Pipeline run for this draft
Generated note
The video involves 9 people, so it cannot show all stalls are tracking error. Self-reported deficits can be inaccurate; plateaus may involve adaptive thermogenesis, lower expenditure and hunger changes. https://x.com/i/status/2055697411301925107 https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2019.00850/full https://www.ncbi.nlm.nih.gov/books/NBK576400/
Source post
"I ate in a calorie deficit but didn't lose weight." No, you did not. Jeremy Ethier ran a small metabolism experiment using astronaut-grade accuracy. What he found was that out of 9 people, not one had a fast or slow metabolism. Everyone burned almost exactly what their body weight predicted. The most interesting part he found was that most people underestimated how much they ate by 500–1,000+ calories a day. A skinny guy who couldn't gain weight guessed 2,000. He was actually eating 3,300. An overweight girl who wanted to lose weight guessed 1,800. She was at 3,000. Not one had an abnormal metabolism. Every single one had a tracking problem. Studies actually show people underestimate their daily calorie intake by 20–50%, with the gap widest among those who need accuracy most. People are not broken, but their tracking is. Fix that and the math takes care of itself.
Pipeline steps
16 steps- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✓
- ✕
generate_note.pre_filter
- Started
- May 16, 2026, 19:43:23
- Finished
- May 16, 2026, 19:43:29
- Duration
- 5.71 s
Input snapshot
{
"post_text": "[Target Post]\n\"I ate in a calorie deficit but didn't lose weight.\"\n\nNo, you did not.\n\nJeremy Ethier ran a small metabolism experiment using astronaut-grade accuracy.\n\nWhat he found was that out of 9 people, not one had a fast or slow metabolism. Everyone burned almost exactly what their body weight predicted.\n\nThe most interesting part he found was that most people underestimated how much they ate by 500–1,000+ calories a day.\n\nA skinny guy who couldn't gain weight guessed 2,000. He was actually"
}Output snapshot
{
"has_factual_claims": true
}