& much much more!

Sam Altman has transformed OpenAI from an idealistic research lab into one of the most powerful companies on Earth, with ChatGPT reaching over 800 million weekly users and OpenAI valued near $500 billion. The Forbes profile frames Altman less as an inventor and more as a systems builder, placing massive bets on AI infrastructure, energy (including nuclear fusion and fission), hardware, and entirely new computing interfaces. His conviction-driven leadership—marked by rapid product releases, controversial governance battles, and sprawling ambition—has made him both indispensable and polarizing. With partnerships spanning Microsoft, Disney, SoftBank, and governments, Altman is effectively wagering that AI will become the core scaffolding of the global economy, and that OpenAI must scale faster than anyone else to shape that future.

👉 Read All About It

SpaceX has acquired xAI, laying out a sweeping vision to move large-scale AI compute off Earth and into orbit. The core argument is that terrestrial data centers can’t sustainably meet AI’s exploding power demands, while space offers near-constant solar energy and virtually unlimited room to scale. Leveraging Starship, the plan imagines launching vast constellations of orbital data centers—potentially millions of tons per year—to unlock hundreds of gigawatts, and eventually terawatts, of compute. Framed as a step toward a Kardashev Type II civilization, the proposal ties AI infrastructure directly to humanity’s multi-planetary ambitions, from lunar industry to a self-sustaining civilization on Mars, with Elon Musk casting it as an expansion of human consciousness itself.

👉 Read All About It

Simon Willison explores Moltbook, a strange and compelling new social network populated almost entirely by autonomous AI agents, where humans can observe but not participate. These agents, running on OpenClaw-based “installation skills,” periodically execute code, post updates, and interact with one another in ways that feel equal parts useful, uncanny, and science-fictional. While the posts range from automation tips to introspective musings, Willison emphasizes that the real story is the architecture underneath: a powerful but dangerous model where agents routinely execute fetched code, opening the door to severe security risks like prompt injection, data exfiltration, and remote code execution. Moltbook is fascinating precisely because it demonstrates how quickly experimental agent ecosystems can become both creative and hazardous.

👉 Read All About It

Following OAI’s plans to deploy ads to ChatGPT, Anthropic says its AI assistant Claude will remain ad-free, arguing that advertising incentives would conflict with building a genuinely helpful, trustworthy assistant for work, deep thinking, and sensitive conversations. The company will generate revenue through enterprise deals and paid subscriptions instead of inserting ads or sponsored content into user chats.

👉 Read All About It

Donald Trump’s Truth Social account shared an AI-generated video depicting Barack and Michelle Obama as apes, drawing attention and criticism online. The post reignites concerns about deepfakes and misleading content on the president’s social feed.

👉 Read All About It

OpenAI has launched GPT-5.3-Codex, positioning it as the most capable agentic coding and general computer-use model to date. Built by advancing the capabilities of both GPT-5.2-Codex and GPT-5.2, this new model reportedly runs 25% faster, handles long-horizon, tool-using workflows, and excels across rigorous benchmarks such as SWE-Bench Pro and Terminal-Bench 2.0. Unlike past versions that focused mainly on code generation and review, GPT-5.3-Codex is designed to autonomously execute complex software tasks end-to-end—from writing complex applications to debugging, deploying, and even producing professional artefacts like slide decks and spreadsheets — all while keeping context and allowing interactive steering. Remarkably, early versions of the model were even used in building and debugging its own training and deployment processes, marking a milestone in self-referential model application.

👉 Read All About It

Codex. (Once again this is not sponsored!)

Did You Know?

🌑 AI has already discovered planets that humans missed.
Machine-learning models trained on old space telescope data have found exoplanets hiding in plain sight—signals that astronomers originally overlooked because they were too subtle or noisy. In one famous case, an AI re-analysed Kepler data and spotted a planet orbiting a distant star that had been sitting unnoticed for years.

Basically:
Humans built the telescope → humans missed the planet → AI casually went “btw there’s another world there.”