AI - image generation using new powerful Mac or generate online?

bigbigJim

Registered Member
Joined
Nov 16, 2023
Messages
70
Reaction score
29
Hi,

Looking to start with AI image generation to create designs for selling. Used my M1 8gb Mac mini and it's holding up although I understand more intense AI work requires a more powerful machine. Do I need to invest in more capable hardware such as an M4 Mac Studio or is it possible & economical to do this using an online service. If anyone can offer some light on this for a newbie I'd very much appreciate it. Thanks.
 
For local generation you need Nvidia GPU with a lot of VRAM, both AMD and Intel are unreliable for AI stuff, RTX 3060 being the bare minimum. Online generators are pretty crap, but for simple stuff they might just work.
 
For local generation you need Nvidia GPU with a lot of VRAM, both AMD and Intel are unreliable for AI stuff, RTX 3060 being the bare minimum. Online generators are pretty crap, but for simple stuff they might just work.
That's all I needed. Appreciate it thanks.
 
Hi,

Looking to start with AI image generation to create designs for selling. Used my M1 8gb Mac mini and it's holding up although I understand more intense AI work requires a more powerful machine. Do I need to invest in more capable hardware such as an M4 Mac Studio or is it possible & economical to do this using an online service. If anyone can offer some light on this for a newbie I'd very much appreciate it. Thanks.
Renting a gpu online is the best and cheapest option you have available. There are places like vast.ai or Runpod where you can run a low powered gpu online for a couple cents/hr. Makes sense if you do not have to run a high powered gpu 24/7
 
It will depends on your needs.
For example, if I need to generate photos just to put on blog posts, I wouldn't buy a new laptop just for that.
But if you need to do something else that requires super high quality photos, it may be necessary to have a powerful hardware to create higher quality photos
 
What's your opinion on a top Mac vs an online generator?
its completely up to your current needs and operation still with latest mac you cant scale or get high quality images without investing in gpu if you can continue your work with online generators then i would say just keep going without further investments If you have any more doubts or questions pm me ill get back to you!
 
you need a powerful mac to run heavy AI jobs, m2 will do fine but you are way behind in tech level, m3 will be your mid point or go all in for the m4 this keeping in mind how apple just discard software and computers easily so you can stay more time in a middle ground without the rush to update again or run behind with the AI updates at long run is better you upgrade your hardware for the long run.
 
Top Mac (M3/M4 series) vs. Online AI Generators
If you consider Mac Studio (M3 Ultra) or the future M4 series for AI generation, their performance is indeed strong, but they are still not as good as NVIDIA RTX GPUs in the field of AI generated images, because Apple GPUs do not support CUDA, and most AI models (such as Stable Diffusion, Llama) are still mainly optimized for NVIDIA's architecture.

Mac Studio (M3 Ultra or future M4)
✅ Pros:

High performance: M3 Ultra (76-core GPU) has good speed within the Apple ecosystem and is suitable for ordinary AI tasks.
Local AI: Supports Core ML + Metal to run Stable Diffusion, but not as fast as RTX 4090.
Energy-saving and low noise: quieter than traditional PC workstations.
❌ Cons:

No support for CUDA, resulting in slower AI generation and incompatibility of some models/plugins.
Limited VRAM: Even though the M3 Ultra is powerful, its 128GB unified memory is still not as AI-friendly as the RTX 4090's 24GB dedicated VRAM.
Expensive: Much more expensive than a PC with the same performance.
Online AI generators (Midjourney, DALL·E, Leonardo, Runway)
✅ Pros:

Zero hardware investment, suitable for beginners or lightweight work.
Excellent pre-trained models (Midjourney, DALL·E strong stylization).
No technical experience required, directly generate with the web interface.
❌ Cons:

Not customizable, can't train your own style model (such as LoRA, DreamBooth).
High generation cost, long-term use is more expensive than buying a GPU (high-quality models require subscription).
Limited by the platform, such as Midjourney can't adjust seed parameters, DALL·E generation quality is limited.
The best choice depends on your needs
1️⃣ If you are a professional designer who wants to customize AI generation and do large-scale creation
➡ It is recommended to buy an RTX 4090-level PC, or rent a cloud GPU (RunPod/Vast.AI). The AI generation efficiency of the Mac ecosystem is still not fast enough, and the RTX 4090 is far ahead.

2️⃣ If you are an ordinary user who only does occasional AI generation
➡ Online AI generators + Mac are a good choice, such as Midjourney, DALL·E, Leonardo. You don’t need to invest in hardware, suitable for non-technical users.

3️⃣ If you must use a Mac and want to run local AI
➡ M3 Ultra is the best Mac solution at present, but the AI generation speed is still inferior to RTX 4090. You can use **Stable Diffusion WebUI (Core ML version)** or Diffusers, but it is not suitable for efficient professional creation.

Conclusion
Top-end Mac models (such as the M3 Ultra) are suitable for general AI tasks, but not for high-end AI-generated images.
If AI generation is your core business, it is recommended to choose an RTX 4090-class PC or cloud GPU.
Online AI generators are suitable for entry-level or non-technical users, but have limited customization capabilities and may cost more than the hardware.
 
Top Mac (M3/M4 series) vs. Online AI Generators
If you consider Mac Studio (M3 Ultra) or the future M4 series for AI generation, their performance is indeed strong, but they are still not as good as NVIDIA RTX GPUs in the field of AI generated images, because Apple GPUs do not support CUDA, and most AI models (such as Stable Diffusion, Llama) are still mainly optimized for NVIDIA's architecture.

Mac Studio (M3 Ultra or future M4)
✅ Pros:

High performance: M3 Ultra (76-core GPU) has good speed within the Apple ecosystem and is suitable for ordinary AI tasks.
Local AI: Supports Core ML + Metal to run Stable Diffusion, but not as fast as RTX 4090.
Energy-saving and low noise: quieter than traditional PC workstations.
❌ Cons:

No support for CUDA, resulting in slower AI generation and incompatibility of some models/plugins.
Limited VRAM: Even though the M3 Ultra is powerful, its 128GB unified memory is still not as AI-friendly as the RTX 4090's 24GB dedicated VRAM.
Expensive: Much more expensive than a PC with the same performance.
Online AI generators (Midjourney, DALL·E, Leonardo, Runway)
✅ Pros:

Zero hardware investment, suitable for beginners or lightweight work.
Excellent pre-trained models (Midjourney, DALL·E strong stylization).
No technical experience required, directly generate with the web interface.
❌ Cons:

Not customizable, can't train your own style model (such as LoRA, DreamBooth).
High generation cost, long-term use is more expensive than buying a GPU (high-quality models require subscription).
Limited by the platform, such as Midjourney can't adjust seed parameters, DALL·E generation quality is limited.
The best choice depends on your needs
1️⃣ If you are a professional designer who wants to customize AI generation and do large-scale creation
➡ It is recommended to buy an RTX 4090-level PC, or rent a cloud GPU (RunPod/Vast.AI). The AI generation efficiency of the Mac ecosystem is still not fast enough, and the RTX 4090 is far ahead.

2️⃣ If you are an ordinary user who only does occasional AI generation
➡ Online AI generators + Mac are a good choice, such as Midjourney, DALL·E, Leonardo. You don’t need to invest in hardware, suitable for non-technical users.

3️⃣ If you must use a Mac and want to run local AI
➡ M3 Ultra is the best Mac solution at present, but the AI generation speed is still inferior to RTX 4090. You can use **Stable Diffusion WebUI (Core ML version)** or Diffusers, but it is not suitable for efficient professional creation.

Conclusion
Top-end Mac models (such as the M3 Ultra) are suitable for general AI tasks, but not for high-end AI-generated images.
If AI generation is your core business, it is recommended to choose an RTX 4090-class PC or cloud GPU.
Online AI generators are suitable for entry-level or non-technical users, but have limited customization capabilities and may cost more than the hardware.
Super helpful thank you.
 
Back
Top