Denoise · Twitter

The conversation is maturing from individual exploit discoveries to systematic frameworks published by actors like @AnthropicAI and @GoogleDeepMind.

The focus is on formalizing agent security, with major labs publishing red team frameworks and responsible disclosures for patched jailbreaks.

• 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.

A new developer workflow is consolidating around the terminal, with @karpathy providing the narrative, @AnthropicAI the product, and @swyx the early benchmarks.

Anthropic's release of Claude Code 1.5 defines the day, marking a significant push toward terminal-native coding agents as the new developer interface.

• 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 visible as @OpenAI, @LangChainAI, @vercel, and @replit all ship components for a standardized agent deployment and execution stack.

Major platforms are shipping SDKs and deployment harnesses focused on agent orchestration, tool calling, and multi-worker 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 agents dominate the conversation, @MistralAI continues its strategy of releasing foundational, open data assets to the community.

The category is quiet except for a significant open dataset release for web OCR from MistralAI.

• Mistral AI@MistralAIrising

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

The primitives are advancing, with @GregKamradt's framework and @mem0ai's specialized memory layers suggesting vector retrieval alone is no longer sufficient.

The discussion is evolving from simple RAG towards 'context engineering' and more complex, layered memory architectures for 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.

While AI agents dominate headlines, established SaaS tools like @NotionHQ and @linear are shipping deterministic automation, representing a parallel path to productivity.

Workspace automation sees quiet but practical releases, with Notion and Linear both shipping new auto-triage and workflow features.

• 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 field is moving from anecdotal tricks (@dotey) to industrial-scale benchmarking (@weights_bienses) to find the efficient frontier of prompt design.

Prompt engineering is professionalizing through large-scale, systematic benchmarking and the distillation of production-tested patterns.

• 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 highlights a key bottleneck for training next-generation agents: ensuring the quality and safety of large-scale synthetic datasets.

A niche but critical discussion emerges on dataset curation for agent training, specifically on filtering out synthetic data that harms generalization.

• Jerry Liu@jerryjliu0repeated

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