### System details
CPU: Intel Core i5-8600
GPU: NVIDIA GeForce RTX 2060
RAM: 16GB
Kernel: 6.17.9-arch1-1
OS: Omarchy (Arch Base)

### What's wrong?
**Description**
I am experiencing random hard freezes while using Omarchy. The screen freezes completely, and neither the mouse nor the keyboard responds. I am forced to perform a hard shutdown (holding the power button) to recover. This has happened multiple times recently.

**Steps to Reproduce**
1. Log in to the desktop session.
2. Run a heavy multitasking workload (e.g., 2 Docker Compose projects, Antigravity AI, Vitest, and Chrome with many tabs).
3. System freezes randomly without a clear trigger.

**Observations**
- No specific app seems to trigger it immediately, but it happens under load.
- This appears to be a regression as it affects multiple machines.

### Logs from Freeze
I captured the logs from the previous boot (`journalctl -b -1`) which ended in a hard freeze.

**1. Critical GPU Error at Boot (Current Boot)**
I am seeing Nvidia I2C timeouts immediately upon rebooting after the crash, suggesting the GPU or driver is in an unstable state.
```
Jan 04 18:37:32 omarchy kernel: nvidia-gpu 0000:01:00.3: i2c timeout error e0000000
```

**2. Hardware Lockup Warning (~15 mins before freeze)**
The kernel reported a CPU interrupt stuck for too long. This "perf interrupt" warning is often a precursor to hard lockups and points to a kernel/driver stall (likely GPU related).
```
Jan 04 15:55:55 omarchy kernel: perf: interrupt took too long (3180 > 3162), lowering kernel.perf_event_max_sample_rate to 62000
```

**3. Memory Exhaustion Context (~40 mins before freeze)**
I experienced a significant OOM event where `vitest` consumed ~11GB RAM and was killed.
```
Jan 04 15:28:48 omarchy kernel: oom-kill:constraint=CONSTRAINT_NONE... task=node (vitest),pid=144756,uid=1000
Jan 04 15:30:16 omarchy systemd[1239]: app-Hyprland-antigravity-8095e96d.scope: Consumed ... 10.9G memory peak
```
*Note: While my system was under heavy load (Antigravity + Docker + Chrome + Vitest), the system successfully killed the memory-hogging process (Vitest). The hard freeze occurred ~40 minutes later, suggesting the OOM was a stressor but the root cause of the lockup is likely the GPU driver stability under this load.*

### Assessment
This appears strongly related to **#3899** (Chromium GPU/Ozone errors).
- I am running a heavy Chrome session (which uses GPU).
- I am on an Nvidia GPU (RTX 2060).
- I am seeing GPU timeout errors and kernel stalls.

It seems the Nvidia driver regression (referenced in #3899) is causing full system lockups when the GPU is stressed by Chrome + general system load, rather than just crashing the browser.