I am barely breathing And I can't find the air Don't know who I'm kidding Imagining you care And I could stand here waiting A fool for another day I don't suppose it's worth the price You're worth the price, the price that I would pay RIP Duncan Sheik

TG-AI-F! I spent the day in Pittsburgh yesterday. I discovered that from data center backlash to AI doom, politicians like Pennsylvania Governor Josh Shapiro are scrambling to keep up with a debate that’s getting wilder by the day.

Let’s get one thing straight. Israelis don’t “view themselves” as victims of 10/7. They ARE the victims of the Hamas-led 10/7 massacre. Over 1,200 innocent people were murdered. Over 250 people were taken hostage, including children and Holocaust survivors. People were slaughtered, raped, abused, burned alive, their bodies desecrated – and all of this was proudly livestreamed by gleeful terrorists for the world to see. The entire nation of Israel continues to live with those scars. That’s not a matter of perspective. That’s just fact.

It's exactly three months since I launched Ground Level AI! What a ride it's already been: On June 16, I said Ground Level AI would be focused on one big question: What happens when AI meets the real world? AI is an overwhelming beat, but what matters most is still underreported amid the hype cycle. The massive infrastructure buildout, the reality for enterprise companies, the security stakes, the policy decisions, the societal consequences, the geopolitical fault lines: these are the stories that will define how this technology actually lands.  That's all happening right now, swirling amid today's headlines, and it's been exciting (and humbling) to report on it all. Over the past three months, I've covered everything from AI data centers and AI-orchestrated ransomware attacks to the AI talent wars for forward-deployed engineers; the AI coding agent hangover; AI's culture wars and the AI safety vs. cybersecurity collision. I've traveled to San Francisco to listen to experts discuss open source AI and the latest in enterprise adoption; to Las Vegas to be the first to report on the OpenAI-Hugging Face briefing at Black Hat; to Iowa to ride an AI-powered tractor; and I'm headed to Pittsburgh and then Amsterdam this week to moderate discussions about AI trust; AI and creativity; and real-world intelligence. I'm tantalizingly close to 5,000 Ground Level AI subscribers now, and the momentum seems to be there for a publication and podcast that looks beyond the model launches, product news, funding rounds, and predictions about the future. I'm most interested in digging into the messy stuff: The infrastructure AI requires, the realities of enterprise deployment, the security and governance challenges, the policy and geopolitical battles, and the workers and communities and industries being remade by it. That's where I intend to keep reporting. Want to follow where I'm headed? Meet me at

I don't live in SF. Is this peak SV culture? Anthropic, a frontier lab planning an IPO that insists its technology could potentially kill all humans, hosted a cozy conversation yesterday about sci-fi with an author who writes about AI dragons and sentient yeast. (and yes, I know Jack Clark writes sci-fi for his indie newsletter)

Who gets a seat at the table to decide if frontier AI is safe? Dario Amodei wants independent evaluators to help “pace the frontier.” But the divide between AI safety and cybersecurity complicates the question: Who should do the evaluating?

If you’re looking for prophetic warnings or promises about AI over the past few years, you don’t have to look far. AI companies and their leaders have long been fluent in a language that can sound downright biblical. Lately, I’m exhausted by all of it. The AI industry seems to exist in a permanent state of prophecy where either abundance is coming or catastrophe is looming, or both at the same time.

no, the world is not waking up to "rogue" AI. I have been asked about OpenAI-Hugging Face by a hairstylist, a cab driver and an old friend. they didn't understand anything about it and calling it "rogue AI" as if humans had nothing to do with it is unhelpful.