Petr Polak

It’s over. Anthropic just banned OpenClaw. Uncensored thoughts: 1. Massive mistake that will come back to bite them 2. Open source needs to win. If you have a local model running on your Mac mini, no corporation will ever be able to ban you 3. ChatGPT 5.4 is the best model. But it sucks compared to opus in OpenClaw. I will continue to pay for Anthropic api 4. I have no doubt the next OpenAI model will be optimized for Openclaw and be excellent 5. In 6 months the local models will be as good as opus 4.6 and all of this will be forgotten 6. It’s feels like from a consumer sentiment perspective things have flipped for OpenAI and Anthropic. They were the darlings when Opus 4.5 came out 7. Going to the Kanye concert right now please don’t spoil the stage or set list in the replies 8. The best openclaw set up is now Opus as the orchestrator, then much cheaper models as the execution layer. If you do this properly you won’t be paying much more than $200 a month. I’m using Gemma 4 and Qwen 3.5 for execution on my DGX Spark and Mac Studio

Starting tomorrow at 12pm PT, Claude subscriptions will no longer cover usage on third-party tools like OpenClaw. You can still use these tools with your Claude login via extra usage bundles (now available at a discount), or with a Claude API key.

We're aware people are hitting usage limits in Claude Code way faster than expected. Actively investigating, will share more when we have an update!

NASA spent $4.1 billion to launch Artemis II, the first crewed moon mission since 1972. Seven hours in, Commander Reid Wiseman called Houston because Microsoft Outlook was broken on his tablet. He’d already tried turning it off and on again. It didn’t help. “I have two Microsoft Outlooks and neither one of those are working,” Wiseman said from Orion, currently on a 10-day trip around the moon. Why two Outlooks on one device? Because Microsoft ships two different email apps and calls them both “Outlook.” This has confused office workers for years. Apparently it follows you to space. Wiseman asked Houston to remote into his tablet and fix it. So NASA did exactly what your company’s IT department does, logged into his machine, poked around, got it working. Except the user was on his way to the moon. The tablet is a Microsoft Surface Pro loaded with Microsoft 365. NASA has a $100 million contract for those licenses across the agency. The crew uses them for email, schedules, and checking mission info. Same software your office runs, same headaches your office has. The computers actually flying the spacecraft are a different animal entirely. Two custom flight computers, built to survive the kind of radiation that would fry a normal laptop in seconds, running software built by Lockheed Martin. 20,000 times faster than what the Apollo astronauts had. If one computer dies, the other takes over instantly. These control life support, steer the ship, and handle every communication back to Earth. The Surface Pro runs Outlook. One half of Orion’s computing can survive cosmic radiation thousands of miles past the moon. The other half can’t stop two email apps from crashing at the same time. I looked up the history of space software bugs and it puts the Outlook thing in perspective. In 1999, NASA lost a $328 million Mars probe because one engineering team measured thrust in pounds and another used the metric system. Nobody caught it. The spacecraft came in too low and burned up in the Martian atmosphere. In 1962, a single missing symbol in a line of code destroyed a probe called Mariner 1 just 293 seconds after launch (under five minutes). That typo cost about $190 million in today’s money. Those bugs destroyed entire missions. Wiseman’s Outlook got fixed in a few hours. So did a toilet fan that jammed two hours after launch (in zero gravity, a broken toilet is a real problem). The computers keeping four humans alive at 25,000 mph around the moon never flinched.

True story - fed the same prompt to Gemini Pro & Grok 4.20 re a Zoning Issue (commercial land in CA). Grok gave me a simple, clean answer. Gemini gave me an opposite answer with far more detail & references. I fed each answer into the other model -- BOTH held their ground. I called the City -- the City Agreed with GROK!!! Sometimes, more is NOT better. Well done.

Thank you to everyone who spent time sending us feedback and reports. We've investigated and we're sorry this has been a bad experience. Here's what we found:

Digging into reports, most of the fastest burn came down to a few token-heavy patterns. Some tips: • Sonnet 4.6 is the better default on Pro. Opus burns roughly twice as fast. Switch at session start. • Lower the effort level or turn off extended thinking when you don't need deep reasoning. Switch at session start. • Start fresh instead of resuming large sessions that have been idle ~1h • Cap your context window, long sessions cost more CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 We're rolling out more efficiency improvements, make sure you're on the latest version. If a small session is still eating a huge chunk of your limit in a way that seems unreasonable, run /feedback and we'll investigate