What is the best way to learn new skills?
This topic is genuinely contested — read the individual model responses for different perspectives.
Models agree on
- ✓Set clear learning goals
- ✓Break the skill into smaller sub-skills
- ✓Apply the 80/20 rule to prioritize high-impact components
- ✓Identify and address potential barriers
- ✓Use structured learning resources (e.g., Coursera, mentors, hands-on practice)
The most effective way to learn new skills combines structured goal-setting with evidence-based learning techniques. Here’s the best approach:
- ·
Set Clear Goals: Define specific, meaningful objectives that align with your interests or career needs. Both models emphasize starting with well-defined targets.
- ·
Break Down the Skill: Deconstruct the skill into smaller, manageable sub-skills. This makes the learning process less overwhelming and more actionable.
- ·
Apply the 80/20 Rule: Focus on the 20% of sub-skills or activities that will deliver 80% of the results. Prioritizing high-impact components accelerates progress.
- ·
Identify Barriers: Proactively recognize potential obstacles (e.g., time, resources) and plan workarounds to avoid derailment.
- ·
Use Structured Resources: Leverage reputable platforms like Coursera for university-backed courses or other proven methods (books, mentors, hands-on practice). Hermes highlights diverse learning tools, while Seed specifically recommends credentialed online options.
- ·
Practice Deliberately: Dedicate consistent time to focused practice, seeking feedback from experts to refine your technique.
- ·
Apply Skills in Real Contexts: Reinforce learning by using the skill in practical scenarios, tracking progress, and adjusting strategies as needed.
- ·
Stay Motivated: Reward milestones to maintain momentum. Hermes adds the importance of patience and perseverance.
Seed uniquely notes the value of choosing skills with long-term impact (e.g., critical thinking, digital literacy), while Hermes provides a more detailed framework for execution (e.g., progress tracking, real-world application). Both agree there’s no universal "best" method, but a systematic, prioritized approach yields the strongest results.
Follow-ups
You just saw open-source models answer
Want GPT-5, Claude, Gemini & more on the same question?
Sign in free to run any question against frontier models — side by side, same synthesis, honest comparison.