Claude and Drupal QA
I had Claude compare 50 pages across two versions of a Drupal site, at desktop and phone widths, and it finished in about ten minutes. The results were impressive but not quite perfect.
Andrew Maddox
AI news, analysis and practice from inside higher ed. Daily AI briefings as well as some longer form essays and experiments.
I had Claude compare 50 pages across two versions of a Drupal site, at desktop and phone widths, and it finished in about ten minutes. The results were impressive but not quite perfect.
What if the U.S. House had 1,000 seats instead of 435? I built a simulator to find out. Small states lose some of their edge in presidential elections, but no winner from 2004 to 2024 would change, and the implications for gerrymandering were a bit different than I imagined at first.
The Mavs and Stars aren't just leaving downtown. They're following a center of population that's been walking north since 1970, and it started with how Dallas treated its southern half.
Voters didn’t just hand Democrats Congress and countless state positions—they backed data center limits in more than 30 communities, changing the calculus on future AI infrastructure.
AI-assisted bug hunting pushed Drupal to 256 security advisories in five months, nearly 4x all of 2025.
We ran out of milk, so I had Claude, Anthropic's AI assistant, build our Walmart cart from our order history. Here's how it went, the steps to try it yourself, and a prompt you can use.
Twenty years ago, Cameron Moll made the case for realigning a site instead of redesigning it. I finally took that advice because of Claude Code.
I used AI (Claude Opus 5.5) to build a custom WordPress theme in a weekend — about 10 minutes of hands-on time versus the 20 hours it used to take.
John Gruber muses on Meta's new AI strategy: There are four big players in consumer AI: Google, Meta, OpenAI, and Anthropic. I don’t think Zuckerberg mentioned any of the others…
Mark Zuckerberg’s newest manifesto casts him as a defender of the individual against concentrated technological power. He argues that “superintelligence” should be widely distributed, enabling people to create, compete, and…
Anthropic is officially a public company. Its Nasdaq debut — an $81 billion offering that opened well above its $83 IPO price before cooling off — is a landmark test of whether public markets will pay up for costly AI companies. OpenAI, meanwhile, is working to win back investor confidence ahead of its own possible IPO. Elsewhere, Google's AI search features are reshaping where ads land and how much advertisers spend, and Meta's Muse agent faces privacy questions after reportedly exposing information across users' accounts.
Anthropic priced its IPO at $83 a share, valuing the Claude maker at about $1.5 trillion. Demand blew past the shares available. Now the public-benefit mission, the safety commitments and the push to expand computing capacity all get a lot more scrutiny, with employees and advocates pressing for guardrails. OpenAI’s Sam Altman and Greg Brockman make a similar case: keep building, but add stronger safeguards and monitoring.
Google’s turning Gemini into an assistant that actually does things: calling businesses, booking appointments, managing calendars and making purchases, with confirmations and spending limits meant to keep you in control. OpenAI’s new GPT-6.2 models chase the same kind of multistep work. And GPT-6 Cyber helped uncover a critical Java flaw that Oracle has now patched. Agents that can act bring real productivity gains. They raise the security stakes, too.
Nvidia’s hunting for a permanent CEO after Jensen Huang’s reported departure, and it’s pulling back from startup financing at the same time. That puts the AI boom under a microscope: Nvidia’s earnings and Anthropic’s IPO this week will test investor nerve, and infrastructure startups could feel the funding squeeze first. Also today: tighter U.S. controls on overseas access to advanced chips, California’s lawsuit against OpenAI and data centers’ growing appetite for power. Plus, SpaceX caught Starship, a milestone for recovering the spacecraft for potential reuse.
Nvidia's China business is under a brighter spotlight. The question: did its partner network help Chinese firms get chips, possibly through diversion, and what did company leaders know? Treasury Secretary Scott Bessent wants the board to act and answer, while new reporting points to gaps in chip controls and oversight. Elsewhere, a four-hour FAA ground stop and a Microsoft outage that hit more than 70 countries showed just how fragile critical systems can be.
What a week. Jensen Huang resigned as Nvidia’s CEO amid a criminal probe, jolting the company at the center of the AI-chip boom. CFO Colette Kress is now interim chief while investigators look at whether Nvidia chips were diverted to China (the allegations around Huang and Megaspeed remain unproven). Meanwhile, Anthropic set its IPO range at $70 to $78 a share — up to a $1.3 trillion valuation — even as investors question AI spending and returns. It also shipped the faster Claude Fable 5.2, while keeping Mythos locked down. Markets bounced back, but financing and data-center costs aren’t going anywhere.