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

Anthropic Adds Bank of England to Growing List of Regulators Alarmed by Their Announcement

TLDR: UK financial regulators are rushing to evaluate cybersecurity risks from Anthropic's new Mythos AI model after the U.S. Treasury Secretary and Federal Reserve Chair held an emergency meeting with the CEOs of America's biggest banks about the same tool.

The Story:

The Bank of England is putting Anthropic's Claude Mythos model on the agenda for its next meetings with banks and regulators, according to Reuters. The FCA, UK Treasury, and National Cyber Security Centre will all participate. This follows an urgent meeting days earlier where Treasury Secretary Scott Bessent and Fed Chair Jerome Powell briefed CEOs from Bank of America, Citi, Goldman Sachs, Morgan Stanley, and Wells Fargo on the model's capabilities. Anthropic says Mythos can find and exploit security weaknesses across major operating systems and web browsers, including flaws that went undetected for 30 years. The company limited its release to a small group through a controlled program called Project Glasswing, with partners including JPMorgan, Apple, Google, Microsoft, and Nvidia. Canada's central bank convened its own lenders around the same issue, making this a coordinated three-country response.

Its Significance:

Whether intentional or not, Anthropic is running one of the best hype campaigns in AI right now. Saying they built something so powerful that governments had to hold emergency meetings about it. That framing does two things at once. It positions Anthropic as the responsible safety leader in the industry (they flagged the risks before anyone else did, restricted access voluntarily, and briefed world governments directly). And it generates massive interest in a model most people can't touch yet. Some observers have speculated there's another reason for the limited rollout: Anthropic may not have the compute resources to release Mythos broadly even if it wanted to. True or not, the scarcity only adds to the mystique. Competitors are left explaining why they didn't do the same, while Anthropic controls the conversation. Don't expect to use this model anytime soon. This also sets the stage for exclusivity and extremely high pricing.

QUICK TAKES

How a Dancer With ALS Used Brainwaves to Perform Live

The story: In December 2025, professional dancer Breanna Olson, who lives with ALS, performed on stage in Amsterdam using only her brainwaves. Dentsu Lab's Project Humanity captured her neural signals and translated them in real time into choreography performed by a digital avatar. An estimated 200 million people worldwide live with serious physical disabilities that make standard computer interfaces unusable.

Your takeaway: Most tech is built for people who can swipe, tap, and type. This project shows that brain-computer interfaces aren't just lab experiments anymore. They're putting people back on stage.

Palantir CEO Says AI 'Will Destroy' Humanities Jobs, But Vocational Workers Will Thrive

The story: Palantir CEO Alex Karp told the World Economic Forum in Davos that AI will eliminate humanities-based jobs and that only two types of people have a clear future: those with vocational training and those who are neurodivergent. Youth unemployment for ages 16 to 24 hit 10.4% in December. Anthropic co-founder Daniela Amodei directly disagreed, saying humanities skills will be "more important than ever."

Your takeaway: Two major AI company leaders giving opposite advice tells you something: nobody knows for sure. But the trend toward valuing hands-on skills alongside creative thinking is getting louder. If you're planning your career, building something you can do with your hands and something you can do with your brain isn't bad advice.

Connecticut Pushes One of the Country's Strongest AI Employment Laws

The story: Connecticut's Senate Bill 435 would require employers to notify workers whenever AI tools are used in hiring, firing, promotion, or performance decisions. It also mandates bias testing and gives employees a private right to sue for algorithmic discrimination. If passed, it takes effect October 1, 2026.

Your takeaway: Companies already use AI to screen resumes, rank candidates, and flag employees for termination. This kind of oversight is going to spread to more states. The bigger picture: the decisions that affect your job are increasingly being outsourced to AI, and lawmakers are starting to catch up.

TOOLS ON OUR RADAR

📝 AFFiNE Free and Open Source: A next generation collaborative workspace that combines your documents whiteboards and databases into one seamless and privacy focused environment. (Alternative to Notion)

🌀 Maze Freemium: A creative generative art platform that operates directly within Discord allowing you to explore new worlds and generate high quality images through simple chat commands.

🐰 Orbiit Freemium: A dedicated research tool that allows you to easily track competitor changes and monitor ongoing market activity so your team always stays informed about the latest industry shifts.

🔮 Scenario Edge Freemium: An intelligent financial analysis platform that helps you stress test your investment portfolios using advanced scenario based simulations to prepare for various macroeconomic conditions.

TRENDING

New Tools Aim to Make AI 'Vibe Coding' Safer for Crypto - The ASI Alliance and developer platform Matterhorn announced a partnership to add security auditing tools to AI-generated smart contracts. Their "Vibe-Audit" system combines AI security scanning with human review before any code goes live on the blockchain. They're targeting 20,000 developers on ASI:Chain in 2026.

MIT's CompreSSM Makes AI Models Smaller While They're Still Training - MIT researchers built an algorithm called CompreSSM that figures out which parts of an AI model aren't pulling their weight after just 10% of training. It then removes those parts and keeps training the smaller, faster version. On image classification tests, compressed models trained up to 1.5x faster while keeping nearly the same accuracy.

16 New Startups Join MIT's Hard-Tech Accelerator - MIT.nano's START.nano program more than doubled its new company count in 2025, adding 16 startups working on semiconductors, quantum computing, energy, and health. The accelerator now supports over 30 companies, almost half with MIT roots, by offering discounted lab access and connections to the MIT network.

AI System Keeps Warehouse Robots From Crashing Into Each Other - MIT and Symbotic built a system that uses deep reinforcement learning to manage traffic among hundreds of warehouse robots. It decides which robots get the right of way at each moment and reroutes them before bottlenecks form. In simulations, throughput jumped 25% compared to existing methods.

College Students Build 'Acutis AI,' a Catholic-Grounded Chatbot - Brothers Peter (21) and Thomas (19) Cooney, students at the University of Dallas and Baylor University, built an AI platform that answers moral questions using the Catechism, encyclicals, and the Summa Theologica as its source material. It also lets parents monitor their kids' chats and set time limits. Named after St. Carlo Acutis, the teen saint known for using technology to spread his faith.

Apple Reportedly Racing to Launch AI Smart Glasses by Late 2026 - Bloomberg reports Apple is accelerating work on smart glasses with built-in cameras, microphones, and speakers. The device will compete with Meta's Ray-Ban smart glasses and run on Siri. Apple CEO Tim Cook has reportedly made the project his highest priority. The company also recently lost its AI chief John Giannandrea to retirement.

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

📊 Build a daily habit tracker that shows your entire month as a color-coded grid you can click to fill in as you go

What this does: Creates a 30-day visual grid with one row per habit and one column per day. Clicking any past or current day toggles it between complete (colored) and incomplete. The app automatically calculates each habit's completion rate and current streak, and locks out future days so you can't check off things that haven't happened yet.

What this looks like:

WHERE WE STAND(based on today’s news)

✅ AI Can Now: Detect up to six different types of hidden atomic defects inside a semiconductor without breaking it open, at concentrations as low as 0.2%

❌ Still Can't: Deploy that detection in a factory setting quickly, because the neutron-scattering technique requires specialized facilities that most manufacturers don't have

✅ AI Can Now: Translate a person's brainwaves into real-time choreography performed by a digital avatar on a live stage

❌ Still Can't: Give users with physical disabilities fine-grained control over complex tasks through brain-computer interfaces alone, since the technology is still limited to high-level signals rather than precise motor commands

FROM THE WEB

RECOMMENDED LISTENING/READING/WATCHING

Synchronic (2019) - Movie

Two New Orleans paramedics keep showing up to the same kind of scene: young people dead or badly hurt in ways that don't make sense, all connected to a new designer drug called Synchronic. When one of them tries it himself to figure out what's happening, he discovers it doesn't do what anyone thought. Directed by Justin Benson and Aaron Moorhead, who have a gift for making small budgets feel enormous. It's a time travel film that treats the concept more seriously than most, grounding the science in actual physics research and never letting the high concept swallow the human story underneath it.

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.}