protobluf

In the future, you’ll be able to accomplish a goal by just giving Claude an outcome and a budget. That’s the direction Anthropic is building in with its new Managed Agents features, announced at this week’s Code with Claude developer event. The basic idea: Claude, wrapped in a computer in the cloud, that you can spin up, scale, and manage as needed. Anthropic is taking on the infrastructure that kills most agent products, and making sure that it scales to meet the needs of agents running 24/7. On this week’s AI & I from @every, I talk with Angela Jiang (@angjiang), head of product for the Claude platform, and Katelyn Lesse (@katelyn_lesse), head of engineering for the Claude platform, about what Anthropic is building and what it takes to make agents reliable in production. We get into: - Why the "build a generic harness, hot-swap any model behind it" playbook is already outdated. Angela points to eval data on Memory where the same task across different harnesses performed drastically differently. - The infrastructure wall every team hits in production—and why Katelyn thinks “my sandbox died and took the agent with it” is the real reason internal agents don't ship. - Why Anthropic is so bullish on using file systems and skills within Claude, including Angela's argument that those early design choices can compound for years. This is a must-watch for anyone trying to take an agent past the demo and into production. Watch below! Timestamps: How the Claude platform evolved from API to agents: 00:01:48 The primitives that make up Claude Managed Agents: 00:04:09 Why the harness and the model are becoming a single unit: 00:10:37 The infrastructure wall that kills most agent projects in production: 00:18:49 Why team agents need a different shape than individual productivity tools: 00:24:49 How Anthropic's legal team uses an agent to review marketing copy: 00:26:36 Using multi-agent orchestration for advisor strategies, adversarial pairs, and swarms: 00:34:24 How to measure agent success with outcome and budget as the end state: 00:35:50 What the platform looks like a year from now, when Claude writes its own harness: 00:39:11

It’s hard to escape the sense that the administration simply does not have a coherent strategy for Iran, and that what we are witnessing instead is a form of strategic improvisation. Washington does not want a war, yet it also does not want to appear politically weak by accepting terms associated with Iranian demands. The result is a prolonged state of limbo that deepens uncertainty in the global economy while doing little to convince Tehran to fundamentally alter its position. The constant policy shifts, contradictory public messaging, and oscillation between threats and diplomatic signaling all point to a deeper problem: the inability to build a sustainable and coherent Iran strategy. On one hand, there is fear in Washington of appearing to concede to Iran. On the other, there is growing recognition that escalation alone is unlikely to force the Iranian leadership to capitulate or accept maximalist American demands at the negotiating table. At the core of this problem lies a deeper issue — a persistent misunderstanding of how Iran’s leadership perceives pressure, deterrence, and strategic endurance. Many policymakers in Washington continue to approach Iran through frameworks developed for other geopolitical rivals, assuming that economic pressure, military threats, or diplomatic isolation will eventually produce strategic surrender. But the Iranian system has repeatedly demonstrated a high tolerance for prolonged pressure and a willingness to absorb significant costs in pursuit of regime survival and long-term strategic objectives. This disconnect helps explain why U.S. policy often appears reactive rather than strategic: alternating between coercion and de-escalation without a clearly defined end state. The administration understands the risks of military escalation, especially after years of instability across the Middle East, yet it has also boxed itself into a political environment where compromise is easily framed domestically as weakness. The consequence is a policy caught between incompatible objectives: avoiding war, avoiding concessions, maintaining deterrence, reassuring allies, and preventing regional escalation, all without a realistic framework for changing Iranian behavior. #IranWar‌

Problem, @claudeai reached it's limits while i was in middle of a feature, wanted to continue with @opencode or codex I went with opencode using codex model 😆 But it messed up, there were 20-25 files in total and it messed up so bad. It was because i was not able to share the context effectively i guess. So i made this tool, so share sessions between agents via MCP, called SESH `bun run sesh` currently supports claude and opencode, let me know if anyone wants it, i will opensource it. 100% local, no external server connection.

מסך הברזל של זיני: מאז כניסתו של ראש שב"כ לתפקיד, הדוברות נסגרה כמעט לגמרי, ועיתונאים מאוימים בתביעות דיבה. כך, לציבור הישראלי נמנעת גישה לארגון חשאי עם כוח עצום - וזה רק עניין של זמן עד שהאמון יישחק | הפרשנות של @roysharon11 #מהדורתכאןחדשות עם @TaliMoreno_

Ukraine's combat-proven Magura drone boat just got its U.S. debut — Green Berets used it to sink a ship at a live-fire exercise 100 miles from Taiwan. Read more: defence-blog.com/u-s-special-fo…