Inspiration : A few months ago, I found myself struggling to keep up with the relentless pace of innovation in AI, Web3, and quantum computing. The landscape was fragmented—information scattered across countless newsletters, feeds, and platforms. Most newsletters simply recycled the same headlines, and the audience was split: some focused only on AI, others on Web3 or quantum. There was nowhere to view these future-defining technologies through a unified lens. That’s when the idea hit me: what if there was a platform that brought these worlds together, providing a single point of view across disciplines? I believe the next generation of the internet will emerge at the intersections of these domains. This quest led to The Rift—and at its core, Pulse.
What it does : Pulse is a fully automated engine that keeps you at the bleeding edge of emerging tech, tracking over 150 companies. Designed for maximum efficiency, it monitors, curates, and distills news from a vast array of sources, delivering concise summaries of everything from AI model launches to blockchain breakthroughs and quantum advances—all in one place, in real time. With Pulse, it takes less than four minutes a day to stay ahead of the curve.
How we built it : The journey started as a personal mission to understand the shifting tech landscape. Early on, I realized that AI alone wasn’t enough for reliable curation—so I leaned into Make.com, discovering powerful automation capabilities. I began with sourcing, but getting consistent curation from AI was tough. After weeks of fine-tuning agentic workflows, experimenting with different models and balancing cost with accuracy, I landed on a curation engine that just worked.
Real-time access via Sonar Pro was a game changer. Once curation was nailed down, I tackled analysis and editorial. Sonar Reasoning Pro proved crucial here, offering real-time enrichment and a strong macro perspective. Tuning it took patience and a lot of iteration—but after months of work, Pulse delivers analysis that’s both timely and insightful.
Challenges we ran into : Tuning instructions for Sonar Reasoning to extract real macro impact was tricky—especially with single-shot workflows. Injecting real-time enrichment data helped, but required extra data engineering work. Once I mapped out a consistent analysis framework, I could send articles downstream for further summarization and copy by other LLMs, tightening the loop.
Accomplishments that we're proud of : I’m proud that Pulse is built almost entirely on no-code platforms, with just a few code snippets running on Replit. It’s proof that no-code and AI together can build real, scalable products. I’ve been a systems engineer for 20 years, but never a coder—until now. I built something that can already scale to hundreds of sources, and soon will be fully autonomous.
What we learned : LLMs are powerful, but their non-deterministic nature creates real challenges for consistent output. This project made it clear: solid data engineering is crucial. Carefully structuring and tuning instructions makes all the difference if you want dependable AI workflows.
What's next for The Rift Pulse : Pulse is just the start. As a core content engine for The Rift—an AI-native media platform—Pulse will keep evolving as models and frameworks improve. My vision is a fully autonomous, agentic news engine: adapting to new industries, tones, topics, and frequencies on its own. Pulse is laying the foundation for AI-native news and media—and the journey is only beginning.
Built With
- airtable
- anthropic
- make.com
- openai
- perplexity
Log in or sign up for Devpost to join the conversation.