Introducing coglayer - a symbiotic intelligence thinking partner:
🤖 Semi-Autonomous: Interacts with you by asking questions & guiding your thought process
🔗 Interconnected Prompts: Adapts to your needs with a self-instructing system
🧩 Self-Structured: Personalizes & focuses on your preferences for a tailored experience
🎯 Consistent Alignment: Always stays aligned with what's best for you without getting off track
Try it for free today (no waitlist required).
I've just played with @coglayer to generate a long non-fiction article. I have my own glue-and-tape system that leverages multi-turn, multi-agent, web research, section-by-section process to generate long articles. So, I have a direct comparison of outputs for an article on the same topic.
Here are my thoughts:
- Coglayer's output is nice, dense and consistent - to me it indicates that there might be a plan (table of contents) generated before writing
- clarifying questions before starting the writing process are very cool - it's a trick I need to incorporate as well
- organization errors (sections written partially then started again to continue) indicate that Coglayer isn't writing section by section and gluing stuff together, but tries to write it until it reaches token limit or so - it might be that it helps maintain consistency
- ability to pull in scientific information with correct citations is FANTASTIC, information is relevant (maybe Semantic Scholar?) and it seems to be up to date - this part I'm struggling the most in my system when leveraging LLMs for writing (what kind of RAG system they have?!)
I'm sold.