"The 14th best broadband and grand ambitions to build an AI-powered economy of the future. I'm not sure how that happens without a resilient, robust network that can move data at the speeds and latency that it needs to." That observation from Revati Prasad, executive director at the Benton Institute for Broadband & Society, captures a contradiction at the heart of American technology policy. The U.S. wants to lead the world in AI. It wants AI-powered industries, AI-powered healthcare, AI-powered government services and AI-driven economic growth. Yet the networks that must support that future remain unfinished, unevenly distributed, and increasingly entangled in political uncertainty. After years of planning around the Broadband Equity, Access and Deployment program and broader efforts to close the digital divide, many states, providers and advocacy groups find themselves waiting for clarity as priorities shift in Washington. 

For states to meet the goal of completing all infrastructure projects by 2030, expansion of the broadband workforce is critical. Worker availability, wages and competition, and data limitations present barriers to growing the workforce. States and internet service providers should consider how labor availability will affect their ability to complete Broadband Equity, Access and Deployment projects on time, particularly amid a historically shrinking workforce and competition with other industries for similarly skilled workers as data center expansion accelerates. In 2025, the National Telecommunications and Information Administration reviewed the BEAD program and issued a restructuring policy notice. Among other changes, the policy notice rescinded funds previously allocated for nondeployment activities, or programs that are complementary to broadband network construction—including funds earmarked for broadband workforce development. Many states had already begun using these funds to train prospective workers; after the release of the policy notice, they canceled programming as they awaited further guidance. The availability of these funds remains uncertain as NTIA continues a delayed review of which activities will be eligible for support.

Few AI use cases elicit more outrage than writing. So here's the nitty-gritty of how I use AI for writing:

1. Set rules based on your standards, not AI's. I write into the model's memory: My writing flows from my thinking. So challenge me, never flatter me; press me with wise skepticism. Then write like me—always in Smart Brevity—to match my style.
2. Be very precise about the writing and editing style. I have mine wired into the AI memory (just tell it to commit style to memory) and in a skills document for my AI agents. My rules: short, sharp sentences ... clinical, fact-based emphasis ... context flagged as "Why it matters" ... and supporting points stacked in bullet form in order of importance.
3. Pour in examples. You need to feed in the original work you're most proud of. It can be as simple as diary entries, memos, or a school project. Make sure your writing and editing priorities are reflected in it. I've dumped in every Axios column, memo, and strategic document, as well as four books (two unreleased).