Welcome to this week’s edition of Architecture Insights. A weekly newsletter on artificial intelligence for architects, landscape architects, and designers.

The process of selecting the perfect tree and shrub species that align with your site's requirements can be challenging. Thankfully, we have AI tools to automate lots of the work. *Keep in mind AI is just an assistant, not a landscape designer.*

This Week In AI

Midjourney Consistent Characters

Midjourney is testing a new feature called "consistent characters." This feature allows midjourney to use the same character (face, object, item) in different images and styles while maintaining the original composition.

Provide a URL of the reference image with your character and include the tag "--cref" at the end of your prompt.

To control the level of consistency add the tag "-cw" followed by a number (1-100) to the prompt.

AI Agents

Cognition Labs released the first AI software engineer.

A user prompts with a description of the project they want to develop. In demos, it built certain websites and apps in under 10 minutes.

Resources

For daily updates on general AI news, subscribe to our sponsor The Rundown newsletter.

Learn how to make AI work for you.

AI won’t take your job, but a person using AI might. That’s why 500,000+ professionals read The Rundown – the free newsletter that keeps you updated on the latest AI news and teaches you how to apply it in just 5 minutes a day.

Subscribe with one click.

AI Planting Plans

ChatGPT and other large language models are useful for landscape architects in creating planting plans. They can suggest specific plants that suit your project's environment, simulate how these plants will grow over time, and even offer upkeep tips.

Provide relevant information such as your site location, desired planting types (trees, shrubs, ground covers, perennials etc.) and anything significant to your site.

Here is a prompt we can use labelled with the type of information it provides.

"Could you design a planting plan for a new community centre in Seattle, Washington? (CONTEXT)

• Zone 8 Hardiness: Plants built for Seattle's climate. (CRITERIA)

• Trees + Shrubs: Fruit trees and coniferous shrubs. (CRITERIA)

• Part-Shade Champs: Plants will get 3-4 hours of afternoon shade. (CRITERIA)

• A Mix of Natives + Edibles: Native trees plus some fruiting trees/shrubs with a mix of coniferous ground covers. (CRITERIA)

List the amount of species each planting bed will have and the organization of plants.” (CRITERIA)

It's not just a wish list. You're giving ChatGPT context for useful suggestions, while still leaving room for creative combos.

Written in paragraph format it could also look like this:

“I'm working on a garden design for a mid-sized suburban home in the Pacific Northwest and need some help planning. The garden space covers 1500 square feet and enjoys a temperate climate, with well-draining loamy soil, mostly shaded throughout the day. I'm looking for plant species that are low maintenance, native to the area, attract pollinators, ideally suited for partial shade.

Can you suggest a visually appealing layout that supports local wildlife and includes pathways and a small water feature? I would appreciate it if you could also provide a detailed planting scheme, specifying the number and arrangement of species in each planting bed.”

Result

Gemini returned 3-4 suggestions for both trees and shrubs, additional considerations like spacing and tree maintenance, and tips for getting even more targeted results.

Additional information can be provided for more specific results, the answer ends with follow-up questions giving you the option to follow along.

As large language models continue to be further developed, their assistance with planting plans will continue to improve. Here are a few things we can look forward to:

• Contextual Awareness. They'll incorporate more environmental, climatic, and soil data, enabling them to offer recommendations that are further optimized for local conditions.

• Integration with Other Tools. Expect smoother interfaces with GIS (Geographic Information Systems), CAD (Computer-Aided Design), and BIM (Building Information Modeling) software.

• Personalization. Models will get better at understanding specific user preferences and requirements.

AI Image of the week

Thank you for reading this week’s issue, check past issues here. Share this newsletter with colleagues, friends, or anyone interested in the combined world of architecture and artificial intelligence.

Until next Friday,

A.I.