Denoise · Twitter

The focus is shifting from prompt injection in chatbots to complex, stateful exploits in agent orchestration layers, a concern shared by Anthropic, Google DeepMind, and security researchers.

Major labs are disclosing and addressing agent jailbreaks, while researchers are developing new frameworks for red-teaming autonomous 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.

The competition between OpenAI and Anthropic is now centered on autonomous terminal agents, with developers like @karpathy and @swyx providing early analysis on this workflow shift.

Anthropic's release of Claude Code 1.5, a terminal-native agent, has sparked intense discussion on the future of IDEs versus agent-driven developer workflows.

• 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.

A clear convergence is forming around the 'agent worker' as a new infrastructure primitive, where protocol standards from OpenAI and LangChain meet the deployment runtimes of Vercel and Replit.

Major platforms including OpenAI, Vercel, and Replit are releasing SDKs and runtimes for agent orchestration, rapidly standardizing deployment patterns.

• 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.

While agent orchestration dominates today's discourse, Mistral's foundational data release underscores the ongoing, critical need for high-quality, specialized training data in the multimodal domain.

Mistral AI released a large-scale, 100M-row web OCR dataset to facilitate the training of new multimodal models.

• Mistral AI@MistralAIrising

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

The conversation led by figures like @GregKamradt is moving from 'RAG vs. long context' to a more nuanced 'context engineering' that treats memory as a tiered system.

Developers are pushing beyond simple RAG, exploring complex memory architectures and caching strategies for long-context models and autonomous agents.

• 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 theme of automation is appearing in parallel tracks: cutting-edge AI agents and practical, rule-based automation in established SaaS tools like Notion and Linear.

Workspace automation features are shipping in tools like Notion and Linear, while Temporal highlights its use case for durable agent orchestration.

• 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.

Tooling from DSPy.ai and Weights & Biases is formalizing prompt optimization into an engineering discipline, enabling programmatic search over prompt variations.

The practice of prompt engineering is becoming more systematic, moving from anecdotal tricks to large-scale, automated 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.

As agent capabilities expand, @jerryjliu0 points out that the bottleneck is shifting back to data quality, specifically avoiding 'poisoning' from flawed synthetic data.

The discussion focuses on the critical but often overlooked challenge of curating high-quality synthetic data for training reliable agents.

• Jerry Liu@jerryjliu0repeated

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