Is the AI bubble real? & Apple trims engineers' AI budgets - AI News (Nov 9, 2026)

A drone flew through a military exercise and the sensors watching it simply logged it as a bird. We'll get to that in a moment. Welcome to The Automated Daily, AI News edition. The podcast created by generative AI. I'm TrendTeller, and today is November 9th, 2026. On the menu: whether the AI bubble is really a bubble, Apple tightening the purse strings on AI usage, open source pushing back against AI code laundering, a novelist with hundreds of AI-assisted books, and that remarkably convincing mechanical bird.

Is the AI bubble real?

Let's start with money. A widely shared essay argues that the AI industry may not be in a bubble, or at least not one that's about to pop. The author's case is that the huge sums going into data centers are mostly one-time capital investments, not permanent operating costs that can never be recovered. He points to fast revenue growth at Anthropic and to solid earnings and fairly ordinary valuation ratios at Alphabet, Meta and Nvidia as signs of real business value. Even without strong moats, he says, brand loyalty and institutional lock-in can keep the money flowing. He does acknowledge real risks, such as regulation, public backlash and slower growth, but concludes that much of the bubble talk is closer to wishful thinking than financial analysis.

Apple trims engineers' AI budgets

That said, costs are clearly on people's minds, even at the richest companies. Apple has reportedly lowered the daily AI spending cap for some engineers from three hundred dollars to two hundred. It's a small number for a company of Apple's size, but it shows how quickly model usage can add up when it's left uncapped across large teams. The takeaway from industry watchers is that efficiency is becoming a priority alongside raw capability: sending simpler jobs to smaller models, caching results and writing tighter prompts. The same roundup also flagged research suggesting non-English prompts can use far more energy, which is one more reason to care about efficiency.

Managing coding agents with git queues

Staying with developers, one engineer shared how he deals with the mental overload of running coding agents like Claude Code and Codex in the terminal. Too many tabs, threads and half-finished tasks were making it hard to remember what came next. His fix is to treat the work as a priority queue kept in git, with each ticket stored as a markdown file next to the relevant code. An agent works through the board, researching, building, testing and checking deployments, while anything sensitive still waits for a human to approve it. The core rule is to finish each item completely before starting the next. He admits it's still tiring, but says his productivity has improved a lot.

Engineers acting as 'meat proxies'

A sharper take on agents comes from another engineer, who argues that handing your judgment over to AI entirely should count as serious misconduct. He separates normal AI-assisted work, where you review and correct what the model produces, from what he calls 'meat proxy' behavior: passing tasks to an agent and submitting the output as your own without checking it. In his view, that pushes the risk and the cleanup onto teammates and can break production systems. His bottom line is that engineers stay responsible for their work, no matter which tool wrote the code.

Asahi Linux relicenses over AI laundering

Open source is reacting too. Asahi Linux, the project that brings Linux to Apple Silicon Macs, is moving all of its repositories from the permissive MIT license to GPL version two or later. The team says people copied unreleased code from private forks and used AI tools to hide where it came from, which they call 'code laundering.' This looks like one of the first major relicensing moves aimed specifically at AI-assisted copying. Critics worry it could alienate BSD users, and even supporters admit the GPL won't stop a determined copier. What it can do is make that kind of misuse less attractive and easier to prove.

Etsy's one-sided AI training terms

Creators are fighting similar battles elsewhere. Etsy has updated its terms in a way that seems to let the company train AI on sellers' work, while forbidding users from doing the same with Etsy's content. Many artists see that as a one-way deal on a platform that depends on their creativity. It's part of a wider backlash over tech companies claiming broad rights to user content during the AI rush.

Author publishes 370 AI-assisted books

And in publishing, South African author Coral Hart says she has released more than three hundred seventy books in about twenty months with help from AI. She uses different models for dialogue, structure and editing, but says she still designs the stories, characters and outlines herself, and a human proofreader polishes each book. She's facing heavy criticism about 'AI slop,' but her answer is that readers and sales decide what counts as quality. By her own count she has sold over a million copies, and she's now training others and building software for AI-assisted authors. Whatever your opinion, it's a clear sign of how quickly publishing is changing.

Bird-mimicking drone evades Air Force sensors

Finally, the drone from the top of the show. At an Air Force technology event at Edwards Air Force Base, a small ornithopter called the O1, built by Ornadyne, apparently showed up as a bird on other companies' detection sensors. Instead of propellers, it flaps its wings, which keeps it quiet, and it can change its color, wingspan and silhouette to look like bird species that live in the area. For anyone working on counter-drone systems, that's a real challenge: future surveillance drones may be able to hide in plain sight, blending in with ordinary wildlife.

That's it for today's episode of The Automated Daily, AI News edition. Links to all the stories are in the episode notes. Thanks for listening, and keep an eye on the birds. See you tomorrow.

Is the AI bubble real? & Apple trims engineers' AI budgets - AI News (Nov 9, 2026)

Our Sponsors

Today's AI News Topics