Most of us are using LLMs for work these days. Personally, I use them across multiple domains:
- Software Engineering – Tools like Copilot, Cursor, and Claude help streamline coding, debugging, and refactoring.
- Game Design—Perplexity is great for finding references, while GPT assists with documentation and game design documents and acts as a “brainstorming companion.”
- Art Direction – My ugly sketches paired with DALL-E or MidJourney help me create better references and make clearer art requests for artists.
- Data Analysis – Here, my choice is Claude while still in the process of testing.
- Language Learning – Let’s be honest, HSK textbooks are dry and outdated. GPT + DeepSeek makes learning Chinese and Serbian much more interactive and engaging.
- Social Media - I still prefer to write texts myself, but I need a good editor to help with grammar and style.
Since we rely on these tools so much, it makes sense to understand them better. You don’t need to be an LLM researcher, but knowing basic concepts like tokens, context windows, or fine-tuning can significantly improve your use of them.
It’s like understanding steel types when choosing the correct hammer or knowing about shoes to adapt your running technique for different weather or road conditions.
For those interested, I highly recommend Andrej Karpathy’s latest video, “Deep Dive into LLMs like ChatGPT.” 🚀 Warning: it’s 3.5 hours long, but trust me, it’s way more interesting than a Hollywood blockbuster (at least for me — I usually fall asleep 30 minutes into any action movie, but here, I couldn’t stop watching).
I’d love to hear your recommendations for other videos, books, or podcasts on LLMs!
#AI #LLM #Productivity #GameDev #AIforWork #DeepLearning