@llamaindexrepeated

  Knowledge graph retrieval walkthrough: when semantic vector search misses, graph hop beats it every time.

The autonomous workflow pattern is being normalized in mainstream tools like Notion and Linear, mirroring the capabilities of dedicated AI agents.

General-purpose SaaS tools are integrating agent-like automation features for workspace management and productivity.

• Notion@NotionHQrising

  Notion workspace automation is out of beta. Auto-fill tables, chained updates across databases, and a new audit log surface.

• Linear@linearrising

  Linear now auto-triages incoming issues. Quiet launch, but already our favorite workspace feature of the year.

• Temporal@temporaliorepeated

  Orchestrating agents with durable workflows: replayable, resumable, and multi-worker by default. Walkthrough from our infra team.

• James Clear@jamesclearrepeated

  The best habit tracker is the one you actually open. Three open-source alternatives worth trying.

The approach to prompting is maturing from anecdotal hacks (@dotey) to industrial-scale optimization, evidenced by @weights_biases' large-scale study.

The focus is shifting from sharing individual prompt 'tricks' to discussing systematic, large-scale benchmarking of system prompts.

• dotey@doteyrising

  Five prompt tricks learned this week from reviewing 200 production prompts. Short thread.

• Weights & Biases@weights_biasesrising

  System prompt benchmarking at scale: we ran 40k variants across 6 frontier models. The efficient frontier is not where you think.

@jerryjliu0's thread highlights that data quality, not just quantity, is a critical bottleneck for training capable and generalizable agents.

A focused discussion on the importance of dataset curation for agent training, specifically filtering harmful synthetic data.

• Jerry Liu@jerryjliu0repeated

  Dataset curation for agent training: how we filter synthetic data that looks good but poisons generalization.