In the age of AI, what kind of knowledge management tools do we need?
When we talk about knowledge, what exactly are we referring to? The American bestseller “Building a Second Brain” presents the CODE principle, which involves completing the personal digestion from data, information, knowledge to wisdom through the steps of Capture, Organize, Distill, and Express. In the digital age, with the explosive growth of information, knowledge management is particularly important. Users have clear needs: first, to record and remember important knowledge amid information overload; second, to effectively filter information sources; third, to transform complex information into valuable knowledge, form an individual knowledge system and keep it up to date.
However, in the era of large language models, new problems arise: Do we still need knowledge management? This depends on the differences between people and the knowledge pedigree of large language models. The “creation” of large language models is based on what is known and there will be “machine hallucinations”. Moreover, there are problems in terms of knowledge timeliness, depth, privacy and security, and personalized needs. Therefore, knowledge management is still needed in the era of large language models.
In the era of large language models, what kind of knowledge management tools do we need? Technology is a help, and users have key pain points in knowledge management. The knowledge management process can be divided into four dimensions: data collection and collation, information search and filtering, knowledge refinement and circulation, and multi-person collaboration and sharing. Evernote has created a vertical large language model focused on the field of knowledge management – Evernote’s large language model. Its core advantages lie in hybrid deployment, public and private domain data management, complementing modular intelligent agent functions and obtaining feedback to nourish the large language model. Evernote’s large language model does not yet have capabilities such as text-to-image generation, but it will consider connecting to third-party large language models to provide such capabilities and services.
2024 is the first year of AI application. Evernote wants to be an AI super application, taking into account six elements: data, model, carrier, interaction, scenario, and user. Currently, Evernote has several implementation scenarios:
– Evernote AI Creation: That is, AI writing. In April 2023, intelligent writing capabilities were launched. AI can participate in multiple writing scenarios and will also actively suggest complete prompts.
– Evernote AI Reading: Through AI’s intelligent analysis of notes, documents, etc., automatic summaries are generated, active questioning accelerates information efficiency. Web articles and scanned documents can be automatically summarized, intelligently analyzed and generate intelligent mind maps.
– Evernote AI Knowledge Assistant: Integrates different tools, decomposes and plans daily tasks, provides users with summaries and reflections, and allows multiple intelligent agents to collaborate to find solutions.
– Smart hardware products: Such as conference headphones, AI e-ink office notebooks, smart pens, etc., solve the problem that information cannot be digitized and improve the efficiency of information compilation and organization.
Evernote hopes that through the overall functional layout, it can help users manage data, information and knowledge from different dimensions, generate greater efficiency under the empowerment of AI, and take us to the depths of knowledge.
