What do AIs dream of when they close their eyes?

👋 Dear Dancing Queens and Super Troupers,

 This is it. The countdown has begun. Unless something unexpected happens, GPT‑5 is dropping this week. And even OpenAI’s CEO Sam Altman is starting to sweat.
He’s tested it. He felt dumb. And he’s warning us: “Get ready for bugs and outages, because everyone’s going to pile on.”
Somehow it’s reassuring to know that the man who compares his own model to an atomic bomb keeps pressing the button anyway. All good, then.

But GPT‑5 isn’t coming alone: it brings supercharged reasoning powers, mini and nano versions for embedded use, and above all, it unifies everything OpenAI has built so far.
It’s faster, smarter, and… free for the whole world (in theory). A tech revolution, yes — but also a serious political and social Russian roulette.

Meanwhile, as this artificial brain prepares to answer your emails, code your apps, and pick your vacation pics, one expert seriously suggests making it president.
By 2032. He talks about AI governance, algorithmic transparency, and augmented leadership. You laugh? He doesn’t. And honestly, looking at some governments today… can we really say it’s such a bad idea?

On the other side of Silicon Valley, Nvidia is promising a gold rush of new millionaires, thanks to AI — even more than during the internet boom.
The code is simple, the tools are here. Even a high schooler in Oklahoma got it: he handed $100 to ChatGPT, and his trading bot made +23% in just four weeks!

Maybe that explains Apple’s sudden rush to catch up — with its own AI chatbot focused on web search?

 Anyway, here’s what’s on the radar this week:

👉️ GPT‑5 is coming — and even its creators are panicking 👏​​​​​​

👉️ An AI in the Elysée Palace by 2032? 🥴

👉️ Nvidia declares the next gold rush 💥​

👉️  This teen just beat Wall Street with a trading bot 👀

👉️ Siri might finally get smart (for real this time) 📱​​​

If we've forwarded this letter to you, subscribe by clicking on this link !

Fact-based news without bias awaits. Make 1440 your choice today.

Overwhelmed by biased news? Cut through the clutter and get straight facts with your daily 1440 digest. From politics to sports, join millions who start their day informed.

Sign up now!

🔥 If you have 15 minutes

1️⃣ GPT‑5 is coming — and even its creators are panicking

The summary: OpenAI is preparing anything but a quiet launch: GPT‑5 is landing this summer, and Sam Altman is already warning the curious — this is going to be intense. Three formats (Flagship, Mini, and Nano), a redesigned architecture to simplify the user experience, and reasoning skills on steroids. It might blow your mind — and maybe the servers too.

Details :

• A neural trilogy: GPT‑5 will be released in three formats — full Flagship for heavy-duty use, Mini for lighter tasks, and Nano for embedded applications via API. Three sizes, one shared intelligence — think of it as an AI Matryoshka.

• Merged under the hood: Say goodbye to the split between GPT‑4 and o3 — everything is now unified. A simpler yet more powerful AI, capable of cold reasoning across complex contexts, with memory that doesn’t crash after ten messages.

• Expanded memory, sharper brain: GPT‑5 barely forgets anything. Its contextual memory has been upgraded like a cognitive exoskeleton, allowing long conversations without repetition or drift.

• Altman keeps it real: No AGI just yet — but we’re getting closer, he says. He anticipates a launch phase that’s “chaotic but amazing,” with probable server overloads, minor bugs… before it all runs like a quantum clock.

• An AI for the public good: With no party, ego, or personal ambition, AI could become a neutral referee in public decision-making. Built to prioritize the greater good, it could offer leaders a rational mirror — immune to political or emotional pressure. A kind of algorithmic wisdom, able to raise red flags when logic breaks down.

Why it's important: With GPT‑5, OpenAI is hitting a strategic milestone: streamlining its product line, delivering a snappier AI, and trying to silence the competition through innovation. Claude, Gemini, Grok? They might start coughing.

GPT‑5 is like going from Google Docs to a cosmic co-pilot — one that understands, codes, writes, predicts, and follows up. And it works at a speed that, according to Altman, is “kind of scary.”

2️⃣ An AI in the Élysée Palace by 2032?

The summary: An AI expert predicts that artificial intelligence could join forces with human leaders as early as 2032 to help guide public policy. This scenario — dubbed “AICracy” — envisions an AI designed to assist leaders by analyzing laws, forecasting policy impacts, distributing resources fairly, and tracking disinformation in real time. Not a digital head of state, but a well-calibrated, ultra-rational co-pilot.

Details :

• Presidential-grade software:
  After decades designing systems for critical environments, the expert argues that modern governments are becoming too vulnerable to bias and irrational decisions. By 2032, he believes AI could become the most reliable and impartial tool in the hands of power.

• AICracy: augmented governance:
  Humans stay in charge, but the AI acts as a hyper-efficient co-pilot. It would spot legal inconsistencies, simulate policy outcomes, guide equitable resource allocation, and detect real-time manipulation through misinformation.

• Built-in limits and caution:
  AI feels no ambition, fear, or ego — but it entirely depends on the ethics, transparency, and datasets embedded by its creators. It enhances clarity, but does not invent it. Leaders remain the moral compass and value-setters.

• An AI for the common good:
  With no party lines or personal ambitions, AI could become a neutral referee for public decision-making. Designed to prioritize collective interest, it would offer leaders a rational mirror — free from emotional or political pressure. A form of algorithmic wisdom, capable of issuing warnings when logic falters.

Why it's important: With trust in leadership crumbling and data complexity exploding, the idea of AICracy feels more and more plausible. By 2032, the question may no longer be “Can AI govern?” — but “Why are we still relying solely on blindly human leadership?”

Picture a digital assistant at every meeting, tracking every chart, spotting flaws and trends — never losing focus. That’s the AI-conductor: emotionless in execution, precise in orchestration

3️⃣ Nvidia Announces the New Gold Rush

The summary: Jensen Huang, co-founder and CEO of Nvidia, makes a bold prediction: artificial intelligence will create more millionaires between 2025 and 2030 than the Internet did between 1995 and 2015. The claim, published on August 3, 2025, on Atlantico, comes from Philippe Bourcier and Don Diego De La Vega.

Details :

• AI = More Millionaires in 5 Years:
  According to Huang, AI will generate more millionaires in five years (2025–2030) than the internet did over two decades. He calls AI “the great technological equalizer,” breaking down barriers to innovation.

• “Everyone’s a programmer now”:
  Today, you don’t need to know Python or C++. All you need is to talk to an AI. Anyone — artist or engineer — can now build a product or start a company.

• The Two-Factory Model:
  Huang envisions every company operating two factories: one for the physical product, and one entirely AI-driven — like Tesla already does.

• Small Teams, Massive Impact:
  Tiny teams of 150 researchers can generate $20–30 billion in value. That’s roughly $200 million per person — an unprecedented return.

Why it's important: This vision marks a deep shift: wealth creation is now based on intellectual efficiency, not manpower. When launching ideas becomes accessible to everyone, getting in early could turn employees into billionaire founders.
Without a single physical factory, a lone hacker-cyborg could turn any idea into a profit engine. AI blending creativity, computation, and strategy — the ultimate millionaire-maker, silently working from your pocket.

4️⃣ This Teen Just Beat Wall Street With a Trading Bot

The summary:  Nathan Smith, a high school student from Oklahoma, built a trading bot powered by ChatGPT that delivered a stunning 23.8% return in just four weeks, outperforming the Russell 2000 and even top investors.

Details :

• Blazing Fast Returns:
  In only four weeks, the ChatGPT-powered bot hit +23.8%, beating market pros and major benchmarks like the Russell 2000.

• Lightning-Fast Code Generation:
  With ChatGPT, Nathan didn’t spend weeks coding. The bot was born after just a few prompt sessions — a real-world example of how AI can automate and optimize financial strategies.

• Micro-Cap Mastery:
  The bot focuses on micro-cap stocks — small companies with big growth potential. That’s where it shines, capitalizing on volatility and momentum.

• Not-for-Profit Project:
  Nathan isn’t monetizing his bot. His goal is to push the limits of AI in finance, not get rich. He even shares the results open source, championing an educational and experimental approach to automated trading.

Why it's important: This performance challenges the entire model of human trading. A teenager with a few clever prompts can now compete with — or beat — seasoned Wall Street pros.
The world of algorithmic trading is being reshaped by bots and accessible prompts.
Picture this: a teen hacker plugs ChatGPT into the financial markets, types three prompts, and out comes a trading robot that outperforms Wall Street’s finest.
Now that’s a bot legend — silent, fast, and ready to humble the CAC 40 elite.

5️⃣ Siri Might (Finally) Get Smart

The summary:  Apple is working on a simplified in-house chatbot, designed to rival ChatGPT. According to Engadget, a small internal group named Answers, Knowledge and Information (AKI) is developing a system that can answer general questions using machine learning and web crawling. Apple is entering the mainstream AI assistant race — in its own way.

Details :

•  A discreet but strategic team:
  The project is led by the new AKI division, composed of engineers specializing in information retrieval. Apple doesn’t want to just plug in ChatGPT — it wants to build its own homegrown alternative.

• Stripped-down chatbot, maximum efficiency:
  The bot is said to be “stripped-down” — light interface, quick answers, no fluff. The goal is an assistant that answers directly, no visual clutter, but with the reliability of a well-calibrated search engine.

• Future Siri integration:
  According to Bloomberg’s Mark Gurman, this project could form the foundation of a new, more direct version of Siri — one that doesn’t redirect you to other services or search engines. A launch is reportedly targeted for Spring 2026.

• Privacy guaranteed — the Apple way:
  Apple is banking on its privacy-first approach. The chatbot would run on proprietary models, complying with Apple Intelligence security standards, with no data leaks to third-party services.

Why it's important:  Apple is taking a major step: evolving from Apple Intelligence to a true autonomous conversational agent. While OpenAI, Google Gemini and others dominate, Apple is pushing back with its own streamlined answer engine — no external boost needed.
Imagine a Siri that doesn’t say “Here’s what I found on the web,” but just answers, instantly. A no-drama assistant — minimal, but as efficient as a well-written line of code. It’s Apple’s silent chatbot rebellion: clean, focused, deadly smart.

❤️ Tool of the Week: Showrunner – The AI That Creates Your Own TV Show

Showrunner is an AI platform developed by Fable Studio that lets you create animated series episodes in minutes, from a simple prompt.
Yes — you’re the showrunner… without ever hiring a crew.

What it’s for :

• Create an animated episode in minutes: script, visuals, voice acting, editing — all generated automatically.

• Insert yourself into the story: send a selfie, and you become a character in your own series.

• Parody or reinvent universes: spoof South Park, build a series about your startup, or reimagine Black Mirror your way.

• Share and monetize: Showrunner is partnering with studios so creators can eventually monetize their content.

How to use it?
 Go to fablestudio.com/showrunner (or showrunner.show) to create your first episode for free. Premium subscriptions start at $10/month!