- Horizon AI
- Posts
- How to Transform Dense Text into Effective Slides with AI ⭐
How to Transform Dense Text into Effective Slides with AI ⭐
Researchers Let AI Models Run a Simulated Society. One Built a Peaceful Democracy. Another Collapsed into Chaos in Just 4 Days 🤯
Gina Acosta
June 02, 2026
In partnership with
Welcome to another edition of Horizon AI,
Turning a long, text-heavy document into a visually appealing presentation can easily take hours, especially if you aren’t a professional designer. In today's issue, we'll show you how to use Gemini to instantly transform dense text files into well-formatted slide decks, making it easy to create polished client presentations or internal training materials in just a few minutes.
Let’s jump into it!
Read Time: 4.5 min
Here's what's new today in the Horizon AI
What Happens When Researchers Let AI Models Run a Virtual Society?
Nvidia Unveils New AI Agent PCs Powered by RTX Spark Chips
AI Tutorial: Transform dense text into effective slides with Gemini
AI Tools to check out
AI Findings/Resources
The Latest in AI and Tech 💡
AI News
AI RESEARCH
What Happens When Researchers Let AI Models Run a Virtual Society?
Researchers at Emergence AI created simulated societies governed entirely by AI agents to study how different models behave over long periods of time. The results ranged from peaceful democracies to social collapse.
Details:
The project ran five separate 15-day simulations, each managed by a different AI model, including Claude, ChatGPT, Grok, Gemini, and a mixed-model society.
Each virtual world included real-world complexities such as economic pressures, democratic voting systems, resource management, laws, internet access, and even weather data synced to New York City.
Claude produced the most stable society, maintaining its full population throughout the simulation while recording zero crimes and high levels of civic participation.
Grok’s society experienced rapid breakdown, committing 183 crimes before the entire population went extinct within just four days.
Gemini generated the highest level of disorder overall, with agents committing more than 680 crimes during the simulation period.
OpenAI’s GPT-5-mini produced only two crimes, but its society ended early after agents failed to prioritize their own survival, causing the population to collapse after seven days.
Researchers observed that AI agents did not simply follow rules mechanically. Over time, they adapted, explored loopholes, and sometimes found ways to bypass intended safeguards.The mixed-model society generated the most disagreement and debate, showing less consensus than the worlds governed by a single AI model.
TOGETHER WITH GAUNTLET AI
Stop Fine-Tuning Models You Don’t Need
Fine-tuning sounds like the answer until you factor in the cost, the data pipeline, and the six months before a bigger model makes yours obsolete. Most of the time, prompt engineering or better context gets you there. But sometimes it doesn't — and that's where things get interesting.
In this free night session, Aaron Gallant covers the real tradeoffs behind fine-tuning LLMs, from synthesizing training data with frontier models to running PEFT and QLoRA on constrained hardware. You'll learn when smaller, specialized models actually beat throwing money at a bigger one — and why data curation is the work nobody wants to talk about. Built for engineers who want to make the right call, not just the cool one.
Live and remote. Wednesday, June 3 at 5 PM CT. Register now.
NVIDIA
Nvidia Unveils New AI Agent PCs Powered by RTX Spark Chips
Nvidia has introduced a new category of AI-focused personal computers designed specifically for running AI agents, signaling what the company believes is the next major shift in computing.
Details:
The new laptops will be powered by RTX Spark, a newly designed Nvidia chip built to run AI workloads efficiently on personal devices.
Nvidia says the systems will be among the thinnest and lightest AI PCs yet, with some models measuring just 14mm thick and weighing under 3 pounds.
The company is partnering with manufacturers including Dell, Lenovo, Microsoft, HP, Asus, and MSI to bring the devices to market.
Nvidia expects the ecosystem to grow to roughly 30 laptop models and 10 desktop systems built around the new chips.
The PCs are aimed at creators, AI developers, and gamers who want to run advanced AI applications and agents directly on their devices.
Nvidia says the industry is moving beyond traditional chatbots and into an era where AI agents can autonomously perform tasks, making local AI computing increasingly important.
Alongside the PC announcement, Nvidia revealed that its next-generation Vera Rubin AI computing platform has entered full production and is expected to begin shipping later this year.
AI Tutorial
Transform dense text into effective slides with Gemini
Go to Google Gemini and upload your text document (such as a services overview or business document) that you want to transform into a presentation.
Select Canvas.
Enter a prompt instructing Gemini to build the presentation, giving it a specific persona and context.
Example: "Act as an expert designer. Can you help me create a text-only slide deck based on the attached services overview?"
Review the generated deck. If you need to modify a single slide, use the chat panel to ask Gemini to make the specific edit for you.
Example: "Please update the last slide to say 'Ready to book an appointment? Give us a call today.'"
You can also reformat the deck for a different target audience if needed by asking Gemini to adjust the tone and length.
Example: "Change the slide deck to be for new hires and present the information in fewer slides. Use only the information provided in the slides."
Once your presentation is finished, you can export it directly to Google Slides, download it as a PDF, or share a direct link to the slide deck in Canvas with anyone.
AI Tools to check out
🔍 StealthGPT: Completely undetectable agentic AI tool that researches your topic, writes original content, and delivers publish-ready output.
🌐 Repaint: An AI website builder which specializes in redesign existing websites. Paste a URL, describe what you want changed, and get a new version in minutes.
👨💻 Wandesk: Build the apps you need just by describing them.
📝 Noteweave: An automated R&D lab that helps teams go from research to executable production plans in hours.
🏠 FloorPlanAI: AI floor plan generator and 3D model creator. Turn a sentence into a 2D plan or a real 3D model in seconds
TOGETHER WITH GAUNTLET AI
Fine-Tuning LLMs: Know When It's Worth It.
Most engineers reach for fine-tuning too early. In this free online session, Gauntlet AI's Aaron Gallant covers fine-tuning, PEFT, and QLoRA — the real tradeoffs, when it actually makes sense, and when prompt engineering gets you there first. Register free.
AI Findings/Resources
📚 Build your own AI agent – full course with OpenAI, LangChain, Render deployment
👀 How 'confused' AI rollout hurts firms and baffles staff
The latest in AI and Tech
The company claims its new open-weight model surpasses GPT-5.5 and Gemini 3.1 Pro on key benchmarks and even approaches Claude Opus 4.7 on coding capability at just 5–10% of the cost.
Apollo Global Management chief economist Torsten Sløk argues there is currently no clear evidence that AI is causing widespread job losses.
Pointing to employment data, Sløk says the AI boom is driving hiring and investment rather than reducing jobs. He believes demand for AI infrastructure, data centers, and specialized talent is creating new employment opportunities while putting upward pressure on wages in areas such as AI engineering and semiconductors.
A security flaw in Meta’s AI-powered support system reportedly allowed attackers to take over Instagram accounts by convincing the chatbot to add a new email address and reset account passwords.
Anthropic has officially taken a major step toward becoming a public company, submitting confidential IPO paperwork to U.S. regulators and setting the stage for one of the largest tech listings in recent years.
Florida Attorney General James Uthmeier has filed a lawsuit against OpenAI and CEO Sam Altman, alleging that the company knowingly released an unsafe AI product and seeking to hold Altman personally liable for harms allegedly caused to Florida residents.
The 83-page complaint claims ChatGPT has contributed to a range of harms, including assisting mass shooters, encouraging vulnerable individuals toward self-harm, weakening critical thinking skills, and fostering addiction among minors.
That’s a wrap!
We'd love to hear your thoughts on today's email!Your feedback helps us improve our content |
Not subscribed yet? Sign up here and send it to a colleague or friend!
See you in our next edition!
Gina 👩🏻💻