Good morning and thank you for joining us again!

Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated, and published with the intention of making AI news and technology more accessible to everyone.

THE FRONT PAGE

Amazon's AI Anime Experiment Lasted Less Than a Week

TLDR: Amazon quietly released AI-generated English dubs for several popular anime over Thanksgiving weekend, pulled them within days after fierce backlash from fans and voice actors, and may have done so without approval from Japanese rights holders.

The Story:

Amazon's AI dubbing pilot program, announced back in March for a handful of obscure licensed films, expanded over the holiday weekend to include anime titles like Banana Fish, No Game No Life Zero, and Vinland Saga—labeled as "AI beta" in Prime Video's language settings. The quality was immediately panned: monotone delivery, odd pacing, zero emotional range in scenes that demanded it. Clips went viral and voice actors mobilized. Daman Mills, who voiced Kaworu Nagisa in Amazon's own Evangelion dubs, called the move "AI generated garbage" and said he'd refuse future Amazon work if the decision stood. The National Association of Voice Actors branded it "AI slop." Within days, Amazon removed the English AI dubs from most titles, though some Spanish versions remain. And here's where it gets messier: Kadokawa told Anime News Network it hadn't approved an AI dub "in any form," while HIDIVE—which already had a human-voiced dub for No Game No Life Zero—said it was "not aware in advance" and is now investigating.

Its Significance:

The fan backlash will fade. The rights holder situation won't. Japanese studios are notoriously protective of their IP and slow to trust new partners—moving forward without clear authorization could set back Amazon's anime ambitions for years. The timing also stings: Crunchyroll's CEO recently confirmed the platform won't use AI for voice acting, giving fans a clear alternative. And the economics make Amazon's choice baffling—Mills pointed out on X that dubbing pays around $125-150 per hour under SAG-AFTRA rates, barely a rounding error for a company Amazon's size. Amazon hasn't commented publicly, leaving open whether this was a pilot gone wrong, or a preview of where streaming is headed.

QUICK TAKES

Micron Kills Crucial Consumer Brand After 29 Years to Focus on AI

The story: Micron announced it will stop selling its popular Crucial brand of RAM and SSDs to consumers by February 2026. The company is redirecting all manufacturing capacity toward enterprise AI customers, including high-bandwidth memory for data centers. Crucial has been a go-to choice for PC builders since 1996.

Your takeaway: AI's hunger for specialized memory chips is reshaping the entire hardware market. Fewer choices and higher prices may be coming for everyday consumers as manufacturers prioritize data center demand.

Microsoft Slashes AI Sales Growth Targets as Customers Resist Unproven Agents

The story: Microsoft reportedly cut sales growth targets for its AI agent products, with some quotas slashed by up to 50% earlier this year. The company denies lowering aggregate quotas, but enterprise customers are struggling to adopt tools from Microsoft Foundry (formerly Azure AI Studio) for building autonomous AI agents. Private equity firm Carlyle cut spending on Copilot Studio after the software couldn't reliably pull data from other applications.

Your takeaway: Despite billions in AI infrastructure spending, companies are finding that autonomous AI agents aren't ready for the complex, multi-step tasks they're being sold on—and security concerns are adding to the hesitation.

1,000+ Amazon Employees Sign Open Letter Warning AI Will "Do Staggering Damage"

The story: Over 1,000 Amazon employees signed an open letter claiming the company is abandoning climate goals to build AI, using the technology to cut jobs, and helping create "a more militarized surveillance state." The letter came as Amazon plans to spend nearly $150 billion on data centers over 15 years and announced cuts of up to 30,000 corporate jobs as part of AI-driven restructuring.

Your takeaway: Internal resistance to AI is growing at major tech companies, with employees pushing back against the pace of rollouts and demanding more say in how the technology is used.

Security Researcher Finds 100,000+ Confidential Legal Files Exposed in AI Tool

The story: A security researcher reverse-engineered Filevine, a billion-dollar legal AI platform, and discovered he could access over 100,000 confidential files from a law firm with zero authentication required. The vulnerability exposed documents protected by HIPAA and court orders, internal memos, payroll data, and client files. Filevine fixed the issue after responsible disclosure.

Your takeaway: As law firms rush to adopt AI tools, they're handing over their most sensitive data to platforms that may not have basic security in place—a reminder that the AI gold rush is outpacing security basics.

Financial Times Invests in AI Platform, Licenses 13 Years of Archives for AI Agents

The story: FT Ventures invested in RavenPack and signed a licensing deal that gives the AI analytics company access to the Financial Times' full news archive dating back to 2012. Hedge funds and banks can now build AI agents that analyze FT journalism for sentiment, context, and market-moving narratives. The FT called it their first distribution partnership built specifically for generative AI.

Your takeaway: Major publishers are finding new revenue streams by licensing their archives for AI training, turning decades of journalism into fuel for financial AI agents.

TOOLS ON OUR RADAR

TRENDING

New Video AI Model Handles Both Creation and Editing – Chinese company Kling AI released Video O1, which can generate videos from text and edit existing footage in a single tool. Users can swap subjects, change weather, or adjust styles without manual editing work.

Marine Robotics Firm Resumes Search for Missing MH370 Plane – Texas-based Ocean Infinity will restart its deep-sea hunt for the Malaysia Airlines flight that vanished in 2014. The company's improved underwater robots will search a targeted 15,000-square-kilometer area starting December 30.

Noninvasive Blood Glucose Monitor Could Replace Finger Pricks – MIT researchers built a device that measures blood sugar using light instead of needles. Tests showed accuracy similar to sensors implanted under the skin, and a wearable version is now in clinical trials.

New AI Video Approach Matches Quality of Current Models – Researchers introduced STARFlow-V, a 7-billion-parameter video generator that works differently than standard AI video tools. The system can create videos from text, extend existing clips, and edit footage using a single model.

VC Firm Maps Out Vision AI Opportunities Before Robotics Era – Bessemer Venture Partners outlined where they see AI vision software creating value: construction progress tracking, healthcare monitoring, manufacturing quality checks, and home assistants that understand physical spaces.

TRY THIS PROMPT (copy and paste into Claude or Gemini(click build first))

Habit Stack Designer: Build a personalized morning and evening routine using habit stacking science

What this does: Creates a habit-building tool that uses the proven technique of "habit stacking" (linking new habits to existing ones) to build sustainable morning and evening routines—with streak tracking, swipe gestures, and research-backed routine examples.

What this looks like:

WHERE WE STAND (based on today’s stories)

✅ AI Can Now: Generate and edit videos in a single unified model, handling multiple inputs at once
❌ Still Can't: Consistently pull data from different business applications without reliability issues

✅ AI Can Now: Power underwater robots precise enough to search 15,000 square kilometers of ocean floor
❌ Still Can't: Autonomously complete complex multi-step business tasks that enterprises are willing to pay for

✅ AI Can Now: Analyze 13 years of premium journalism to detect market-moving narratives and sentiment patterns
❌ Still Can't: Guarantee basic security—a billion-dollar legal AI tool was found with zero authentication protecting 100,000+ confidential files

FROM THE WEB

Traditional video production requires jumping between 5-10 different tools — one for generation, one for editing, one for masking, one for color grading, one for effects. Each handoff introduces errors and eats time. The model addresses "the pipeline problem" where errors accumulate as assets move between different specialized models.

This approach replaces the traditional, labor-intensive workflow of manual rotoscoping, masking, and keyframing, which often requires specialized software and significant time investment.

RECOMMENDED LISTENING/READING/WATCHING

SHORT FILM: Sunspring (2016)

This nine-minute sci-fi film was written entirely by an AI after being trained on dozens of sci-fi screenplays. The result is beautifully bizarre. The dialogue makes no sense, characters say things like "I don't know what you're talking about, but I'm not a bright future.”

Director Oscar Sharp and AI researcher Ross Goodwin fed the AI a corpus of sci-fi scripts and had it generate a screenplay. Then they got actual actors—including Thomas Middleditch from Silicon Valley, to perform it completely straight. The film is fascinating because it shows you exactly what AI creativity looked like in 2016: technically impressive but fundamentally alien when there is no human editing or creativity directly involved.

Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!

-James

By the way, this is the link if you liked the content and want to share with a friend.