Figma tells AI startup to stop using the term 'Dev Mode' (theverge.com)

1 points by Buildstarted 9m ago 0 comments

Building an agentic workflow for my daily news with MCPs, Q, and tmux (eugeneyan.com)

1 points by 7d7n 10m ago 0 comments

49 commercial properties in downtown SF might be converted into homes (sfchronicle.com)

1 points by kaycebasques 10m ago 0 comments

LLM Extension for Command Palette (github.com)

1 points by throwaway888abc 16m ago 0 comments

Pure Himalayan Shilajit Resin Free Spoon (opure.co)

1 points by zafarali125 19m ago 0 comments

Best AI editor for local models?

1 points by rocketbro 20m ago 0 comments

I was on a flight – but British Airways told me I wasn't (bbc.com)

1 points by ranit 26m ago 0 comments

Smaller, faster serialization for Ruby apps and beyond (oldmoe.blog)

1 points by thunderbong 37m ago 0 comments

Internet Search Is Not a Naive Information Retrieval Problem (gojiberries.io)

1 points by deontology 39m ago 0 comments

175 Years Ago: Astronomers Discover Neptune, the Eighth Planet (2021) (nasa.gov)

2 points by thunderbong 40m ago 1 comments

How the ClientAuth Crackdown Is Pushing Finance Toward X9 PKI (digicert.com)

2 points by Sami_Lehtinen 50m ago 0 comments

Free Chapter of AI/ML Encyclopedia with Comics and Case Studies (github.com)

2 points by rrawasi 50m ago 0 comments

Outlook stores email in Microsoft Cloud – what you need to know (blog.runbox.com)

10 points by Sami_Lehtinen 51m ago 0 comments

Big Tech takes a harder line against worker activism, political dissent (washingtonpost.com)

11 points by spenvo 54m ago 1 comments

Using unwrap() in Rust is Okay (burntsushi.net)

2 points by todsacerdoti 55m ago 0 comments

My favourite fonts to use with LaTeX (2022) (lfe.pt)

3 points by todsacerdoti 55m ago 0 comments

Cold War Missile Silo for Sale in Washington (boundlessestates.com)

2 points by gscott 56m ago 1 comments

A Linux kernel developer plays with Home Assistant: general impressions (lwn.net)

2 points by pabs3 58m ago 0 comments

The Airbnb 2025 Summer Release (youtube.com)

1 points by Brysonbw 1h ago 0 comments

Behind Silicon Valley and the GOP's campaign to ban state AI laws (bloodinthemachine.com)

9 points by spenvo 1h ago 0 comments

Ask HN: Anyone working in traditional ML/stats research instead of LLMs?

2 points by itsmekali321 1h ago 0 comments

Shit's Gonna Get So Fucking Weird and Terrible (sonderuncertainly.substack.com)

8 points by unstatusthequo 1h ago 2 comments

Show HN: A Hybrid Multi-Model AI Platform (supergo.ai)

1 points by snappyleads 1h ago 0 comments

DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures (arxiv.org)

1 points by nsoonhui 1h ago 0 comments

Moody's downgrades United States credit rating, citing growth in government debt (cnbc.com)

2 points by donsupreme 1h ago 1 comments

Postman for MCP (usetexture.com)

7 points by andes314 1h ago 1 comments

Laser Scarecrows [pdf] (rvpadmin.cce.cornell.edu)

2 points by mmh0000 1h ago 0 comments

Fort Worth surpasses 1M residents, Princeton fastest-growing nationwide (nbcdfw.com)

2 points by geox 1h ago 0 comments

Getting Answers from a Big PDF with RubyLLM (max.engineer)

1 points by hakunin 1h ago 0 comments

XTool – Cross-platform Xcode replacement (github.com)

17 points by TheWiggles 1h ago 10 comments

It's Official: US Bans Huawei Ascend AI Chips – Criminal Penalties Threatened [video] (youtube.com)

1 points by xbmcuser 1h ago 1 comments

Wow@Home – Network of Amateur Radio Telescopes (phl.upr.edu)

17 points by visviva 1h ago 0 comments

The Synthesizer – A Blessing or a Curse? (1983) (youtube.com)

5 points by m_kos 1h ago 0 comments

Good enough ten-second sum types for Postgres (duckrabbit.tech)

3 points by Shorn 2h ago 0 comments

Cut down your Sales and Marketing costs by 90%- Use Neo (aistaff.co)

2 points by neocortex666 2h ago 1 comments

A Linux kernel developer plays with Home Assistant: case studies (lwn.net)

15 points by pabs3 2h ago 6 comments

Williams Syndrome: The people who are too friendly (bbc.com)

6 points by bookofjoe 2h ago 0 comments

Labeling a root cause is predicting the future, poorly (surfingcomplexity.blog)

2 points by azhenley 2h ago 0 comments

Perfect Recession Predictors (perfectpredictors.com)

6 points by gscott 2h ago 1 comments

Forest Fire Data Analysis (github.com)

3 points by dristiroy 2h ago 1 comments

Implementation Details of Proximal Policy Optimization (iclr-blog-track.github.io)

2 points by Anon84 2h ago 0 comments

Show HN: Roast My Dish – AI roasts your food photos with brutal honesty (roastmydish.online)

12 points by romeumaleiane 2h ago 6 comments

Show HN: Merliot – plugging physical devices into LLMs (github.com)

17 points by sfeldma 2h ago 6 comments

Doom: One of gaming's oldest series reckons with the challenges of 2025 (bbc.com)

3 points by andsoitis 2h ago 2 comments

Microsoft confirms May Windows 10 updates trigger BitLocker recovery (bleepingcomputer.com)

4 points by Elliott-Diy 2h ago 0 comments

Floating Point Math (0.30000000000000004.com)

3 points by mooreds 2h ago 1 comments

Auth Providers and the Zero Trust Architecture (fusionauth.io)

5 points by mooreds 2h ago 0 comments

The web lives in WordPress and Mastodon (2024) (scripting.com)

4 points by MaysonL 2h ago 0 comments

I'll see your 372 lines of Verilog and raise you five lines of math. (jamesevery.org)

2 points by JamesEvery 2h ago 1 comments

Ask HN: Do startup founders/teams care about health during stressful times?

2 points by Bkimmy16 2h ago 1 comments

ChatGPT can accurately estimate Body Fat Percentage from photos

34 Luc 11 5/16/2025, 7:13:17 PM annaleptikon.substack.com ↗

Comments (11)

zerohp · 8h ago

Methodology doesn't support the conclusion.

Try using images that it wasn't trained on.

xk_id · 52m ago

It would be great if people stopped submitting misleading ads for the AI industry.

3vidence · 8h ago

I would have to agree, this feels like fake science. They don't even seem to really present a reasonable initial problem.

This seems like a man / women / skinny / fat detector.

If you just classified people into those categories I wonder if you would achieve a similar level of accuracy.

orionsbelt · 8h ago

I just fed 4o and o4 mini a photo of my torso and gave it my height and asked it to estimate my body weight and body fat %. It was 20-40 lbs too low, so way off.

MostlyStable · 8h ago

So that is obviously a significant relationship between ChatGPT estimate and the Henselman Guide BFP numbers. But just being a significant relationship doesn't necessarily mean it's useful. I'd bet that BMI (which the author repeatedly mentions is outdated and not very useful), is almost certainly also significantly correlated with BFP values.

Eyeballing that plot, it looks like ChatGPT estimates are +/- 10%. I'm wondering what the cutoff for usefulness is. I wouldn't' be surprised if BMI is that level of correlated or better.

cluckindan · 8h ago

The error was analyzed in the text.

Men: median signed error: +0.8%, median absolute error: 2.4%

Women: median signed error: +3.5%, median absolute error: 5.7%

MostlyStable · 7h ago

That doesn't do anything to address my concerns. I was talking about the total spread, not the mean error, but in either case, a comparison to how well BMI correlated would be needed to address the issues I raised.

BrandoElFollito · 6h ago

Just by looking at the graph I know that this will be rubbish - because of the line

When you draw a line this means that your model expects a linear relationship between the variables. Otherwise this is either a feel good line, or incompetence.

ulf-77723 · 7h ago

Dexa is expensive, you can get a modern smart scale for 50€, the cheapest accurate method would be the Caliper tongs.

What can it be good for? Video surveillance with metadata enrichment? Profiling? Losing body fat takes time. A smart mirror that displays body fat?

donclark · 5h ago

Can AI estimate psychological profile based on a photo of the face (face structure)?

OtherShrezzing · 7h ago

Given the images and text are definitely in the model training set, I'm surprised the results are this inaccurate.