This isn’t dot-com 2.0, there’s no “dark compute.” Every chip, every GPU, every watt of energy is being used the moment it comes online. In AI, demand isn’t lagging infrastructure… infrastructure is racing to keep up with demand.

It continues…ever since OpenClaw release, Anthropic has been on a war path in response. Scheduled tasks, remote access, now this… Anthropic can now control your mouse, keyboard, and screen. Combine that with Claude Code, MCP connectors, remote execution, etc They’re

Just connected Claude Code and my OpenClaw to Google Stitch via API so I can prompt them to create projects and screens. Also set it up to be triggered as part of an automated workflow (not really doable with Loveable). And the export from Stitch to Figma is flawless. Solid

Thank goodness, some completely wasteful compute capacity has been recovered.

If your plan is to wait until AI gets better, consider this: Give two kids a bike.  One already knows how to ride. The other still needs training wheels. Now give both kids a faster, lighter, more advanced bike. Who benefits more? The one who already knows how to ride takes

Those panicking about AI compute constraints are missing half the equation. 💡 Jense laid out the math last week. Then Google Research just changed the denominator completely. We are witnessing an unprecedented explosion in computing demand. The shift from basic generative AI

Just me, or is Claude slower, dumber, and buggier since they started the whole “extra usage” campaign?

Most important skill of the next decades will be agent literacy. Someone asked Chamath what he'd tell his kids to focus on in a world of AI. His answer: philosophy, psychology, history, physics, and English writing. Not coding bootcamps. Not "learn to prompt." Just

SaaS stacks are bloated, and CFOs are starting to take notice. Teams are paying for platforms they use at 5%, while lean, AI-powered workflows can be built faster and tailored to exactly what they need. The future of work is smarter, hybrid stacks built for your actual business

My wife had an idea last week, and wanted to share because I think it will be useful for other couples and families. "Can we get a new Bible verse and short devotional sent to both of us every morning?" Took me about 6 minutes to build and add some content for our life/rhythms

My son is claw pilled

My kids are growing up in a way I never could have planned. Different cultures, different perspectives, and a childhood I couldn't have drawn up for them on paper. It traces back to one decision… choosing a less certain path when everyone around me was doing the “next right

AI and blockchain are not rivals. They're infrastructure for each other. But here's the biggest what most people are missing… Every TAM model ever built assumes the end users are humans. Which implies human speed, hours, and transaction volume. When billions of AI agents

Trying to develop this attribute more every day.

Wow!!! Breaking news everybody 🚨

Me & my co-founder Nate de la Montaña on every call: We're piping intake forms through Snarfbyte for intelligent triage, which kicks over to Blorpstack to auto-generate project scaffolds and screen flows. From there Quibnose handles the data mapping and API stitching between the

Excuse me, what?…there must be a typo here

Every company's AI journey starts the same way… The CEO tells his directors they need to be using AI. They get ChatGPT or Claude subscriptions. People start playing around with it… using it for emails and maybe summarizing some documents. Maybe they make some funny office

AI tools have never been easier to use. ChatGPT, Claude, Copilot… your team can get real value from these today without writing a single line of code. But here's what I keep seeing: A CEO uses AI to draft emails faster. A sales rep uses it to write follow-ups. Someone in ops