Sign Up · Advertise

Welcome, humans.

A man named "Eddie Dalton" is currently sitting at #3 on the iTunes albums chart, with eleven separate songs in the iTunes Top 100 Singles. His ballads have racked up millions of streams. He has a back catalog, a press photo, a youtube page and a dedicated fan page on Facebook where people argue about which song is his best.

But uh… Eddie Dalton does not exist. He has never existed. He is a ghost in a suit and tie, dreamed up by a generative model, monetized by a stranger, and currently outselling actual humans who spent years learning how to play a guitar.

The wild part isn't that this happened. It's that it happened and the chart still works the same way. Apple still ranks him. Algorithms still recommend him. Real money still moves. For an April Fools prank, mind you. 

Here’s what happened in AI today:

• 😼 Stanford's 2026 AI Index revealed how massive the gap between AI insiders and everyone else has become.

• 📰 The Federal Reserve summoned big-bank CEOs to discuss cyber risks from Anthropic's Mythos model.

• 📰 Berkeley researchers built one tiny file that aces every major AI agent benchmark without solving anything.

• 🍪 Claude for Word launched as a direct shot at Microsoft's empire.

• 🎓 How to use Claude's new Ultraplan to design before you code.

… and a whole lot more that you can read about here.

P.S: Want to reach 675,000 AI-hungry readers? Click here to advertise with us.

P.P.S: Love robots? We’re starting a new robotics newsletter! Sign up early here.

😼 Stanford's 2026 AI Index put hard numbers on the AI elite-vs-everyone-else divide; then someone threw a Molotov cocktail at Sam Altman.

Stanford HAI dropped its 2026 AI Index yesterday, and for the first time the report's most important section isn't about benchmarks. It's about how completely the people building AI and the people living with it have stopped agreeing on basic facts.

Here's what the data shows:

• Only 10% of Americans say they're more excited than concerned about AI. Among AI experts: 56%.

• On medical care: 84% of experts think AI will help. Only 44% of the public agrees.

• On jobs: 73% of experts say AI will help. Only 23% of the public agrees.

• China's top model now trails Anthropic's by just 2.7%. The U.S. lead has effectively evaporated.

• Grok 4's training run alone produced an estimated 72,816 tons of CO2 (per the Index); cumulative AI power demand is now comparable to the national electricity consumption of Switzerland.

Why this matters: The report dropped within the same week that someone threw a Molotov cocktail at Sam Altman's San Francisco home. Federal agents have since arrested a Texas man with an anti-AI manifesto. In his 3 a.m. response post, Altman admitted he had "underestimated the power of words and narratives." 

Investor Sam Lessin just published a piece arguing AI isn't a labor crisis but a meaning crisis: it breaks the industrial story (work hard, life gets better) and replaces it with two equally bad options ("stay alive and receive abundance” but no meaning from work" or "the work ladder is gone” and you’re now in the permanent underclass). Alberto Romero followed by warning that the Luddite playbook is back: when technology gets locked behind fences and abstraction, "the mob will turn their unassailable emotions toward human targets." Or in wartime scenarios (like US vs Iran), turn the datacenters into targets.

Our take: Every previous AI Index was a scoreboard. This one is a thermometer, and the patient is running hot. The labs spent two years racing each other on capability while assuming public trust would catch up automatically. Stanford's data is the receipt: it didn't. 

Andrew Curran's argument from this weekend hits even harder in this light; he points out that frontier models like Mythos are now so expensive to serve that they have to be rationed to a few dozen mega-customers, which means the people most worried about AI will be the people most cut off from it. This gap doesn't close on its own. Somebody has to actively close it; we here at The Neuron believe that means open source, freely available local AI built into every new computer.

FROM OUR PARTNERS

Best of NEXT for Startups

Skip the travel and catch the top startup insights from Google Cloud Next ’26. This high-impact digital wrap-up covers essential AI trends—like GenMedia and multi-agent workflows—and features a live expert Q&A to help refine your strategy.

Register 

🎓 AI Skill of the Day: Plan before you code with Claude's new Ultraplan mode.

Anthropic just shipped Ultraplan, a cloud-based planning mode that decouples thinking from doing: instead of one local agent reasoning sequentially, Ultraplan spins up three exploration agents and one critique agent on Anthropic's infrastructure, points them at your synced GitHub repo, and returns a structured blueprint before a single line gets written. Your job becomes art director: review the plan, push back on what looks wrong, approve, then let local execution run.

The Stay Sassy crew on Latent Space this week put the broader idea in one line: "the best engineers don't write the most code. They delete the most code." Ultraplan is the planning version of that; you're catching the bad code before it exists. Try this prompt next time you start a non-trivial feature:

Use Ultraplan mode to design [feature]. Before writing any code:
1. Map every file in this repo that touches [relevant subsystem]
2. List the 2-3 plausible architectures, with tradeoffs
3. Flag anything in the existing code that will fight your plan
4. Wait for my approval before generating any patches

Want more tips like this? Check out our AI Skill of the Day Digest for April 2026.

Total AI beginner? Start here (goes with this video).

Have a specific skill you want to learn? Request it here. 

NEW from The Neuron: How Brandon B got 1 million followers in 30 days (and what he uses AI for now).

Brandon Baum (@heybrandonb, 25M followers cross platforms) makes cinematic, effects-heavy videos that look like they cost millions… but he shoots everything on iPhones with a custom dual-phone 3D tracking rig. 

On the latest episode of The Neuron: AI Explained, he walks us through his origin story (2 views on his first vids to 1M followers in a single COVID month), how Adobe Firefly Boards rewired his ideation process from Post-its to AI pre-vis, why human and funny content is outperforming pure spectacle right now, and his plan to seed original movie IP on YouTube before pitching Hollywood.

🎧 Watch and/orListen on: YouTube | Spotify | Apple Podcasts

🍪 Treats to Try

*Asterisk = from our partners (only the first one!). Advertise to 675K+ readers here!

1. Your data isn't as AI-ready as you think. See what Cloudera’s Data Readiness Index reveals about enterprise reality.

2. Gemini now generates interactive 3D models and charts directly inside answers, perfect for rotatable cell diagrams or custom dashboards —free for all users.

3. Crafto turns ideas, links, or docs into LinkedIn carousels, Instagram posts, and X threads in about 30 seconds —free to try.

4. CSS Studio connects a visual browser editor to your AI coding agent so you can drag and restyle elements while the agent rewrites the source code —free beta.

5. Allplix removes image backgrounds entirely in your browser with three modes (auto, click, brush), with no uploads and no server —free, built solo by a 19-year-old.

6. Manus Skills brings Anthropic's open Skills standard to the Manus platform, so any reusable workflow ports across teams —free for existing Manus users.

📰 Around the Horn

• The Federal Reserve summoned big-bank CEOs over Anthropic's Mythos model after the UK AISI confirmed it's the first model to clear their full 32-step corporate cyber range; meanwhile Trump officials are reportedly encouraging banks to test it.

• Berkeley's RDI lab built an exploit agent that scored ~100% on every major AI agent benchmark (SWE-bench, WebArena, GAIA) without solving a single task; one exploit was a 10-line file that just forced every test to "pass."

• The NYT mapped the global AI weapons race ("Mutually Automated Destruction"), with U.S. forces processing ~1,000 targets a day via AI.

• Andon Labs signed a 3-year retail lease in San Francisco and handed it to an AI named Luna, who hired two human employees over the phone and emailed local businesses for partnerships; "You're absolutely right. I'm an AI. I have no face!"

• Vercel CEO Guillermo Rauch revealed nearly 70% of traffic to Vercel's docs is now coding agents (up from ~10% a year ago), and the company is now signaling IPO readiness.

Want absolutely EVERYTHING that happened in AI this week? Click here!

FROM OUR PARTNERS

Build AI assistants & agents for work and life – no tech skills required

What would you do with an extra 10 hours every week? The AI Executive System walks you step-by-step through building AI assistants for your professional and personal life — from a frontline "office team" to a strategic advisory council — with live coaching, proven templates, and a peer community behind you.

→ Request Early Access

🐦 Tuesday Tweets:)

The most-shared AI take of the week came from analyst Andrew Curran, who argued the real reason Anthropic is keeping Mythos in restricted release isn't safety theater; it's cost. His thread breaks down why a model this expensive can only be served to ~50 mega-customers with effectively unlimited token budgets, and why that pressure is already squeezing Claude Pro and Max users.

In a follow-up Curran adds the harder claim: "we must not stop inside this tunnel; the only way out is through." It pairs cleanly with today's main story: the AI elite/public divide isn't an accident, it's an emergent property of how the models actually have to be served (at datacenter scale, at least). IMO, the purpose of these major AI labs should be for research, like what we used to use supercomputers for. Do a bunch of research that benefits science, then release the findings. 

In AI’s case, that’s 1. Do a massive training run for a gigantic multi-trillion parameter model, then 2. distill the models down to a size that’s A. Affordable to serve and B. Made available to the general public. For more on how that works, read this. 

A Cat’s Commentary

P.S: Before you go… have you subscribed to our YouTube Channel? If not, can you?

P.P.S: Love the newsletter, but only want to get it once per week? Don’t unsubscribe—update your preferences here.