Hey everyone! Our discussion on the Dolphin Gemma project, sparked by @Pierre-Henry Isidor , inspired me to share an article I came across: “The Ethics of Using AI in Scientific Research” by Resnik and Hosseini (2024): https://link.springer.com/article/10.1007/s43681-024-00493-8. It offers essential guidance for those exploring AI in science or business. Why It Matters: The article highlights AI’s transformative potential, such as AlphaFold’s breakthrough in protein modelling, alongside challenges like biased COVID-19 data models. For businesses, it provides strategies to implement AI responsibly, preventing issues like unfair hiring systems or unclear customer analytics. Practical Insights for Businesses: - Ensure Fairness: Regularly audit AI systems, as scientists do in research, to maintain equitable outcomes in hiring or marketing. - Promote Transparency: Clearly communicate how AI influences decisions, such as loan approvals, to build customer trust. - Engage Stakeholders: Involve customers in AI development, mirroring scientists’ community consultations, to align with real-world needs. Your Thoughts: How can businesses adopt these principles to use AI ethically? Have you observed effective or problematic AI applications in your industry? Please share below! #EthicalAI #AIinBusiness #ResponsibleTech

There is this research from ChromaDB that talks about the different strategies for chunking text without loosing valuable context for optimal RAG systems, but I'd recommend checking out Adam Lucek's YouTube channel where he demonstrates with practical examples these strategies. He also made available a Github repository (QuicKB) where he puts in practice all this. I think it's a really valuable resource. Don't miss the link to the video below! Cheers! 🙂

Hi All, as I tackle the first few courses on here which seem really well written and useful but I have a nagging question that I am hoping you all can help me answer. I am looking to learn AI and become an expert in my workplace by implementing it to save time and money on every aspect I can, from simple automation, to full company org style agentic systems with image process etc, My question - do I REALLY need to learn this level of coding and knowledge to do that if I an to rely on mainstream AI platforms and workflows? or am I wasting my time. I am a generalist anyway, so specialising on things to a deep level is not my thing, I tend to join the dots on things and let people that are more clever than me, do the detailed stuff. I see AI as that person/entity that will help me do that. What are you thoughts and experiences? I I would rather concentrate my time and effort into the area that will help the most, if I do not need this deep knowledge then I won't but interested in your thoughts. Cheers, Neil.

"Hey Data Alchemists! AI agents are evolving fast, from coding assistants like OpenAI’s Codex CLI to edge-optimised models like DeepSeek-R1. But let’s get real—building or using AI agents can be a steep learning curve. If you could instantly master ONE skill to level up your AI agent game in 2025, what would it be? For me, it’s prompt chaining—crafting workflows that make agents solve complex tasks reliably (like avoiding those pesky hallucinations we all hate) (https://cobusgreyling.medium.com/chaining-large-language-model-prompts-81091daccaad). What’s yours? Maybe it’s fine-tuning open-weight models like Llama 4 Scout, debugging agent failures, or even explaining AI decisions with XAI? Drop your pick below and share a quick tip or challenge you’ve faced with it. Let’s spark some ideas and help each other grow! P.S. Inspired by the ‘How to Build Effective AI Agents’ post