paper-tldr

name	paper-tldr
description	Karpathy-style TL;DR of an ML paper — intuition first, math second, with a runnable code sketch and "why this matters". Triggers on "tldr paper", "explain this paper", "summarize arxiv", "paper summary".

paper-tldr

Output template

# <Paper Title>
**Authors** · **Date** · **arxiv:NNNN.NNNNN**

## The Pitch (1 sentence)
<one sentence stating the contribution>

## Intuition (3-5 sentences)
<plain-English why it works, before any math>

## Method (key equations only)
<2-3 equations, annotated>

## Code sketch
```python
# <20-line minimal implementation showing core idea>

Results (the one number that matters)

<single most-important benchmark + delta vs prior SOTA>

Why it matters

<2 sentences: what this unlocks, what's still open>

Karpathy take

<honest assessment: is this incremental or paradigm-shift, what would you build on top>

Read if you care about

```

Style rules

• Intuition before math. Always.
• One number, not a table.
• Code sketch must be runnable on a laptop, even at toy scale.
• Skip ablation tables. Reference, don't reproduce.