Agustín C.

Yeah, n8n is great. But if you’re building AI agents specifically—this might be the open-source alternative you’ve been looking for. Langflow lets you design and deploy AI-powered agents with a drag-and-drop UI. No code. Just plug in your LLMs, tools, vector DBs, and go. ��� Fully open-source → 62K+ stars → Built for AI workflows from the ground up If you’re serious about AI agents, this isn’t just a toy—it’s infrastructure. https://lnkd.in/ehB_tP3Y #Langflow #AI #Agents #OpenSource #n8nAlternative #LLMops #DevTools #Automation #NoCode

147

7 Comments

Over the past six months, our team at Development Seed has been collaborating with the Land & Carbon Lab and World Resources Institute, with support from Bezos Earth Fund to build globalnaturewatch.org. We're exploring what an IDE for maps could look like in the future - this first version is our attempt at building agents and reimagining map-based interfaces. We drew inspiration from tools like Cursor and Windsurf - how they understand user context, recommend next steps, and predict actions through auto-edit and suggested workflows. We're applying similar thinking here. The agent uses a simple ReAct framework with specialized tools for specific tasks. Our goal is to ground AI responses in trusted data from LCL and WRI. Building trust in AI workflows is central to what we're creating. All data and code are open. I've learned so much about my area by using this over the past few weeks. In the demo, I discovered that Kandhamal, Rayagada, and Gajapati districts account for 60% of tree cover loss in Odisha - all in under a minute. These kinds of insights can help people ask the right questions and take timely action. Try it out, join the waitlist, or reach out if you'd like early access or want to discuss further. Read more here: https://lnkd.in/gQ2SRbRN

162

7 Comments

Last year we found over 4,000 secrets in GraphQL endpoints. This included: - 1,396 access tokens - 49 passwords - 2 credit card numbers And thousands of other secrets leaked through stack traces All publicly accessible. GraphQL makes it way too easy to expose sensitive data. At BSides SF, we shared a simple takeaway: → Use the GraphQL Secure by Default checklist. Start with basics and then focus on the GraphQL-specific best practices: - Depth query & cost limits - Rate limiting - Aliasing secrets management And if you want to go deeper: - Field-level RBAC - Federation boundary hardening - Continuous DAST (e.g. Escape) - Solid observability Want the full security checklist? Drop a “GraphQL” below.

63

6 Comments

Here’s a list of papers I’d study before planning / developing a generative music model. They’ll give you a strong foundation. You’ll also get an idea of the state of the art, and what to expect. 1. Museformer: Transformer with Fine- and Coarse-Grained Attention for Music Generation  2. Mousai: Text-to-Music Generation with Long-Context Latent Diffusion 3. MusicLM: Generating Music From Text 4. Simple and Controllable Music Generation 5. Masked Audio Generation using a Single Non-Autoregressive Transformer 6. Long-form music generation with latent diffusion 7. Video2Music: Suitable Music Generation from Videos Using an Affective Multimodal Transformer Model 8. ACE-Step: A Step Towards Music Generation Foundation Model You have to be comfortable with the topics below to follow these papers: - Transformers  - LLMs - Diffusion / flow models - Encoder architectures - Spectrogram / Mel-spectrogram representations — 💼 Check my music tech advisor service: https://lnkd.in/divEFrTA 📩 Follow my content: https://lnkd.in/d7MPrPQT #GenAI #AIMusic #GenerativeMusic 

164

8 Comments

We’ve been working in AI evaluations for a long time, starting with Ragas. Now, scaling evaluation is becoming the foundation for training the next generation of frontier models, using methods like RLVR. We’ve raised $2.5M in pre-seed funding and are launching Vibrant Labs to benchmark and build environments that improve the long-term capabilities of AI agents, starting with web and computer-use agents. Visit us at vibrantlabs.com.

455

66 Comments

Growing eyes on the back of your head might sound like a scene out of a sci-fi movie—but it’s actually happening in a lab at Tufts University. Incredible work from Michael Levin’s team is showing us what’s possible when we understand and manipulate bioelectricity—the electrical signals that cells use to communicate and organize. By adjusting these signals, they’ve not only grown functional eyes in unusual places like the backs and tails of frog embryos, but also reprogrammed cancer cells to stop growing out of control—all without changing their DNA. This kind of research opens the door to a whole new world of regenerative medicine—think growing limbs, healing organs, or even reversing disease—simply by “talking” to our cells in their native electrical language. We're stepping into a future where biology and technology merge in ways we’ve only dreamed of. ✨

38

4 Comments

News of Claude Code coming to Slack is already breaking, and it reflects how quickly AI-assisted development is advancing. Developers will now be able to move from a Slack conversation directly into generating code, fixing issues, or submitting pull requests without switching tools. This keeps context clear and reduces many of the delays that typically slow teams down. What interests me most is how AI is shifting from a simple coding assistant to a true workflow partner. When communication and execution happen in the same environment, engineering teams gain alignment, speed, and more space to focus on higher-value work. At BairesDev, we are already seeing how these capabilities reshape the pace and quality of software delivery. The impact is early, but it is becoming more visible every day. How are you approaching AI’s role in your development workflow as these tools continue to evolve? TechCrunch has the story: https://lnkd.in/g5t_SenA

73

1 Comment

🌐 AI isn’t the main story anymore. Adoption is. After dozens of conversations with startups, investors, and donors working with AI, I found myself circling back to a view I had in 2022. It's more relevant now than ever — and more debatable. There are two kinds of adoption: Narrow – where AI agents are tailored to solve very specific, complex tasks. It’s painful. Setting up LLMs for such tasks feels like teaching a three-year-old with perfect memory but no real understanding. Whoever cracks a scalable methodology here (for expert teams) will be sitting on a gold mine. Mass – integrating LLMs into everyday products to automate routine user tasks. Here, success depends less on model quality (we're nearing diminishing returns) and more on product integration. 🔮 My personal bets: Google/Gemini has a shot, despite lagging behind technically — thanks to Android’s massive user base and iteration speed. Apple could still surprise us, but they’re slow out of the gate. Samsung is making smart moves, but they’re still tied to Android. Meta? I don't believe in their use case for AI adoption. Russia's AI experiments? Let’s just say — I’m not worried. 🏆 As for LLM chat products, I’m still bullish on OpenAI. Whether standalone or absorbed by Microsoft/Apple, they’ll stay at the core of the ecosystem. I also expect Microsoft, Oracle, and Amazon to dominate enterprise-level AI infrastructure. Everyone else? Most will be acquired. But the real assets will be engineers and chips — not business models. 🛋️ Just some couch thoughts. Curious what others are seeing. What’s your take on AI adoption today? #AI #LLM #Startups #ProductStrategy #Adoption #OpenAI #GoogleGemini #EnterpriseAI #TechTrends2025

86

5 Comments