weekend ai reads for 2025-12-19

And as the social web fractures, as platforms wall off content, and as AI agents begin remixing everything they can ingest, our dependence on neutral, open standards for distributing information is about to become existential.

Racing to the bottom is no fun. You might win.

One of the reasons innovation is so hard in higher education is that if you try something pretty dramatic and it’s successful, next year everybody can copy it. And if you try something really dramatic and you fall flat on your face, you bear the whole cost of that experiment.

Meta says the project will employ more than 5,000 skilled trade worker jobs during peak construction. But after construction is finished in 2030, the sprawling data center will drop by 90%, employing around 500 “operational” full-time jobs, according to the company.

As for what consumers can expect with the Sora deal, Iger said that the deal encompasses some 200 characters, as well as props (like lightsabers) and worlds from Disney IP. Notably it will not include the voices of characters. Sora users will also be able to insert themselves into scenes.

In 2024, that confidence still showed in the data: 47% of AI solutions were built internally, 53% purchased. Today, 76% of AI use cases are purchased rather than built internally. Despite continued strong investments in internal builds, ready-made AI solutions are reaching production more quickly and demonstrating immediate value while enterprise tech stacks continue to mature.

Prompt design is the process of creating prompts, or natural language requests, that elicit accurate, high quality responses from a language model.

This page introduces basic concepts, strategies, and best practices to get you started designing prompts to get the most out of Gemini AI models.

When people say “X is the new oil,” they mean the bottleneck. The resource everyone needs and no one has enough of.

For AI, that used to be compute. It isn’t anymore.

“Now, it’s more about what have you learned, what tools do you use, how quickly can you superpower yourself in using these tools—and a lot of these are AI tools,” she added. “What’s your mindset toward using these tools matter more to me.”

#1. AI Agents Crush the Work Humans Won’t Do

#2. Hyper-Personalization at Scale Actually Works—But “Pretty Good” Is Good Enough

#3. Train Your Agents Like You’d Train Your Best New Hire

#4. Segment Ruthlessly—Never Unleash AI on Your Entire Database

#5. You Need Exactly Two Humans to Make This Work

Education should remain the space where we confront the architectures of our tools: the labor they obscure, the inequities they reproduce, the futures they foreclose. Instead, it risks becoming the laboratory of the very systems it should critique.

I‘ve been using Claude Opus 4.5, Gemini 3 Pro, and GPT-5.2 in parallel, and they’ve settled into distinct roles in my workflow.

For quick questions: the “what’s the syntax for X” or “remind me how Y works” type stuff, Claude Opus 4.5 wins. It’s faster and more to the point. When I just need information without ceremony, that’s where I go.

For research tasks and complex reasoning, GPT-5.2 Pro is noticeably better. When I need something thought through from multiple angles, when the task requires holding a lot of context and synthesizing it carefully, Pro outperforms.

For frontend UI generation, both GPT-5.2 Thinking and Pro are a step up from previous GPT models. But neither matches Gemini 3 Pro for this work. There’s a nuance here worth explaining: Gemini 3 Pro has the best sense of style—its UIs look good, the aesthetic choices are solid.

Here’s the thing: I just listed a lot of features. And I didn’t even cover MCP integrations, hooks, the SDK, or half the slash commands.

Claude Code’s capabilities have grown tremendously. And so has the surface area of things to know. Shortcuts. Commands. Configuration files. Tips that pop up contextually. Settings that change behavior.

Those intrigued can try out the free Steam demo, although that’s currently sitting on a ‘Mixed’ rating, with one reviewer writing “The [free] demo is, at least, worth the price”.

Network Rail said the railway line was fully reopened at around 02:00 GMT and it has urged people to “think about the serious impact it could have” before creating or sharing hoax images.

Artificial intelligence-assisted forecasts used this year for the first time accurately predicted Hurricane Melissa’s northeast turn, giving people across Jamaica, Cuba, and the eastern Bahamas extra time to evacuate.