RSS Amplifier

AI Operator · Jul 8, 2025

#15: Build an AI "Dream Team"

0
Sign in to vote or save

This page did not load. You can still read it on the original site — the toolbar below keeps your place in the directory.

Hey, it’s Marc.

Hey, it’s Marc.

Meta just launched a new AGI unit, “Superintelligence Labs” led by Scale AI’s founder Alexandr Wang. They’re also scooping up top talent from OpenAI, Google, and Anthropic. [NEWS]

Subscribe now

Top Agent signals this week:

  • Visa and Crossmint team up to enable secure AI-driven automated payments

  • Genspark launched Super Agent, automating real tasks using OpenAI APIs

  • Dataiku partners with HPE to help enterprises launch governed AI agents

  • Supervity replaces legacy RPA with AI agents that cut costs by 50%

  • Cursor lets developers assign and track AI coding agents via web

👉 Get your brand in front of 30,000+ decision-makers — book your ad spot now.

Top Boardroom Reads This Week


Power Moves

Oracle’s $30B Cloud Deal Changes the AI Infra Map

Oracle quietly signed a cloud deal worth $30B/year — starting FY28. Revealed via SEC filing, this is one of the biggest annual enterprise cloud wins ever. With multicloud database rev growing 100% YoY and big AI partnerships (OpenAI, IBM, SoftBank), Oracle is becoming a surprise leader in AI infra. [NEWS]

So what? This deal makes Oracle a key player in the AI race. Instead of chasing Amazon and Microsoft, they’re becoming the go-to supplier for top AI companies. It shows the real battle in AI is about who can deliver the best infrastructure—and Oracle just proved it belongs at the top.

Super Agent Hits $36M ARR in Just 45 Days

Genspark launched "Super Agent," a no-code AI workspace that reached $36M in annual recurring revenue in just 45 days — with a 20-person team and $0 ad spend. Built on OpenAI's GPT-4.1 and Realtime API, Super Agent automates real-world tasks like making calls, designing pitch decks, and generating short videos. Users describe a task in natural language, and the agent executes it across tools using 9 LLMs and 80+ integrations. [ANNOUNCEMENT]

So what? This is the new enterprise OS. Internal workflows, sales ops, content generation, and even outbound calls can be abstracted away via AI agents with a no-code solution.


Vendor Spotlight: Sakana AI

Sakana AI has introduced a new system called AB-MCTS (Adaptive Branching Monte Carlo Tree Search) that proves for the hardest problems, a team of cooperating AIs is far more powerful than any single model. [ANNOUNCEMENT]

Why it matters: Siloed models have limitations. But by letting them cooperate in real time, Sakana’s system beats individual models by up to 30% on one of the toughest benchmarks (ARC-AGI-2). It provides a blueprint for solving high-value business challenges that are too complex for today's best standalone AIs.

What is it? The system is a "project manager" that directs a team of different AI models (like those from Google, OpenAI, etc.). Instead of relying on one AI's "opinion," it gets them to work together, leveraging each one's unique strengths.

The shift: Inference-time scaling (aka "thinking harder during runtime") is quickly becoming the new model scaling. Sakana’s approach combines:

  • Trial-and-error reasoning (like debugging code)

  • Dynamic LLM selection based on task type

  • Tree search logic (like AlphaGo) for exploration

What they proved: No single model solved the task. But using o4-mini’s flawed output as a clue, DeepSeek and Gemini found the right answer.
→ Real collective intelligence, not ensemble voting.

How Corporates Can Use Collective AI Intelligence

  1. R&D Acceleration: Use Multi-LLM AB-MCTS to boost complex problem-solving where one model hits a wall — e.g. materials discovery, biotech modeling, chip design.

  2. Autonomous Agents for Ops & Strategy: Instead of relying on a single LLM agent (which may be biased or brittle), deploy multi-model reasoning agents to make high-stakes decisions in:

    • Supply chain optimisations

    • Financial forecasting

    • M&A target analysis

  3. AI for Code & Product Engineering: Enhance AI pair programmers or internal copilots by orchestrating different LLMs:

    • One model for architecture

    • One for code quality

    • One for security scanning

And, the list of use cases is exhaustive.

More on Sakana AI:


Top AI Vendors This Week

👉 We’re tracking 1000s of AI and blockchain vendors so you find the right vendors, partners, and bet on the right tech. Sign up for early access.

Subscribe now


More from 51:

That’s all for today.

Thanks,

Marc & Team


Read on 51ai.substack.com

Comments

Nothing yet. Say the first thing.

    Sign in to join the conversation.