Stargate is the Manhattan Project of our generation

Hi Futurist,

AI is wasting no time making waves in 2025. From OpenAI unveiling its first AI agent designed for non-techies to China keeping pace with the US in the AI race, the year has barely begun, and the announcements are already reshaping the landscape. And with Trump giving Big Tech the green light to innovate without restraint, the stakes for global AI leadership have skyrocketed.

I get it—some of this may seem far off from your day-to-day work. But make no mistake: understanding these developments now is critical to staying ahead in the next two years. Whether you’re navigating the rise of autonomous AI agents or witnessing the emergence of new AI ecosystems like Project Stargate, these shifts will impact your career sooner than you think. Therefore, I am sending you insights, inspiration, and innovation straight to your inbox. Let’s dive into the depths of the digital future together and discover the waves of change shaping our industry.

💡 In this post, we're dipping in:

• 📣 Byte-Sized Breakthroughs: Trump’s return to office has sparked an AI revolution with Project Stargate. OpenAI launched its first AI Agent, while China has nearly closed the gap in the global AI race.

• 🤖 Digital Toolbox: Composio connects AI agents to 250+ tools for seamless integration. Taskade’s Loops feature automates repetitive tasks with ease. LangChain’s Agent Inbox UI enhances human-AI collaboration for dynamic workflows.

• 🧐 In Case You Missed It: Perplexity drops NFL widgets, LumaLabs launches Ray2 for lifelike video generation, and DeepSeek introduces a free, ad-free AI assistant. Meanwhile, Google embeds Gemini AI into Workspace, and Meta plans AR glasses to rival Apple.

Do you have tips, feedback or ideas? Or just want to give your opinion? Feel free to share it at the bottom of this dip. That's what I'm looking for.

No time to read? Listen to this episode of Digital Dips on Spotify and stay updated while you’re on the move. The link to the podcast is only available to subscribers. If you haven’t subscribe already, I recommend to do so.

Quick highlights of the latest technological developments.

Headstory: Stargate is the Manhattan Project of our generation

Donald Trump may be the first AGI president, but he is positioning himself to be remembered as the ASI president too. On his first day back in the Oval Office, he revoked Executive Order 14110—the cornerstone of Biden-era AI regulation—and threw the doors wide open for the tech industry. The message? Innovate at will. America’s future is in your hands. Now it’s clear why Big Tech has fully backed Trump.

The tech world didn’t hesitate. Within days, whispers about Project Stargate became headlines. A $500 billion investment in AI infrastructure, a moonshot to secure America’s dominance in AI for the next half-century. This isn’t just a project. It’s a revolution. And the stakes? Global leadership. Economic supremacy. The rewriting of geopolitics.

Why does it matter? Because China was catching up. Fast. DeepSeek (the Chinese AI model), mentioned in previous Dips and highlighted again in this one, was starting to rattle nerves in Silicon Valley. Open-source, cutting-edge, and backed by Beijing, it was forcing U.S. companies to reveal their cards. Trump’s deregulation flipped the script. The Stargate Project is America’s answer.

Yes, this might sound like a tall tale. Think Manhattan Project meets Apollo Program—but bigger. This isn’t about rockets or nukes. It’s about intelligence. Imagine billions of Einstein-level (AI) minds solving the world’s biggest challenges. Medicine. Engineering. Space exploration. Every sector is about to be transformed by AI super Agents—autonomous AIs that can outthink the smartest humans.

But there’s a catch: exponential change is impossible to grasp. AGI isn’t just coming; it’s knocking on the door. And ASI? It’s not far behind. Recursive AI research—AI improving itself—isn’t science fiction anymore. It’s just a few years away. Perhaps even only months. The January 30th closed-door meeting at the White House is a sign of how serious this is. Sam Altman, OpenAI’s CEO, will brief government officials on what’s next. Insiders say breakthroughs in PHD-level AI super Agents are imminent.

With The Stargate Project, the U.S. is building more than AI models. It’s building an ecosystem. Chips, data-centres, energy, and talent. Texas is ground zero, but the ripple effects will reach every corner of the country. And the world? It’s already playing catch-up. Later in this Dip, we explore how the UK is attempting to launch an AI initiative, but it pales in comparison to what we’re witnessing here.

Exponential technological progress comes with risks. History reminds us that unchecked innovation can lead to unintended consequences. Deregulating AI raises critical questions: How do we prevent misuse of these powerful systems? Will economic benefits be evenly distributed? And how do we ensure AI aligns with human values instead of amplifying existing inequalities?

Here’s the reality: AI is now a national issue. Governments will want more control over training data, infrastructure, and standards. The U.S. isn’t just investing in technology; it’s shaping the global rules of AI. And if done right, this will protect democracy while ensuring AI serves humanity.

You might be wondering, what does this mean for me as a business leader or professional? Well:

• Are you prepared to collaborate with AI? AI super Agents are coming. How will you leverage AI to enhance your expertise or lead your team into this new era?

• Do you understand the impact on your industry? AI will reshape every sector—from healthcare to finance to engineering. What changes are already happening, and how will you stay ahead?

• How will you future-proof your skills or organization? Adaptability is critical. Are you developing the skills, workflows, or culture to thrive in an AI-dominated world?

• Are you aligning with the shift to AI-driven ecosystems? Whether you’re leading a business or building your career, understanding the infrastructure being built (like The Stargate Project) is essential.

AGI is almost here. Exponentials are hard to internalize, but they’re here. The AI Manhattan Project has begun, and a new world is coming. If you’re reading this, consider it a final wake-up call, not a doomsday alert. AI is accelerating. Are you ready to lead in the Intelligence Age?

OpenAI introduces Operator, an AI that works for you

TL;DR

Operator is OpenAI's latest AI agent, capable of independently performing tasks on the web via its own browser. It can shop for groceries, book trips, fill out forms, and more, all while allowing users to monitor and take control when needed. Although still in its research preview, Operator demonstrates immense potential for making repetitive and complex tasks easier, safer, and faster for non-tech users.

Read it yourself?

Sentiment

OpenAI just raised the bar for AI agents with Operator. It’s not perfect yet—there are limitations, like restricted website access and occasional blocks—but the concept feels like a glimpse into the future. Being able to delegate multi-step, browser-based tasks to AI opens up a new level of productivity. It’s especially exciting for those who aren’t developers, as it requires zero API integrations.

My thoughts

While OpenAI isn’t the first to roll out AI (browser)agents like this (Claude and Bytedance are already in the race), Operator sets itself apart by targeting non-tech users. And that’s huge. It’s not just for shopping or booking flights—it has the potential to handle work-related tasks like testing dev environments, scheduling meetings, or generating mockups. Right now, Operator is exclusive to Pro users at €200/month, but it’s only a matter of time before it becomes more accessible via cheaper subscriptions like Plus. OpenAI’s goal is clear: drive adoption by making AI an everyday assistant. Whether for personal or professional tasks, Operator is laying the groundwork for an era where your devices don’t just assist—they work autonomously.

Deepseek R1 reshapes open-source AI

TL;DR

DeepSeek R1 is an open-source reasoning model that competes with OpenAI’s o1 at a fraction of the cost. It’s five times faster, 30 times cheaper, and now supports web reasoning for real-world problem-solving. Built using reinforcement learning instead of human-curated data, R1 sets a new standard for open-source AGI development and accessibility.

Read it yourself?

Sentiment

DeepSeek R1 feels like a turning point for AI. The ability to reason over live web data, coupled with its affordability and open-source availability, makes it revolutionary. Its use of RL training over traditional methods signals a bold shift in how advanced AI can be developed. This model is reshaping the open AI ecosystem, bringing cutting-edge tools to a wider audience. In contrast, many people also suggest that this is a new Chinese Trojan Horse. Similar to Temu, for example. Developing a product as cheaply as possible in the hope that as many people as possible will adopt it, only to then use or misuse all your data. Additionally, questions about President Xi or Taiwan seem to be off-limits.

My thoughts

We have already discussed in this and earlier Dips what this means for geopolitics and the race between the US and China, so we won’t go over that again. DeepSeek R1 is not just another AI model; it’s a movement. It matches o1-mini’s power but is drastically cheaper and faster. Web integration and its COT reasoning process make it incredibly versatile. What’s even more exciting is R1-Zero, trained without human input, proving that reinforcement learning (RL) can teach AI to surpass previous benchmarks autonomously. By being open-source and MIT licensed, DeepSeek R1 unlocks opportunities for smaller teams and researchers to compete at the highest levels of AI development. This is a massive leap forward for open-source AGI, ensuring the technology remains accessible and affordable while pushing the boundaries of what’s possible. And of course, I’m not blind to the Chinese Trojan Horse either. While I applaud the development of these models for encouraging other companies to take action, I would never use it myself.

More byte-sized breakthroughs:

• Introducing perplexity assistant
  Perplexity Assistant simplifies your day with smart reasoning and seamless multi-app actions. It can book dinner, answer complex questions, or set intelligent reminders. Its web browsing powers ensure accurate results for your tasks. Plus, its multimodal capabilities let you interact via text or camera. Everyday productivity, made effortless—now available on Play Store.

• Tasks in ChatGPT lets you schedule smarter
  OpenAI’s new Tasks feature makes scheduling simple and efficient. Set one-time reminders or recurring actions, like daily news updates or project follow-ups. Tasks can suggest frequent actions based on your chat history for a personalized touch. With up to 10 active tasks, ChatGPT helps you stay on track effortlessly. Rolling out now for Plus, Team, and Pro subscribers.

• Microsoft 365 Copilot brings AI to everyone
  Microsoft 365 Copilot introduces powerful AI agents to handle tasks within your favorite apps. From drafting documents to analyzing data and summarizing meetings, Copilot works alongside you seamlessly. These agents adapt to your workflow, offering smart suggestions and handling complex actions. Integrated into Word, Excel, Teams, and more, it’s the next step in effortless efficiency.

A must-see webinar, podcast, or article that’s too good to miss.

Composio - Integrate AI, faster and smarter

Composio is the ultimate platform for connecting AI agents and LLMs with your existing tools. Easily integrate with 250+ tools like GitHub, Salesforce, and Slack, and automate workflows in no time. Perfect for engineers and teams looking to scale quickly without worrying about complex integrations.

Taskade - Automate, optimise, achieve

Say goodbye to repetitive tasks with Taskade’s new Loops feature. Schedule social posts, send bulk emails, analyse surveys, or build multi-step workflows. Automations have never been this easy. Just add a step, set a loop, and watch your work get done.

LangChain - Streamline human-AI collaboration

LangChain’s new Agent Inbox UI brings humans into the loop for AI workflows. Connect to any LangGraph app, manage interrupt events, and guide AI agents in real-time. Ideal for developers and teams handling complex, dynamic tasks with precision.

A roundup of updates that are too cheesy to ignore.

• Perplexity launches @PPLXsports with new NFL widgets, including Playoff tables and homepage updates (US only).

• Perplexity launches Sonar, an affordable search API for real-time info and citations, with a Pro version for advanced features.

• Perplexity moves to merge with TikTok US, aiming to combine generative AI with TikTok's massive user base.

• LumaLabs launches Ray2, a video generative model with 10x compute, delivering lifelike motion and enhanced creative possibilities.

• The UK unveils its AI Opportunities Action Plan to drive growth, create jobs, and attract global AI investment. Aiming to be a world AI superpower.

• DeepSeek App is launched. It offers web search, deep-think mode, and file upload with text extraction. 100% free with no ads.

• LTX Studio launches Auto SFX, a tool that adds immersive, mood-matching sound to your scenes automatically.

• FC Barcelona teams up with Cardano ($ADA) to boost fan engagement using blockchain technology.

• Kinetix introduces Character Motion Control for video diffusion models, letting creators guide character movement with acting videos.

• Google brings AI to Workspace with Gemini, embedded across Gmail, Docs, Sheets, and more—no add-ons needed.

• Relevance AI launches Agent Generator ‘Gen’—build custom AI agents in minutes with no tech skills.

• Codeium launches Windsurf Wave 2, featuring web search, autogenerated memories, and smarter problem-solving tools.

• OpenAI updates custom instructions, letting you easily personalise ChatGPT’s tone, traits, and rules.

• OpenAI updates ChatGPT Canvas. Now compatible with o1 for coding and supports rendering HTML/React for mini-apps.

• OpenAI announces the free tier of ChatGPT is getting O3-mini, and the Plus tier will enjoy extensive O3-mini access.

• OpenAI releases a reference implementation for building multi-agent (voice) apps with the Realtime API

• Kling launches Elements in early access, letting you integrate up to 4 images to create consistent characters, environments, or objects.

• Tencent unveils Hunyuan3D-2, a groundbreaking 3D AI model that generates and animates 3D models.

• Meta reveals plans for AR glasses, smartwatches, and earbuds, aiming to rival Apple’s hardware dominance.

• Bytedance releases UI-TARS, two open-source models (7B & 72B) with a PC/MacOS app to control your computer via vLMS.

• Google updates Gemini 2.0 Flash Thinking, now with code execution, a 1M-token context window, and improved reasoning.

• Google invests $1B in Anthropic, backing its own Gemini competitor while doubling down on advancing AI innovation.

• Lovable launches Figma to Lovable, turning your Figma designs into full-stack apps with just a prompt.

• Spline introduces Spell, a tool to create 3D worlds from images in just minutes. The worlds match the image input and are easy to render.

• Freepik launches MiniMax Reference mode, bringing AI video with character consistency to the Freepik AI Suite.

• Freepik adds sound to AI videos! Create videos with audio in any mode or format.

• Freepik launches Hunyuan V2 Pro mode in its AI Image Generator, offering control and multiple shots in a single generation.

• FilmAgent AI launches the world’s first open-source 3D AI film production studio: scripts, acting, camera angles, set design, all in one tool.

• Krea introduces real-time custom AI models, letting you train styles, characters, or products for instant use in Krea Real-time.

• Krea launches 3D. Turn images into 3D objects and use them in the Real-time tool.

• X Money is here. Creators can now monetize posts, streams, and Spaces. From payments to crypto, X becomes your bank.

• Qwen Chat v0.2 is here, featuring web search, video creation, and image generation, alongside tools like artifacts and document comprehension.

• Hailuo launches MiniMax-01 with 4M-token context, Lightning Attention architecture, and unbeatable cost-efficiency—now open-source.

• Hailuo Audio HD launches T2A-01-HD, a breakthrough in voice synthesis. Clone voices in seconds, add emotional depth, and speak in 17+ languages.

• Mistral unveils Codestral 25.01, a new coding model supporting 80+ languages, 256k context length, and 2x faster performance.

• Devin 1.2 is here with improved repository reasoning, voice messages via Slack, enterprise account management, and usage-based billing.

• Firecrawl launches /extract: write a prompt, get web data—no scrapers needed. Now in open beta and integrations for Python SDK and Zapier.

• Moonshot AI introduces Kimi k1.5, a multi-modal model with o1-level performance.

This was it. Our twenty-eighth digital dip together. It might seem like a lot, but remember; this wasn't even everything that happened in the past few weeks. This was just a fraction.

If you’d like to explore practical, forward-thinking ways to incorporate AI in your business or career—balancing innovation with responsibility—let’s talk. This isn’t about keeping up; it’s about getting ahead—and staying there. Just drop me a message!

In the next Dip, I’ll share the ultimate prompt to get the best out of ChatGPT and boost the quality of your output.

Looking forward to what tomorrow brings! ▽

-Wesley