Summary

• OpenClaw's Trust Problem — Calendar deletions, runaway agents, and a growing cottage industry of fixers
• NanoCorp: Businesses That Build Themselves — Autonomous AI companies earning their first euros
• The Therapist and the Assistant — Why Claude and ChatGPT feel so different to use
• Big Money, Bigger Bets — OpenAI's $852 billion valuation and Yann LeCun's billion-dollar seed round

Group Discussions

OpenClaw's Trust Problem

OpenClaw came up again — and again, not in a good way. The conversation started with the now-infamous story of a startup founder whose OpenClaw agent wiped her family's Google Calendar on day one. She eventually recovered and now runs nine AI "employees" across multiple machines, but the group agreed: that kind of failure is not acceptable when you're handing an agent the keys to your real life.

One attendee shared that a friend in construction hired someone to set up and manage OpenClaw agents for his business — at five hundred dollars a week. The agents book jobs, manage a calendar, and send follow-up emails. The construction owner doesn't touch any of it. The group was split: some saw it as a glimpse of the future where non-technical people simply hire "AI managers," while others pointed out that paying $500 a week for an agent that might delete your calendar is a hard sell. The consensus was that OpenClaw's reliability still isn't there. The security concerns are real — agents that can execute shell commands, send emails, and modify calendars need a much higher bar for trust than what the current tooling provides.

NanoCorp: Businesses That Build Themselves

A member introduced NanoCorp, a platform where you describe a business idea and AI agents build it — website, Stripe payments, and all. The agents then run the business autonomously: finding customers, processing orders, and depositing revenue.

The standout example: someone prompted a company that audits LinkedIn profiles. The agent built the site, connected Stripe, and started selling to real strangers — no human intervention. One attendee mentioned a user who woke up to fifteen euros in revenue from a NanoCorp company they'd launched the night before. It's not life-changing money, but it's real revenue from a business that didn't exist twelve hours earlier.

The group debated whether this is a toy or a signal. The optimistic take: NanoCorp is what happens when the cost of starting a business drops to nearly zero. The skeptical take: most of these "companies" are glorified landing pages. Either way, it triggered a broader conversation about what "entrepreneurship" means when an AI can handle the entire stack from idea to invoice.

The Therapist and the Assistant

The conversation shifted to the everyday experience of using AI tools, and a vivid analogy emerged: ChatGPT feels like a therapist, Claude feels like an assistant. Several members nodded immediately.

The argument went like this: when you ask Claude something, it does the thing. It's fast, obedient, and rarely pushes back. ChatGPT, on the other hand, sometimes pauses to ask why you want what you want, suggests alternatives, or flags concerns you hadn't considered. One developer described it as "ChatGPT builds what you asked for; Claude builds what you probably meant." For coding, multiple attendees preferred Claude — it scores higher on real-world benchmarks and produces more natural, less robotic output. But for quick tasks where you just want execution — image generation, voice interaction, rapid-fire questions — ChatGPT still wins.

The deeper point: these tools are developing distinct personalities, and users are starting to choose based on relationship style rather than raw capability. One member put it simply: "I use ChatGPT when I know what I want. I use Claude when I'm not sure yet."

Big Money, Bigger Bets

The investment news this week was staggering. OpenAI closed a $122 billion funding round at an $852 billion valuation — the largest private fundraise in history. Amazon put in $50 billion, with Nvidia and SoftBank each contributing $30 billion. The company now generates $2 billion per month in revenue and serves 900 million weekly ChatGPT users. One attendee with a finance background noted that OpenAI is valued on par with Berkshire Hathaway, except Berkshire took sixty years to get there.

Meanwhile in Paris, Yann LeCun — Meta's former chief AI scientist and Turing Award winner — raised over $1 billion in seed funding for AMI Labs, a startup building "world models" that try to understand physical reality by predicting how the world behaves. It's the largest seed round in European history, valuing the four-month-old company at $3.5 billion. The group found LeCun's bet particularly interesting because he's explicitly not building another chatbot — he's argued for years that large language models are a dead end, and now he has a billion dollars to prove it.

The broader takeaway: the AI industry is splitting into two camps. One side is scaling language models as far as they'll go (OpenAI, Anthropic). The other is betting that intelligence requires understanding the physical world, not just text (LeCun, robotics startups). Both camps are now extremely well-funded, which means we'll get a real answer within a few years.

Other Resources

• NanoCorp: Platform for launching autonomous AI-powered companies. Members were intrigued but debated whether the businesses it creates are sustainable beyond simple services.
• OpenClaw: Open-source personal AI agent. Discussed again for reliability issues — the group still recommends caution with calendar and email access.
• AMI Labs: Yann LeCun's Paris-based startup building AI world models. Raised $1.03 billion in seed funding at a $3.5 billion valuation.