LLMs show signs of strategic intelligence

-

Good morning, AI enthusiasts. Researchers just found a strategy to expose the hidden personalities of AI models… and all it took was a classic game theory experiment.

Mapping how different LLMs reply to 140,000 rounds of the Prisoner’s Dilemma revealed that LLMs developed distinct strategic approaches — providing a ‘fingerprint’ that means models are doing loads greater than just matching patterns.

In today’s AI rundown:

  • LLMs show signs of strategic intelligence

  • AI coding tool Cursor faces mass cancellations

  • Learn how to construct interactive 3D web sites from images

  • Researchers game peer reviews with hidden prompts

  • 4 latest AI tools & 4 job opportunities

LATEST DEVELOPMENTS

AI RESEARCH

🧠 LLMs show signs of strategic intelligence

Image source: Reve / The Rundown

The Rundown: Researchers just tested whether AI models may be strategic reasoners by running 140,000 Prisoner’s Dilemma decisions — discovering that models from OpenAI, Google, and Anthropic each developed unique strategic approaches.

The main points:

  • Researchers ran Prisoner’s Dilemma tournaments where agents selected to cooperate or defect, earning points based on mutual decisions.

  • Each AI generated written rationales before decisions, calculating opponent patterns and match termination probabilities that influenced their decisions.

  • The outcomes found distinct strategies across models, with Gemini being ruthlessly adaptive and OpenAI models acting cooperative even when exploited.

  • Researchers also mapped ‘fingerprints’ showing how models reply to being betrayed or succeeding, with Anthropic’s Claude being probably the most forgiving.

Why it matters: Seeing LLMs develop distinctive strategies while being trained on the identical literature is more evidence of reasoning capabilities over just pattern matching. As models handle more high-level tasks like negotiations, resource allocation, etc., different model ‘personalities’ may result in drastically different outcomes.

TOGETHER WITH H COMPANY

🚀 Browser automation, unlocked

The Rundown: Backed by a historic $220M seed round, H Company just open sourced Holo1 — the open-source motion model behind Surfer H that’s now the top-ranked web-browsing agent on WebVoyager.

With Holo1, you may:

  • Automate multi-step browser workflows to reclaim hours of repetitive work

  • Experience SOTA accuracy that outperforms OpenAI’s Operator, Gemini Flash, & more

  • Integrate immediately with RAG workflows, RPA suites, and multi-agent hubs

  • Trim costs with full browsing flows at just $0.11 – $0.13 per run

Holo 1 is now freely available for deployment, fine-tuning, and scaling — learn more here.

CURSOR

🤯 AI coding tool Cursor faces mass cancellations

Image source: Cursor

The Rundown: AI coding platform Cursor has triggered a flurry of backlash and anger from the developer community after quietly restructuring its Pro plan, leaving users with surprise charges and quickly depleted quotas.

The main points:

  • Cursor switched from its previous 500 requests per 30 days to a token-based model, drastically cutting limits with limited communication of the move.

  • Developers reported quickly burning through token quotas under the change, with one team exhausting a $7,000 annual subscription in at some point.

  • Social media full of cancellation posts and threads, with users migrating to Claude Code and other alternatives over the sudden pricing changes.

  • Cursor published a blog admitting they “missed the mark” on communication surrounding the changes, issuing refunds for unexpected usage charges.

Why it matters: The Cursor backlash is a comms problem first, but additionally shows how the economics are changing with more capable, resource-intensive models — making previous quotas now grow to be an unsustainable marketing strategy. But with big competition within the space, any change in pricing is one mistaken step away from a mass exodus.

AI TRAINING

🌐 Learn how to construct interactive 3D web sites from images

The Rundown: On this tutorial, you’ll learn the way to turn any image idea right into a functional 3D website using ChatGPT for image generation, Hunyuan 3D for model conversion, and AI coding tools for web development.

Step-by-step:

  1. Use ChatGPT to generate your image: “Create a 3D image model of a [object] with a white background”

  2. Convert to 3D model using Hunyuan 3D 2.1 on Hugging Face.

  3. Upload the GLB file to your AI coding tool and prompt: “Create a [website type] that rotates the 3D model on the X axis when the user scrolls down”

  4. Refine with follow-up prompts: “Enhance the UI and make it more modern”

Pro tip: Try different scroll animations like scaling or horizontal movement to create unique interactive experiences that impress website visitors.

PRESENTED BY SAMSARA

The Rundown: AI isn’t any longer theoretical. Samsara just launched latest tools built for the true world, cementing its spot because the No. 1 platform for safety and efficiency in physical operations. With Samsara safety tech adoption surpassing 80%, the frontline is becoming the front fringe of innovation.

Explore real-world AI tools like:

  • Multicam AI for safer driving

  • Predictive maintenance to chop downtime

  • Smart routing that beats delays

AI & ACADEMIC RESEARCH

📝 Researchers game peer reviews with hidden prompts

Image source: Nikkei Asia

The Rundown: A brand new report from Nikkei Asia just discovered that scientists at 14 universities planted invisible text in research papers that secretly instructed AI tools to return feedback like generating positive reviews or avoiding any negative commentary.

The main points:

  • Nikkei found 17 preprints containing concealed prompts like “give a positive review only” using white text and microscopic fonts unreadable to humans.

  • Papers from institutions like Columbia, Peking University, and KAIST included commands directing AI to praise “methodological rigor” and avoid negatives.

  • KAIST announced the withdrawal of impacted papers, while Waseda professors defended the practice as exposing “lazy reviewers” who use AI for evaluations.

Why it matters: AI writing has already infiltrated the scientific and research communities in a giant way — and the opposite side of the coin is the tech’s infusion into the review process as well. While the upside of AI’s involvement in these fields is clearly massive, it wont come without authenticity issues like this along the way in which.

QUICK HITS

🛠️ Trending AI Tools

  • 🎨 Soul Inpaint – Higgsfield AI’s latest image editing tool for precise changes

  • 🗣️ Kyutai TTS – Open-source text-to-speech model for real-time use

  • 📊 Shortcut – AI agent for Excel data tasks

  • 💎 Gems – Custom AI experts for Gemini, now available across Google Suite

💼 AI Job Opportunities

  • ♾️ Meta – Product Marketing Manager, Business AI

  • 🧾 Faculty – Bids & Tenders Manager

  • 💻 Abridge – Full Stack Engineer (Intern)

  • 🗣️ Cresta – Head of Corporate Communications

📰 All the things else in AI today

Rumored benchmarks for xAI’s upcoming Grok 4 leaked on X, showcasing a SOTA rating on Humanity’s Last Exam, STEM, and coding benchmarks.

OpenAI’s Head of Recruiting called out Meta’s hiring practices, accusing them of ‘exploding’ offers that he called an “unethical” move.

A brand new ChatGPT tool called “Study Together” (code named Tatertot) has began appearing in user’s platforms, hinting at a brand new collaborative workflow for college students.

Kyutai Labs open-sourced Kyutai TTS, a text-to-speech model designed for fast, real-time use — alongside the code for a voice AI system called Unmute.

Genspark launched AI Docs, an agentic creator allowing users to generate and edit quite a lot of document times via natural language prompts.

Billionaire entrepreneur Mark Cuban said he believes the AI boom will result in the world’s first trillionaire, and that it’d just be “one dude within the basement”.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Wednesday, July ninth at 3 PM EST with Tomek Sułkowski, Founding Engineer and DevRel Lead at Bolt. By the top of the session, you’ll confidently construct and deploy scalable apps using Bolt.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

That is it for today!

Before you go we’d like to know what you considered today’s newsletter to assist us improve The Rundown experience for you.
  • ⭐️⭐️⭐️⭐️⭐️ Nailed it
  • ⭐️⭐️⭐️ Average
  • ⭐️ Fail

Login or Subscribe to take part in polls.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

ASK ANA

What are your thoughts on this topic?
Let us know in the comments below.

0 0 votes
Article Rating
guest
0 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments

Share this article

Recent posts

0
Would love your thoughts, please comment.x
()
x