Denoise · Twitter

The conversation is shifting from theoretical exploits to documented, real-world agent jailbreaks, with Anthropic and Google DeepMind leading the public discourse.

Major labs are proactively publishing red-teaming frameworks and responsible disclosures for agent 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.

The competition between OpenAI's Codex-lineage and Anthropic's Claude is now playing out in the terminal, with early benchmarks from @swyx showing meaningful divergence.

The primary interface for AI coding assistance is shifting towards terminal-native agents, led by Anthropic's Claude Code 1.5.

• 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 de facto agent stack is forming, with protocols from OpenAI and Anthropic defining the core logic, while Vercel and Replit provide the serverless execution layer.

Infrastructure providers are racing to support the deployment and orchestration of autonomous agents with new SDKs and runtimes.

• 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 the agent conversation dominates, @MistralAI continues to focus on releasing foundational, open artifacts like datasets to enable community model development.

Mistral AI released a large-scale, open dataset for web OCR training.

• Mistral AI@MistralAIrising

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

A fault line is appearing between "more context" and "smarter context," with figures like @GregKamradt advocating for engineered memory/retrieval systems over larger windows.

The discourse on providing context to LLMs is moving beyond simple RAG towards more complex memory architectures.

• 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 agentic patterns seen in coding are also appearing in productivity tools, with @NotionHQ and @linear abstracting away manual data entry and task management.

Workspace automation is a recurring theme, with Notion and Linear launching features for auto-filling and auto-triaging.

• 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 practice is moving from artisanal "prompt tricks" (@dotey) to industrial-scale optimization, exemplified by @weights_biases's large-scale study.

The focus in prompt engineering is shifting to systematic, large-scale benchmarking to find optimal 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 note highlights a critical but often overlooked infrastructure problem: filtering out "poisonous" synthetic data that harms model generalization.

The discussion centers on the challenges of curating high-quality synthetic data for training agents.

• Jerry Liu@jerryjliu0repeated

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