6 min read

🛎️ New AI Benchmark

More Stories: OpenAI Rides Google’s Rails, Lovable Hits $100M ARR

Good Morning, AI Enthusiasts!

Google’s helping OpenAI scale, Lovable skipped adolescence and became a centaur. Everyone’s winning, but at what cost?



BECHMARK

Top AI Coder Wins $50K for Solving 7.5% of The Test

📌 What’s happening: The Laude Institute, joined by Databricks and Perplexity co-founder Andy Konwinski, just launched the K Prize—a new AI coding challenge designed to test real-world problem-solving. The first-round winner, Brazilian prompt engineer Eduardo Rocha de Andrade, scored only 7.5% and still took home the $50,000 prize. Konwinski says that’s good news: the benchmark is clean, hard, and cheat-resistant—unlike SWE-Bench, which may be bloated from model pretraining contamination.

🧠 How this hits reality: AI’s coding chops aren’t production-ready. K Prize exposes a tough truth: many models may just be regurgitating training data. If your LLM evaluation still leans on SWE-Bench, you're probably overestimating what your stack can do. Dirty benchmarks make for easy demos, not real deployments.

🛎️ Key takeaway: AI benchmarks only matter if the data’s clean—otherwise, you're just grading memory, not intelligence.


TOGETHER WITH PLESK

All-in-One Server Panel

Plesk is a commercial WebOps control panel for managing Windows or Linux servers via an intuitive web interface. It supports website, domain, email, and database management, plus integrated security tools like firewalls, SSL, and automatic updates.

With extensibility through 100+ plugins—including Docker, Git, SSL, and WordPress Toolkit—it suits developers, agencies, and hosting providers.


CLOUD

Google Builds the House, OpenAI Throws the Party

📌 What’s happening: Google CEO Sundar Pichai told investors he’s “excited” to support OpenAI—Google’s biggest AI rival—on Google Cloud. With Microsoft hitting GPU capacity limits, OpenAI has turned to Google and Oracle to fuel model training. Meanwhile, Google Cloud posted $13.6B in Q2 revenue, up 32% YoY, largely driven by AI demand.

🧠 How this hits reality: Google Cloud is becoming the AWS of model training—open for business, even to enemies. That’s great for revenue, but terrible for strategy. Cloud’s biggest customers are the same companies threatening Google Search’s existence. Execs call it “open,” but this smells like déjà vu—Google once helped Yahoo grow, then ate its lunch. Now it’s laying the same track for OpenAI.

🛎️Key takeaway: You can’t be the referee and sell sneakers—unless you plan to rewrite the game.


STARTUP

Lovable Claims $100M ARR in 8 Months

📌 What’s happening: Swedish upstart Lovable claims it has reached $100M ARR in under eight months, joining the Centaur Club just after becoming Europe’s newest unicorn. With only 45 employees, it’s powering 2.3 million active users and over 10 million projects. The company also revamped its pricing strategy, sunsetting its Team plan and introducing a new Business tier aimed squarely at the mid-market.

🧠 How this hits reality: This is the stuff of VC fever dreams—AI-native, brutally efficient, and monetizing at warp speed. Lovable started as a slick prototyping tool for no-code builders, and now it’s eyeing the enterprise. But its “vibe coding” ethos still triggers compliance side-eyes in boardrooms.

🛎️ Key takeaway: Lovable did in 8 months what takes others 8 years—turns out European AI SaaS can move like it’s from California.


QUICK HITS

  • Google's AI Overviews are now used by two billion people monthly, while its new AI search mode has attracted 100 million users in the US and India.
  • Proton has launched a new privacy-focused AI assistant that uses end-to-end encryption and keeps no logs of user conversations.
  • A report from Holistic AI suggests that structured adversarial testing could have prevented the public failures of X.AI's Grok 4 model.
  • Spurred by an exodus from Twitter, rival platform Mastodon has launched in-app donations to fund its growth.
  • Google Photos is introducing new generative AI features that can stylize photos and turn them into cinematic videos.

TRENDING

Daily AI Launches

  • Clearitty released a sales platform that targets truly in-market accounts using intent data and AI scoring.
  • Commitify released commitify.me, an AI agent that calls your phone to keep you accountable.
  • Qwen released Qwen3-Coder, a 480B MoE model that excels at coding with 1M context support.
  • 🎯 atypica.AI automates market research in 10 minutes, simulating consumers to reveal key insights.
  • 📚 Heardly is the Fast Way to read Best Book.
  • 🪶 CopyOwl is the First AI Research Agent, deep research on any topic in one click.

PREVIEW

Meet the Future: Our New Newsletter Brand

Spectacle Service Economy

Where once you pumped gas, now you share popcorn with robots. Where once you waited, now you dwell, observe, connect. Welcome to a world where the grid feeds your car—and your soul.

View the story

Posthuman is a daily futurescape newsletter decoding how humans, AIs, and robots are beginning to co-create society. We explore the strange realities of tomorrow as they unfold today—from AI therapists and robot coworkers to memory markets and synthetic influencers. Each edition unpacks one uncanny signal at a time, tracing the slow merge into human–AI symbiosis. This isn’t sci-fi—it’s the early user manual for the future.


TOGETHER WITH US

AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, boasting over 1 million readers from leading companies such as OpenAI, Google, Meta, and Microsoft. Our Newsletter Brands:

We've helped promote over 500 Tech Brands. Will yours be the next?

Email our co-founder Mark directly at [email protected] if the button fails.


Latest Daily Rundowns
🛎️ Replit AI Did A Boo-Boo
More Stories: Altman’s Threat Pitch, A16Z’s New AI predictions
🛎️ OpenAI Wins Google’s Math
More Stories: MIT’s Robots Self-Aware, ChatGPT Hits 2.5B Daily
🛎️ Cursor Eats Koala
More Stories: ChatGPT Mall, Musk’s Baby Grok
More AI Stories
Windsurf’s Weekend Wipeout: Anatomy of a Silicon Valley Collapse
In the hothouse of Silicon Valley’s AI boom, few stories soared as high—or crashed as spectacularly—as Windsurf’s. Born in the crucible of Y Combinator’s 2022 batch, Windsurf quickly became the darling of developers and enterprise CTOs alike. Its core product: a suite of AI-powered programming
Cursor’s Unorthodox Path: How Michael Truell and Team Ignored the Playbook and Won Big
In the world of startups, there are two kinds of founders: those who religiously follow the gospel of Silicon Valley, and those who toss the rulebook out the window, set it on fire, and dance around the ashes. Michael Truell, Co-founder and CEO of Cursor, is firmly in the latter
Salesforce’s CRMArena-Pro Benchmark: LLM Agents Struggle to Pass the CRM Test
In the world of Customer Relationship Management (CRM), where businesses rely on seamless interactions with customers, Salesforce’s new benchmark, CRMArena-Pro, has delivered a reality check for Large Language Model (LLM)-based AI agents. A team led by Kung-Hsiang Huang, a Salesforce AI researcher, has revealed that these AI agents are,