Sharon Goldman

OpenAI gives first detailed debrief of the Hugging Face incident at Black Hat conference I attended a session where researchers said the company is "consciously slowing down research to enhance security" while a full technical postmortem is still underway. www.groundlevel-ai.c…

OpenAI gives first detailed debrief of the Hugging Face incident at Black Hat conference groundlevel-ai.com

My final story for Fortune as a staffer is the last of four deep-dives I reported on the ground about communities grappling with the AI data center boom: Two mayors, one $10 billion AI data center, and a growing divide in small-town Texas fortune.com/2026/06/16/a...

Two mayors, one $10 billion AI data center, and a growing divide in small-town Texas | Fortune fortune.com

Are the kids going to be all right? A weekend in the redwoods with my 22-year-old nephew made me wonder what the AI era looks like to the generation inheriting it. open.substack.com/pub/sharongo...

Are the kids going to be all right? groundlevel-ai.com

AI companies and their leaders have long been fluent in a language that can sound downright biblical. But lately, I’m exhausted by the AI industry's seemingly-permanent state of prophecy. www.groundlevel-ai.com/p/i-am-exhau...

I am exhausted by the AI industry’s permanent state of prophecy groundlevel-ai.com

This weekend I revisited something I wrote back in my @VentureBeat days, right after OpenAI’s GPT-4 was released. It's amazing how much the AI industry's conversation about openness has changed in just three years... open.substack.com/pub/sharongo...

The fight over open models has entered a new phase groundlevel-ai.com

The OpenAI Hugging Face breach changes a core assumption: AI models must be treated as potential adversaries “It’s such a visceral example of how different the world is now,” he said. “There was no attacker. The agent has become a malicious insider.” www.groundlevel-ai.com/p/t…

The OpenAI Hugging Face breach changes a core assumption: AI models must be treated as potential adversaries groundlevel-ai.com

Question for the AI policy and infrastructure crowd: Is the debate over U.S. AI competitiveness too focused on chips and models, rather than the politics of building AI infrastructure? That's the argument I explore in my latest essay for Ground Level AI: www.groundlevel-ai.com…

America's AI race is being fought on two fronts groundlevel-ai.com

As Meta's AI data center campus grows from $27 billion to $50 billion, I checked back with the residents and entrepreneurs I met in February. Some are thriving. Others are struggling more than ever. www.groundlevel-ai.com/p/exclusive-...

EXCLUSIVE: Meta's mega Louisiana data center is getting even bigger. Six months after my visit, I checked in with local residents groundlevel-ai.com

The people building AI and using it every day are experiencing a different reality from the people who resist it or simply don't use it. That growing disconnect, I believe, is helping fuel a backlash that isn't going away. open.substack.com/pub/sharongo...

The mutual blindness behind the AI backlash groundlevel-ai.com

AI needs a 'Stargate for Data'? A new viral essay from OpenAI researcher Will DePue argues AI's next bottleneck is data. Proposed fixes include buying companies purely for their data, and getting firms to stop deleting theirs. Here's what made my eyes bug out: www.groundlevel-…

AI needs a 'Stargate for Data,' a new viral essay argues. Here's what made my eyes bug out groundlevel-ai.com

The US wants Anthropic to secure its most powerful AI models. But one researcher says that is mathematically impossible open.substack.com/pub/sharongo...

The US wants Anthropic to secure its most powerful AI models. But one researcher says that is mathematically impossible groundlevel-ai.com

1/ One interesting ripple from today's news that Citadel bought Leopold Aschenbrenner's Situational Awareness stock portfolio: I'm told the fund attracted significant participation from people across the AI industry, including a number of Anthropic employees, often through SPVs.

🎉Ground Level AI turns one month old today! 🥳 if you've subscribed, read a story, shared one, had me on a podcast or just encouraged me along the way -- thank you!! 🙏 www groundlevel-ai.com 🎊

At final Black Hat keynote (called a locknote, ha ha) panelists say they are surprised at how the OpenAI - Hugging Face incident debrief, as well as other reporting on AI agent escapees (Anthropic/Meta) - has turned into a marketing/PR play. "Felony humble-bragging," said one.