What does the LaFoto journal cover?

Prompting technique, how diffusion models actually behave, and practical AI photography — the material that does not fit on a product page. Guides are written to be usable at the keyboard rather than to rank and disappoint.

How do I write better AI image prompts?

Order matters more than length. Subject first, then setting, then light, then optics — and cut adjectives that describe feeling rather than appearance. "Cinematic" moves the image less than "85mm, f/1.8, single warm key from camera left" does.

How can I tell whether a photo is AI-generated?

Look at hands, teeth, jewellery clasps, background text and where two patterns meet — high-frequency detail is where models still lose coherence. Metadata is unreliable: it survives generation inconsistently and is trivially stripped either way.

What is the difference between an AI photo generator and an AI photo editor?

A generator makes an image that did not exist from a description. An editor changes an image that already exists, which means preserving everything you did not ask to change. They use related technology and solve genuinely different problems.

Are the guides updated when models change?

The ones tied to specific model behaviour are, because that behaviour moves. Each post carries a reviewed date so you can judge whether advice about a particular model still applies.

Do I need a GPU to generate AI images?

Not for hosted tools, which run the model on their own hardware. You need one only to run open models like Stable Diffusion locally, where the practical floor is around 8GB of VRAM and the real cost is the setup time rather than the card.

The LaFoto Journal

Guides, comparisons, and field notes on AI photo generation.

Data-rich, source-cited writing on AI photo and image generation — for creators, marketers, and designers making real decisions.

AI photography is mostly a vocabulary problem. The models respond precisely to photographic and diffusion terms — focal length, key light, denoise strength, seed — and vaguely to everything else. These guides teach the vocabulary that actually moves an image, then show where each term stops working.

Where to start with AI image generation

If you have never written a prompt, start with prompt structure — subject, setting, light, optics, in that order — because ordering does more for a result than length does. If your images already come back close but wrong, the useful next step is understanding seeds and denoise strength, which is what lets you change one variable and see what it did.

If you are choosing between tools rather than learning to use one, the comparisons and the tested roundup answer that faster than any guide here will.

Featured

Data

AI Photo Generation Statistics in 2026

Market size, adoption, image volume — and the unsettling finding that people can no longer reliably tell AI photos from real. Every figure sourced and dated.

13 min readRead article

All articles

Comparison

Best AI Photo Generators in 2026 (Ranked)

A trustworthy, sourced ranking of the leading AI photo generators — Midjourney, GPT Image, Firefly, Stable Diffusion — with honest pros, cons, and which fits you.

12 min read Guide

How to Write AI Photo Prompts That Actually Work

The anatomy of a strong prompt, copy-ready examples for portraits, products, landscapes and more, and a weak-vs-strong rewrite table you can steal.

10 min read Guide

The Complete Guide to AI Photo Editing in 2026

Background removal, object removal, upscaling, restoration, style transfer and more — what AI photo editing can do, how it compares to Photoshop, and a clean workflow.

11 min read Decision

AI Headshots vs. Professional Photography

Spend $250 on a photographer or a fraction on AI? An honest look at cost, turnaround, and quality — and exactly when each one wins.

9 min read Guide

Text to Image: How AI Turns Words Into Photos (2026 Guide)

From a typed sentence to a finished photo: how text to image actually works, the diffusion pipeline in plain English, and how to steer an AI image generator with words.

11 min read Guide

How to Get Photorealistic AI Images: 9 Prompt Techniques

Nine concrete techniques — camera language, real lighting, deliberate imperfection, and smart negative prompts — that push an AI image generator from obviously fake to genuinely photorealistic.

10 min read

The journal

AI photography and prompting questions

What does the LaFoto journal cover?: Prompting technique, how diffusion models actually behave, and practical AI photography — the material that does not fit on a product page. Guides are written to be usable at the keyboard rather than to rank and disappoint.
How do I write better AI image prompts?: Order matters more than length. Subject first, then setting, then light, then optics — and cut adjectives that describe feeling rather than appearance. "Cinematic" moves the image less than "85mm, f/1.8, single warm key from camera left" does.
How can I tell whether a photo is AI-generated?: Look at hands, teeth, jewellery clasps, background text and where two patterns meet — high-frequency detail is where models still lose coherence. Metadata is unreliable: it survives generation inconsistently and is trivially stripped either way.
What is the difference between an AI photo generator and an AI photo editor?: A generator makes an image that did not exist from a description. An editor changes an image that already exists, which means preserving everything you did not ask to change. They use related technology and solve genuinely different problems.
Are the guides updated when models change?: The ones tied to specific model behaviour are, because that behaviour moves. Each post carries a reviewed date so you can judge whether advice about a particular model still applies.
Do I need a GPU to generate AI images?: Not for hosted tools, which run the model on their own hardware. You need one only to run open models like Stable Diffusion locally, where the practical floor is around 8GB of VRAM and the real cost is the setup time rather than the card.

What the journal covers

Focus: Prompting, models, practical AI photography
Written by: The people who ran the tests
Model advice: Dated, because behaviour changes
Sponsored posts: None
Corrections: Made on the page
Reviewed: Re-checked when models change

Why these guides exist

AI photography is mostly a vocabulary problem

Diffusion models respond precisely to photographic and technical terms — focal length, key direction, denoise strength, seed — and vaguely to everything else. Most disappointing results come from prompts written in adjectives rather than in physical description.

These guides teach the terms that actually move an image, and are equally explicit about where each one stops working. Knowing the limit of a technique saves more time than knowing the technique.

An open large-format book of photographic plates on warm paper, pages held flat by a brass weight

Start somewhere

Where to begin, by what you already know

Start with structure

Prompt order does more for a result than prompt length. Subject, then setting, then light, then optics — and cut every adjective that describes a feeling rather than an appearance.