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 edited, and published with the intention of making AI news and technology more accessible to everyone.

THE FRONT PAGE

AI Fraud: 1 Man, 4 Billion Fake Streams, $10 Million Made

TLDR: A North Carolina man pleaded guilty to stealing over $10 million in music royalties by using AI to create fake songs, then running bots to stream them billions of times across Spotify, Apple Music, Amazon Music, and YouTube Music.

The Story: Michael Smith, 54, of Cornelius, North Carolina, pleaded guilty on March 19 to conspiracy to commit wire fraud. He bought hundreds of thousands of AI-generated songs from an accomplice, uploaded them to every major streaming platform, and set up more than 1,000 bot accounts to play those songs nonstop. Each bot could stream around 636 songs per day. At roughly half a cent per stream, the scheme pulled in about $3,300 a day, over $99,000 a month, and more than $1.2 million per year. The whole operation ran from 2017 to 2024 — seven years — before the platforms and the Mechanical Licensing Collective finally caught on. To stay off the radar, Smith used VPNs and gave his fake artists names like "Calvinistic Dust," with track titles like "Zyzomys." Prosecutors are calling this the first criminal streaming fraud case in the country. Smith agreed to forfeit $8,091,843.64 and faces up to five years in prison. Sentencing is set for July 29.

Its Significance: Every time a bot streamed one of Smith's fake songs, a real artist got slightly less money. Streaming royalties are split across a giant pool, and when fake streams flood that pool, legitimate musicians get a smaller cut. This is the first time someone has been criminally convicted for this kind of AI-powered streaming fraud, which sets a legal line: using AI and bots to steal royalties is now officially a federal crime with real consequences. With AI music tools generating millions of songs per day, don't expect Smith to be the last person who tries this.

QUICK TAKES

AI Company Leaders Are Openly Talking About Government Takeover of AI

The story: Palantir CEO Alex Karp warned at the a16z summit that if Silicon Valley takes white-collar jobs while refusing to work with the military, the government will nationalize AI companies. OpenAI CEO Sam Altman followed by publicly musing that it might actually be better if building AGI were a government project.

Your takeaway: When the CEOs of major AI companies are openly discussing nationalization, it signals just how much pressure these firms are under from Washington. For everyday people, a government-controlled AI industry would look very different from today's private one — and not necessarily better.

The US Military Is Investigating Whether AI Helped Pick the Target That Killed 175 People at an Iranian School

The story: A US Tomahawk missile strike on February 28 destroyed an elementary school in Minab, Iran, killing at least 175 people, most of them young girls. The Pentagon confirmed it uses AI tools to sift through targeting data, but refused to say whether AI played a role in selecting that specific school. A preliminary investigation found that officers used outdated intelligence to generate the coordinates — the school had been separated from a nearby military base all the way back in 2016.

Your takeaway: This case shows the real-world cost when AI speeds up military decisions without enough human checks. Whether AI flagged the target or not, the situation makes clear that when AI builds target lists faster than humans can verify them, deadly mistakes can slip through.

OpenAI Is Planning to Nearly Double Its Workforce to 8,000 Employees by End of 2026

The story: According to a Financial Times report, OpenAI plans to grow from roughly 4,500 employees to 8,000 by the end of 2026. Most new hires will go into product development, engineering, research, and sales. The company is also adding specialists in "technical ambassadorship" roles — people whose job is to help businesses actually put OpenAI's tools to work. This comes as Anthropic is reportedly capturing 73% of first-time enterprise AI spending, putting real competitive pressure on OpenAI's commercial side.

Your takeaway: While much of Big Tech has been cutting jobs, OpenAI is doing the opposite and betting that more people means a faster path to winning the enterprise market. This hiring surge also signals that AI companies aren't just competing on models anymore — they're competing on who can get businesses to actually use them.

TOOLS ON OUR RADAR

🐧 Budibase Free and Open Source: A powerful low code platform that allows you to build custom internal tools and business applications with a secure open source foundation. (Alternative to Retool)

🎞️ *OpusClip Freemium: A highly intelligent generative video tool that automatically takes your long form videos and repurposes them into viral short clips perfectly optimized for social media growth.

✨ Stitch Freemium: A design companion from Google that uses artificial intelligence to instantly transform your rough product ideas into beautiful production ready user interfaces in seconds.

🎙️ Respeecher Paid: A remarkably advanced voice cloning technology that allows you to transform one human voice into another with incredible emotional depth and accuracy for professional film production or creative projects.

TRENDING

AI Cow Collars Could Be Worth $2 Billion — A New Zealand startup called Halter makes solar-powered AI collars that let ranchers herd cattle through a smartphone app instead of physical fences. Peter Thiel's Founders Fund is reportedly close to leading a round that would push Halter's valuation past $2 billion. Beef prices climbed 18% in one year as the US cattle herd hit a 75-year low, and ranchers are desperate for ways to cut labor costs.

Russian Troops Say They're Fighting Actual "Terminator" Robots in Ukraine — A pro-Kremlin lawmaker told Russian media that two Phantom MK-1 humanoid robots from US company Foundation have arrived at the Ukraine frontline, and that Russian soldiers are "at their wits' end." The Phantom can carry rifles and breach doors. Foundation already holds $24 million in US military contracts across the Army, Navy, and Air Force.

Researchers Built a Robot That's Nearly Impossible to Destroy — Northwestern University researchers let AI design a robot from scratch, giving it modular building blocks and one goal: keep moving. The result is a "legged metamachine" that keeps going even when modules break off or sustain damage, finding new ways to move with whatever parts still work.

Cursor Quietly Built Its New AI Coding Model on a Chinese Foundation — Cursor, the $29.3 billion AI coding startup, launched its new Composer 2 model as "frontier-level coding intelligence" without mentioning it was built on Kimi 2.5, an open-source model from Chinese company Moonshot AI (backed by Alibaba). A user spotted the connection in the code. Cursor admitted the oversight, though says 75% of the final model came from its own training and the use was fully licensed.

Amazon's Custom AI Chip Is Now Running Claude — and Soon OpenAI Too — TechCrunch got a rare tour of Amazon's Trainium chip lab in Austin, where 1.4 million Trainium chips are deployed and Anthropic's Claude runs on over 1 million of them. As part of Amazon's new $50 billion deal, AWS will supply OpenAI with 2 gigawatts of Trainium capacity. Apple is also reportedly testing the chips. This could put real pressure on Nvidia's near-monopoly on AI hardware.

Alibaba Wants Tens of Millions of US Users for Its New AI Agent — Alibaba launched Wukong, an enterprise AI agent platform that coordinates multiple AI assistants from a single interface. It's already embedded in DingTalk (20+ million corporate users) and plans to expand to Slack, Microsoft Teams, and WeChat. The platform handles document editing, meeting transcription, approvals, and research. Alibaba is clearly pushing hard into the US despite trade tensions.

Zuckerberg Is Building a Personal AI Agent to Help Run Meta — According to a Wall Street Journal report, Mark Zuckerberg has a personal AI agent in development that retrieves information faster than going through layers of employees. Meta workers are also using tools including "Second Brain," built on Anthropic's Claude, which is being described internally as an "AI chief of staff." Meta is spending up to $135 billion on AI infrastructure in 2026 alone.

TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)

💰 Build a monthly budget planner that tracks spending vs. budget across categories and shows where your money actually goes

What this does:

It's a personal monthly budget tracker where you set your income, then build a list of spending categories (rent, groceries, Netflix, etc.) with a target budget and what you've actually spent so far. Each item shows a progress bar that turns red when you've gone over. The app also rolls up your numbers into category summaries and running totals so you always know what's budgeted, what's spent, what's left, and how much income you haven't assigned yet. Everything lives in a clean dark green interface and you can update your spending numbers directly in the list without any extra steps.

What this looks like:

WHERE WE STAND(based on today’s news)

✅ AI Can Now: Generate hundreds of thousands of songs convincing enough to fool major streaming platforms' fraud detection for seven years straight.

❌ Still Can't: Reliably verify military target data against current civilian maps without human oversight, as the Iran school bombing investigation shows.

✅ AI Can Now: Design robots from scratch by evolving modular structures, producing machines that keep moving even after losing parts.

❌ Still Can't: Replace the full decision-making role of a CEO without a human still in the loop, even as Zuckerberg builds an AI agent to speed up his own workflow.

FROM THE WEB

Yes, they really went with the name Skynet…

RECOMMENDED LISTENING/READING/WATCHING

Nemesis (Boardgame)

For science-fiction lovers. Semi-cooperative survival horror on a spaceship infested with alien creatures; basically the Alien movie as a board game without the license. You have secret objectives, so your crewmates might be working against you. Dripping with tension, and one of the most atmospheric gaming experiences you can have at a table. Supports 1-5 players and runs 90-180 minutes.

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.

^{Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.}