• AI Fire
  • Posts
  • 🌊 China’s AI is Floating in the Ocean

🌊 China’s AI is Floating in the Ocean

Insane & Free World's First AI Design Agent

ai-fire-banner

Read time: 5 minutes

Claude’s gunning for Bloomberg, and Google’s AI just stopped a cyberattack in the wild. Oh, and China is literally putting AI data centers in the ocean. Scary-fast!

LEARNING PARTNER AIRCAMPUS

Master Building AI Agents 👨‍🏭👩‍🏭

No coding. No tech degree. No problem.
On 19th July at 10 AM EST, something different is happening.

Imagine this:
Someone with zero coding skills builds an AI agent that replies to emails, schedules meetings, posts on social media, and runs 40+ apps in perfect sync —
All in one morning.

Not a fantasy. Not a Silicon Valley story.
Just a Masterclass that shows how anyone — yes, anyone — can build smart systems that automate the boring stuff and unlock hours of freedom.

AI agents aren’t the future.
They’re the quiet edge behind every productive creator today.
This is the blueprint.
🎟️ Reserve your seat before it fills.

AI INSIGHTS

claude-vs-bloomberg-ai-just-put-the-25k-terminal-on-watch

Just when financial analysts thought their toolset couldn’t get any more crowded, Anthropic dropped Claude for Financial Services - a serious shot at the aging Bloomberg Terminal.

What makes it different?

  • Claude pulls together external market data (S&P, FactSet, Morningstar) and internal firm data (Databricks, Snowflake)

  • Claude Code helps you run custom financial models or simulations without gluing scripts together at 2AM.

  • No need to Frankenstein your APIs. Claude ships with MCP connectors that work out of the box.

  • Upgraded token + query limits let analysts crunch full 10-K filings or M&A docs without hitting token walls.

Claude’s no slouch on benchmarks:

  • Finance Agent Benchmark: 44.5% accuracy on real analyst tasks. (That’s neck-and-neck with OpenAI’s best; most other models scored under 30%.)

  • Enterprise Impact:

    • Bridgewater: +20% productivity

    • Norway’s sovereign fund: Saved 213K hours

    • AIG: Underwriting review time cut by 5x

These aren't simple “What's Apple's stock price?” questions – we're talking multi-step SEC filing analysis that would make a first-year analyst sweat.

Why It Matters: Besides, Perplexity’s quietly become a powerful alt-tool for analysts, with offerings that feel more like a Bloomberg killer for everyone else. CEO Brian Armstrong even calls it a “10x unlock” for AI-finance integration.

PRESENTED BY JIRA

56% of workers say scheduling a meeting is the only way to get information. With Jira, use AI to automatically add work from Slack, create subtasks, or attach relevant resources. So instead of scheduling a meeting, check the status in Jira. Easy.

TODAY IN AI

AI HIGHLIGHTS

💀 A new report found that 75% companies in the S&P 500 list AI as a risk, covering everything from ethical concerns, operational disruptions, regulatory scrutiny, security.

🖥️ Anthropic’s Thariq Shihipar thinks Claude Code is the best place to build a general agent to automate your work tasks and daily routines. Here’s the exact setup he uses.

🎉 Runway released a new AI motion capture model that can track a subject’s face, hands, even body to quickly transport them into any scene. Watch it in action here.

⚠ Another ChatGPT outage started early July 16, affecting users globally about login loops, blank screens, and lost sessions. Here’s what OpenAI said about this meltdown.

📱 Google’s new business-calling agent will dial up local shops, ask about availability and pricing, and report back to you without you making a call. Good for Gen Zers!?

🌊 China is putting AI data centers in the ocean to cool them naturally using water. The first phase includes 198 servers with $223M, enough to train GPT-3.5 in a day.

💰 AI Daily Fundraising: AWS is investing another $100 million into its Gen AI Innovation Center to help more businesses build real AI agents, after already supporting thousands of companies worldwide.

AI SOURCES FROM AI FIRE

🎉 Today, let's shout out to our legends of June 2025! You’ve just unlocked a 3-month pass to AI Fire Academy → yep, that means full access to 500+ workflows, tutorials, and AI deep dives:

  • eku***@*****vus.edu - adam****[email protected]
    - lgreen@******.edu - calcant****@*****.edu.pe

NEW EMPOWERED AI TOOLS

  1. 🕵️‍♂️ Webhound automatically finds, extracts, and organizes your web data

  2. 🎥 Glyp create hundreds of on-brand influencer videos to scale your traffic

  3. 🧪 AI Playground has all the latest AI video tools in a single place to try

  4. 🌐 Pagy 2.0 builds a website in minutes, for free, and on your own domain

AI QUICK HITS

  1. 🖼️ Here’s a hidden Veo trick to drop real-life objects into your AI scenes

  2. 📉 Scale AI did lay off 14% of its staff, largely in data-labeling business

  3. ⚠️ OpenAI & Anthropic researchers decry ‘reckless’ safety culture at xAI

  4. 🧠 Amazon’s AI agent takes on complex tasks for you to do creative work

  5. 🎯 Meta reportedly scores two more high-profile OpenAI researchers

AI CHART

ai-progress-might-be-moving-much-faster-than-expected

2025 was supposed to be the year of AI agents. But instead of just automating repetitive tasks, we’re now seeing AI make real-world discoveries and catch threats in real time, hinting that we may have skipped a stage in the evolution of intelligent systems.

Key Examples of What Just Happened:

  • Google’s Big Sleep Caught a Live Cyberattack: An AI agent developed by Google spotted and stopped a software exploit before it was used in the wild.

    → Google claims this is the first real-world AI intercept of an active cybersecurity threat.

  • A Self-Driving Chemistry Lab Is Now Live: North Carolina State University launched an AI-powered lab that runs its own chemical experiments.

    • It collects 10x more data than human researchers

    • Could cut material discovery time from years to days

According to METR, AI is now doubling the length and complexity of tasks it can handle every ~4.5 months across 9 benchmarks. AI is getting better at:

  • Robotics

  • Self-driving

  • Mathematics

  • Scientific reasoning

Just 6 months ago, task-handling capacity doubled every 7 months. Now? It’s accelerating even faster. Are we entering post-agentic AI?

→ In labs, in software, and in cybersecurity, “research years” might collapse into “AI weeks.”

We read your emails, comments, and poll replies daily

How would you rate today’s newsletter?

Your feedback helps us create the best newsletter possible

Login or Subscribe to participate in polls.

Hit reply and say Hello – we'd love to hear from you!
Like what you're reading? Forward it to friends, and they can sign up here.

Cheers,
The AI Fire Team

Reply

or to participate.