Denoise · Twitter

The discourse is maturing from basic prompt injection to sophisticated, orchestration-level attacks, a pattern visible in frameworks from GoogleDeepMind and analysis by @MalwareTechBlog.

Attention is squarely on the security of autonomous agents, with major labs releasing red-teaming frameworks and disclosures on patched vulnerabilities.

• 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 direct competition is solidifying between Anthropic's Claude Code and OpenAI's models, with early benchmarks by @swyx suggesting meaningful performance differences.

Anthropic's launch of Claude Code 1.5, a terminal-native agent, dominates the conversation, sparking debate on the future of 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 convergence toward a standard agent stack is underway, with OpenAI's SDK, LangChain's protocol support, and deployment targets from Vercel and Replit all pointing in the same direction.

The entire stack for deploying and orchestrating agents is being built out in real-time, with new SDKs, hosting platforms, and protocol integrations.

• 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 continues its strategy of releasing high-quality, open data artifacts, differentiating itself from competitors who rely on proprietary training sets.

A single major open dataset release for web OCR from Mistral AI marks the key event in this category.

• Mistral AI@MistralAIrising

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

A fault line is visible between scaling raw context (per @reach_vb) and developing structured memory architectures (per @GregKamradt, @mem0ai) to manage complexity.

The discussion has moved beyond simply increasing context window size to engineering more sophisticated memory and retrieval systems.

• 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.

A convergence pattern is clear as non-AI-native platforms like Notion and Linear independently develop similar agentic features, commoditizing basic workflow automation.

Workspace productivity tools like Notion and Linear are shipping similar AI-powered automation and auto-triage 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.

A methodological split is emerging between sharing anecdotal 'tricks' (@dotey) and building systems for exhaustive testing (@weights_biases), with the latter gaining momentum.

The focus is on moving prompt engineering from an art to a science through large-scale benchmarking and systematic optimization.

• 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.

The tweet from @jerryjliu0 reveals that data curation, specifically filtering out harmful synthetic data, is a primary bottleneck for improving agent generalization.

The conversation centers on the critical but difficult process of curating high-quality datasets for training capable agents.

• Jerry Liu@jerryjliu0repeated

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