Watching seasoned programmers like Andrej discover the low hanging fruit in tools like Cursor is like seeing seasoned creatives jump into ComfyUI and discover all that it can already do 🙂
PROMPT ENGINEERING
-
Using XML for organizing inputs instead of markdown fences
By
–
I understand- hence the 2nd part of my response. I don’t think that helps. Claude has read loads of fenced markdown blocks and is good at using them. So I just use xml for organising my inputs.
-
Claude Projects XML Document Integration Best Practices
By
–
If you add them to a Claude project I assume Claude adds xml for each doc. I don’t think there’s a benefit to putting xml within each doc
-
Two-Stage Prompting Technique Enhances LLM Robustness
By
–
5). Enhancing Robustness in LLMs – proposes a two-stage prompting technique to remove irrelevant information from context; it serves as a self-mitigation process that first identifies the irrelevant information and then filters it out…
-
Top ML Papers: GraphRAG, LLMs, Agents and Robustness
By
–
The Top ML Papers of the Week (August 19 – 25): – GraphRAG Methods
– LLMs for Tabular Data
– Automated Agentic Systems
– Enhancing Robustness in LLMs
– Controllable Text Generation for LLMs
– LLM Pruning and Distillation in Practice
… -
Context Documents Matter More Than Expected in AI
By
–
I didn’t find it mattered much at all – context docs were the important thing
-
Midjourney killed stock photography, free guide for 24 hours
By
–
Midjourney killed stock photography! Now people generate images with Midjourney. That's why I created the Midjourney Mastery Guide. And it's free for next 24 hours after that it will be $27. Like & reply "Send" and I'll DM you the guide. (Must follow to receive it)
-
Iterative Refinement Through Multi-turn Conversation With AI
By
–
Indeed, that is an issue. It can be addressed by having a multiturn conversation with the system ("Actually I wanted the text to be green") and allowing iterative refinement/correction that way, or the user can edit the generated code after the system generates a first cut.
-
Methods for diversity and control in TinyStories generation
By
–
Exactly — you can control and monitor how weird it’s getting through the example replacement/editing. Also, sampling from a k-shot pool. All the diversity-promoting methods in TinyStories have equivalents here too.
-
Thread Demonstrating k-shot Prompting Method
By
–
I made another thread to illustrate the k-shot prompting method mentioned above: