@thevarunmayya's reel — transcript & breakdown

Introduces a new, 'weird' prompting trick that promises to revolutionize AI interaction, creating a strong curiosity hook.

“There's a weird prompting trick called ASD STE 100 that's borrowed from aerospace manuals. And once you try it, you'll never prompt AI the same way again.”

Explains the first trick, ASD-STE100, showing real-world application (Andrej Karpathy's tweet) and before/after examples to demonstrate its effectiveness in making AI output clear and readable.

“First, writing. Basically, you ask your LLM to explain something in ASD-STE100. It's a controlled language specification originally developed for aerospace maintenance documentation. LLMs are well-versed in it, and it comes with heavy constraints on clean writing style that makes answers a lot more readable.”

Introduces the second trick: generating diagrams. Visual examples quickly illustrate how diagrams make complex information easy to process, appealing to visual learners.

“Second, diagrams. Even clean writing still has to be read. So instead of writing, ask your LLM to create a diagram. These can be a lot easier to process fast and understand.”

Demonstrates the third trick: generating interactive web pages (HTML). Showing a working SIP calculator created instantly highlights the practical and powerful capabilities of LLMs in front-end development.

“Third, web pages. Just ask for the output in HTML. Karpathy says LLMs are getting really good at front-end. Ask for an SIP calculator that shows compounding year by year, and you get a working page with sliders and animations.”

Reveals the fourth trick: creating custom explainer videos. This highlights the highest level of AI-generated content, pushing the boundaries of what users can achieve.

“Fourth, explainer videos. The format he's most bullish on is fully custom or bespoke explainer videos generated on any arbitrary topic. Try, 'create a 3Blue1Brown style video explainer on X. Use my ElevenLabs API key for audio narration'.”

Summarizes the broader impact of these tricks, showing how AI can automate complex tasks and shift human work towards oversight and higher-level thinking.

“Now, LLMs will do more and more of the legwork autonomously, and our work will rise up the abstractions into oversight and understanding. Intelligence and code are now abundant, so you can ask for large, custom, discardable software artifacts that would have never made sense to create before.”

Encourages viewers to experiment and engage by sharing their own favorite prompting tricks, fostering community interaction.

“Push the boundaries here, and you'll be surprised. What's your favorite prompting trick? Comment below.”