In partnership with
Welcome, busybees
A free AI model with no name attached quietly showed up on OpenRouter this week, and it's driven a good chunk of the AI world into a genuine guessing game. Ox Alpha, as it's labeled, handles coding and long, multi-step agent work well enough that Stripe CEO Patrick Collison called it "very impressive" in public. Nobody will say who built it. Speculation has bounced between a Chinese lab, an unreleased Microsoft model, and a well-funded stealth startup, and as of this morning, still nobody actually knows.
Meanwhile, a 12-person London lab most people haven't heard of just made a genuinely bold claim: its AI agent beat both Anthropic's and OpenAI's much larger models at a real scientific task, using a fraction of the size. Here's what else is worth knowing, plus real roles across engineering, design, and sales you can apply to right now.
Here’s what happened in AI today:
• An anonymous AI model called Ox Alpha has the AI world guessing who built it.
• A tiny London startup says its small AI agent outperformed Anthropic and OpenAI at replicating scientific research.
• Nvidia just partnered with a new data center developer as it keeps expanding its footprint across the AI buildout.
• Nvidia's own product team is arguing the software wrapped around an AI model now matters more than the model itself.
• Anthropic's newest model, Opus 4.6, is facing sharp criticism over how it handles certain content.
• The Department of Justice is investigating Andreessen Horowitz, raising real questions for the venture capital world funding AI.
…and a whole lot more that you can read about below.
• Nobody Knows Who Built the AI Model Everyone's Suddenly Talking About
A free model called Ox Alpha appeared on OpenRouter on Thursday, listed only as a "stealth model" built by "a third-party provider who has chosen to remain anonymous during this preview." It's a reasoning model aimed at coding and sustained agentic work, with a context window over one million tokens, and it's already climbed near the top of leaderboards. Stripe CEO Patrick Collison, whose company just acquired OpenRouter, called it "very impressive" publicly. Early theories pointed to Z.ai, the Chinese lab behind the GLM model family, based on tokenizer and behavioral fingerprinting, but by Friday morning even that theory was losing confidence among the people tracking it closely. Others have floated an unreleased version of Microsoft's MAI.
• A 12-Person London Startup Says It Just Beat Anthropic and OpenAI at Their Own Game
Inherent, a London lab founded by DeepMind alumni, says its agent Faraday outperformed Claude Opus 4.8 and GPT-5.5 at independently reproducing published scientific research. The twist: Faraday runs on a comparatively tiny 27-billion-parameter model, trained via reinforcement learning to develop "research taste." The company raised a $50 million seed round just weeks ago and plans to grow from 12 to 20-25 people by year-end.
• Nvidia Bought Its Way Into Another Piece of the AI Buildout
Nvidia took a minority stake, reportedly worth several hundred million dollars, in Cloverleaf Infrastructure, a company that preps land, power, and site access for data centers. Cloverleaf will now build using Nvidia's own DSX platform. It's the same playbook Nvidia used with its $1.5 billion investment in an OpenAI-linked Ohio data center project earlier this month.
• The Justice Department Is Investigating One of AI's Biggest Investors
The DOJ has spent nearly a year investigating Andreessen Horowitz over whether two partners improperly sit on the boards of competing companies, Databricks and Fivetran, invoking a rarely used 112-year-old antitrust law. Regulators haven't decided whether to pursue enforcement. Why it matters: nearly every hot AI startup takes a board seat from a firm like a16z, so forcing VCs to give up overlapping seats could reshape how funding rounds get structured industry-wide.
Today's Job Board
Founding Product Engineer, Full Stack | b12 Labs
San Francisco Bay Area, California (Full-time)
Help make hard molecules for pharma and biotech at an early-stage startup.
Software Engineering Intern/Co-op | Clearline
Kitchener, Waterloo, or Toronto, ON, Canada (Internship)
Build the AI operating system that helps car dealerships capture missed revenue.
Software Engineer, Fullstack | Aviator
Remote (Full-time)
Build a Google-level engineering productivity suite for fast-moving software teams.
Founding AI Engineer | Voiceops
United States (Full-time)
Help B2C companies learn from every customer interaction using AI.
UX Designer | Coulomb AI
United States (Full-time)
Design the interface for a battery observability platform for electric vehicles.
Staff Software Engineer | Confido
United States (Full-time)
Build the AI-enabled financial automation and intelligence platform used by CPG brands.
Forward Deployed Engineer | Regbase
New York, NY (Full-time)
Help automate corporate compliance work for companies worldwide.
Head of Business Development | Malama Health
San Mateo, California (Full-time)
Grow partnerships for a startup providing doula-led, holistic pregnancy and postpartum support.
Mechanical Engineer Intern, Winter 2026 | GRU Space
San Mateo, California (Internship)
Help build the first hotel on the Moon..
Reply to everything. Edit nothing.
Your inbox is full. Slack is piling up. Client messages need a response yesterday. Typing thoughtful replies to all of it takes hours you don't have.
Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. Speak like you would to a colleague — tangents and all — and get polished output. Emails, Slack, LinkedIn, WhatsApp, whatever's open.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Works on Mac, Windows, and iPhone.
Today's Money Moves 💸
Starcloud Raises $250M for Orbital Data Centers
Starcloud raised $250 million to build data centers in orbit, betting that space-based compute becomes more attractive as launch capacity on Earth gets harder to secure.
How Rillet Became a Unicorn in 48 Hours
AI accounting startup Rillet raised $100 million and crossed a unicorn valuation within 48 hours of the round coming together, one of the fastest unicorn turnarounds of the year.
Today's Growth Recommendation
You hire an agency every couple of years. Vendry does it every day. Tell them your budget and needs, and their experts shortlist 3-4 vetted agencies, free, in under 7 days.
Get Matched Free Below ↓
Find the perfect marketing agencies in <15 minutes
You hire an agency every couple of years. We do it every day.
Tell us your budget and what you need
Our experts shortlist from agencies we've already vetted
You get 3 or 4 agencies worth your time
It costs nothing and <7 days
Skill of the Day
Turn Gemini Into a 24/7 Assistant That Actually Follows Through
Gemini Spark takes Gemini out of one-shot chatbot mode and turns it into an agent that can actually carry out multi-step tasks on its own (currently limited to Google AI Ultra).
Creator Stewart Gauld put it through a real workflow: "Identify business events near me over the next 90 days, add them to my calendar, and email me a summary of the ones I should attend." Spark researched the events, added six straight to his calendar, and sent the summary, no follow-up prompting needed at any step.
Here's the part worth borrowing: any task you've run once can become a reusable "skill."
Just say "turn this into a skill called ___" and it locks in the whole workflow, ready to fire off again with a short phrase. Once that's set, put it on a recurring schedule ("run this every weekday at 9am") so it keeps running quietly in the background without you asking twice.
Identify [category] events near [location] over the next [N] days.
Add them to my calendar.
Email me a summary of the ones I should prioritize.
Invite friends & get instant rewards
If you enjoy The Incredible Edge, share it to unlock free resources when they subscribe.