Obtain Better AI Summaries with Visual and Interactive Outputs
As language models take over execution steps, workers should direct outputs toward high-level evaluation rather than long walls of text. Requesting diagrams or custom visual explainers helps humans review complex conclusions faster.
Try it yourself
- 1Select a detailed subject or dataset that requires analysis.
- 2Prompt the AI model to output the findings as a diagram structure, webpage code, or video script instead of standard paragraphs.
- 3Review the generated visual format to quickly check logic, spot errors, and verify the main points.
The Blend
As artificial intelligence models take on heavier workloads, humans spend less time writing and more time reviewing what software generates. Former OpenAI researcher Andrej Karpathy highlighted this shift in a recent post on X, suggesting that standard text responses are no longer the best way to process complex information. Instead, he recommends asking chatbots to present their conclusions through structured visual formats and custom software tools.
Karpathy noted that asking models to write using simplified technical standards, such as aerospace writing rules, creates far cleaner text. Even better, users can instruct tools to output interactive HTML web pages, visual diagrams, or custom voiceover videos on demand. Because generating code has become fast and inexpensive, workers can create single use websites or short visual lessons just to understand a single difficult topic quickly.
This approach shifts human work from reading tedious walls of text to reviewing tailored visual dashboards. However, it remains uncertain whether non-technical users will adopt the complex prompts needed to trigger these formats, or if software makers will eventually build these rich visual outputs directly into default chat interfaces. If interactive summaries become the standard, learning complex subjects could become far more intuitive for the average employee.
Written independently by AI News Smoothie from the reporting listed below. Facts belong to the original publishers. Follow the links for their full coverage.
Ingredients
- Andrej Karpathy on X: "We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:
Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for … / X
Computer scientist Andrej Karpathy suggests asking language models to output diagrams, interactive web pages, and videos to help humans understand complex information faster.