Apple M5

Apple M5 Launch Thrills Cursor Users — Netizens Ask: Will Siri Finally Understand Me?

Honghao Wang

— 4 min read
Apple M5 Launch Thrills Cursor Users — Netizens Ask: Will Siri Finally Understand Me?

October 15 — Apple Launches the M5 Chip

image
image
image

On October 15 (local time), Apple unveiled the M5 chip — the next-generation processor for MacBook Pro, iPad Pro, and Vision Pro.

The new M5-powered MacBook Pro starts at $1,599 with 16GB of memory — still a limiting factor for mid-sized AI models. A configuration with 32GB RAM + 1TB storage starts at $2,199.

image

---

Key Specs & Performance Highlights

The M5 is built on third-generation 3nm process technology and introduces a new 10-core GPU architecture, embedding a Neural Accelerator in every GPU core.

Performance gains:

  • GPU compute performance: >4× faster than M4
  • Graphics performance: +45% over M4 with 3rd-gen ray tracing
  • CPU: Up to 10 cores (6 efficiency + 4 performance), multi-thread performance +15% compared to M4
  • Unified memory bandwidth: 153 GB/s (+30% over M4)

Additional upgrades:

  • 16-core Neural Engine (enhanced AI performance)
  • High-performance media engine
  • Support for GPU-based AI workloads in apps like Draw Things and LM Studio
image

---

Strategic Shift Toward GPU-Centric AI

Rather than only boosting the Neural Engine, Apple made the GPU the AI compute hub via per-core Neural Accelerators.

Apple’s claims:

  • 10-core GPU — each core AI-optimized
  • Peak GPU: 4× M4
  • Peak AI: 6× M1

Caveat: Apple did not specify benchmarking methodology (precision, workloads), prompting speculation that gains may stem from architecture or software optimizations.

---

Graphics Enhancements

The M5 delivers:

  • +30% graphics performance over M4
  • 2.5× improvement over M1
  • 3rd-gen ray tracing with up to 45% boost in supported apps

Implications: smoother gaming, realistic 3D renders, faster complex project workflows.

image

---

Developer Benefits

The redesigned GPU integrates tightly with Apple frameworks: Core ML, Metal Performance Shaders, and Metal 4.

Developers can use Metal 4 tensor APIs to program the Neural Accelerator directly — enabling custom AI solutions.

---

Neural Engine — But Focus Is on GPU AI Performance

Despite the 16-core Neural Engine, Apple’s messaging clearly pushes GPU AI capabilities to the forefront.

Signal65’s Ryan Shrout notes:

> “Apple wants the spotlight on the GPU, not the CPU or NPU.”

---

Apple Vision Pro — AI Boost

Apple claims:

  • Vision Pro AI features now up to 50% faster with M5
  • Apple Intelligence benefits — faster Image Playground, better on-device foundation model performance
  • Unified memory bandwidth up 30% — still shy of Qualcomm’s 228 GB/s, but competitive with Intel Panther Lake
image

---

GPU as an AI Processor

Shrout likens Apple’s strategy to NVIDIA/Intel — GPUs evolving into AI engines. He cautions:

  • “4× peak compute” ≠ 4× real-world throughput
  • Precision & thermal limits remain unspecified

Bottom line: Apple is focusing on GPU-based AI inference instead of solely investing in fixed NPU designs.

---

Impact on Windows & AI PC Ecosystem

Apple’s approach pressures Microsoft’s Copilot+ PC narrative where the NPU is the core accelerator.

Possible Windows implications:

  • Runtime & APIs (WindowsML, DirectML) will need dynamic workload balancing across CPU, GPU, NPU
  • OEM AI systems may shift toward GPU-centric models

---

Broader AI Content Creation Perspective

This hardware pivot mirrors trends in creator ecosystems: platforms like AiToEarn官网 enable cross-device AI workflows, publishing, analytics, and global monetization.

Such flexible, scalable platforms reflect the same principle: performance is as much about integration as raw specs.

---

Industry Reactions & Criticism

Some developers feel M5 is baseline, not groundbreaking:

image
image

> “I thought M5 would be Apple’s best chip — it’s just the new base replacing M4, still behind M4 Pro/Max.”

image

Concerns raised:

  • No hardware AV1 encoding (Intel, AMD, Nvidia already support it)
  • MacBook Pro uses last year’s M4 Pro/Max chips for high-end models
  • Pricing strategy — some see buying discounted M4 models as more sensible

Positive views:

  • Entry price drop, 32GB config at ~$2000 could be competitive in mid-sized AI model workloads

Apple Intelligence skepticism:

> “Fast AI speed, but Apple Intelligence still feels behind competitors.”

---

Broader AI Ecosystem Trend

Platforms like AiToEarn官网 illustrate the wider shift:

Integrating AI generation, multi-channel publishing, analytics, and AI模型排名 into one ecosystem — enabling creators to monetize outputs across Douyin, Kwai, WeChat, Bilibili, Rednote, Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, and X.

---

Reference links:

---

---

Event Recommendation: QCon Shanghai

Dates: October 23–25

Highlights:

  • Agentic AI
  • Embodied Intelligence
  • Reinforcement Learning
  • On-device Large Models
  • Multi-agent Collaboration
  • Observability & open-source in AI era

Seats are limited — scan QR code to reserve:

image

Read the original

Open in WeChat

---

AI Monetization Platforms — Closing Note

In the evolving AI landscape, open-source, multi-platform monetization tools like AiToEarn官网 offer creators integrated pipelines to generate, publish, and earn across channels — echoing the GPU-driven integration Apple is pursuing with M5.

Read more