10 lessons from a year of getting good with AI

This year we've taught 100s of motivated people to start using AI effectively to achieve their goals.

Here are the top 10 lessons that have helped our community and move us all forward this year.

Its more evolution that revolution, but definitely a shift change.

1. Know your (llm) models

chatGPT is maybe where we started but different AI's will help with your tasks. We use GPT, Gemini, Claude & Mistral - all accessible on Launch Lemonade. Language Model Arena good place to track what's hot > https://lmarena.ai/

2. Know your LLM ‘secret options’

The interfaces for these models are very similar and you can get more out of them by going to the drop down options... eg on GPT select 'THINKING' and within the chatbox, select 'AGENT' etc...

3. Better prompting: RCOTE

Tell your AI's that you like to prompt the RCOTE way...then they will write you better prompts

4. Context - and memory - is king

Other than quality data, the most important thing in all AI interaction is providing context... if your AI has a good 'brain' and 'knowledge base' of your work and goals...the outcomes are much better.

5. Personal Software

We've started creating little bits of software that are personalised to our projects - eg onboarding flows. The era of paid SAAS feels like it could be over.

6. Publish & Share Projects Quickly

We've always liked using tech tools to get ideas live...now this feels even more viable. We published https://goodwithai.org/ on Lovable.

7. Set-up your privacy right

Don't feed the training models with your data (unless you have specific mission). You can adjust all these in your AI's settings. Turn them off. Its like 'opting out'. This will protect what you share.

8. Create like a pro

The creativity debate about AI is understandably intense. But we’re enjoying seeing how nanobana, suno, veo, veed and more are being harnessed to tell new stories. The challenge is avoiding the algo-driven slop. Our sense is that human verified creativity will end up getting more value.

9. Start to automate

This is the one we all want but isn't as easy as we hope. But for those who have built forms, basic websites, email lists and use tools like Zapier... effective automation is close. This is an area for us to focus on next year.

10. Use AI for purpose

Last in our top 10 but the most important. We started exploring AI because we could see its potential upsides for impact and empowerment - as well as the downsides. After a year going deeper into teaching it, we stand by this. Ultimately, the real impact happens, not simply by using AI but WHAT we use AI for. There are so many great examples already - pictured: Winnow (food waste), tracking Snow Leopards in Pakistan and Grey Parrot (smart recycling). We'll keep sharing them.

Want to get good with AI. Join us in January > https://goodwithai.org/

£100 off any cohort with EARLYBIRD2026 (ends December 10th)