Uploaded May 2026 | Updated September 2026, 3 weeks ago
Last month, Linux's #2 maintainer shipped 24 patches to the kernel — all
caught by an AI running on a desktop next to him. No subscription. No
API key. No monthly bill. The kernel is the software that powers your
phone, your router, your car, and half the internet — and these were
real security bugs in real subsystems.
Same month, a different AI subscription quietly billed a developer $200
because of a filename in his git history. Same vendor that found a
16-year-old FFmpeg bug also refused to refund the $200 they had
mistakenly charged. Two visions of where AI is heading. Only one of
them is yours to keep.
This is the story of GKH's "clanker_t1000," AMD's Strix Halo, the new
Linux AI policy, Anthropic's Mythos vs HERMES.md, Zig's anti-AI ban —
all on one map, with the variable that explains who's right.
— Codacus
═══════════════════════════════════════════════
🕒 CHAPTERS
═══════════════════════════════════════════════
0:00 Intro
0:36 Who is Greg Kroah-Hartman
1:17 Clanker T1000 + the SMB bug
2:31 Why the hardware matters (AMD Strix Halo)
3:27 Two more bugs (Nouveau, IO_uring)
4:14 Why these bugs mattered
4:44 The killer quote — accountability
5:40 Subscribe (one-second plug)
5:51 The Zig contrast — opposite direction
6:34 Anthropic Mythos — cloud AI's win
7:13 The HERMES.md kicker — same vendor, two faces
8:14 The take — who's accountable for the output
8:57 What this means for you
9:25 Final thought
═══════════════════════════════════════════════
🔗 LINKS & SOURCES
• Linux kernel AI policy (official):
docs.kernel.org/process/coding-assistants.html
• Phoronix — "Clanker T1000" coverage:
phoronix.com/news/Clanker-T1000-AMD-Ryzen-AI-Max
• Tom's Hardware — Linux's #2 hunts bugs locally:
tomshardware.com/software/linux/linux-kernels-second-in-command-uses-framework-desktop-to-hunt-bugs-with-local-ai
• Anthropic Mythos / Project Glasswing announcement:
red.anthropic.com/2026/mythos-preview
• Simon Willison — Zig's anti-AI policy:
simonwillison.net/2026/Apr/30/zig-anti-ai
• HERMES.md billing flaw — Consumer Rights Wiki:
https://consumerrights.wiki/w/Anthropic_Claude_Code_HERMES.md_billing_flaw
• HN thread on the $200 charge:
news.ycombinator.com/item?id=47952722
═══════════════════════════════════════════════
⚙️ HARDWARE in this video:
AMD Ryzen AI Max+ 395 (Strix Halo) on Framework Desktop
16 Zen 5 cores · 40 RDNA 3.5 CUs · 128GB unified LPDDR5X
═══════════════════════════════════════════════
🟢 If you're tired of being the QA team for AI you didn't ask for,
subscribe — I'm doing more videos like this. The fight isn't AI
vs no-AI. It's who's accountable for the output.
#linuxkernel #localai #vulnerability #cybersecurity #StrixHalo #aisecurity #opensource #opensourceai
Last month, Linux's #2 maintainer shipped 24 patches to the kernel — all
caught by an AI running on a desktop next to him. No subscription. No
API key. No monthly bill. The kernel is the software that powers your
phone, your router, your car, and half the internet — and these were
real security bugs in real subsystems.
Same month, a different AI subscription quietly billed a developer $200
because of a filename in his git history. Same vendor that found a
16-year-old FFmpeg bug also refused to refund the $200 they had
mistakenly charged. Two visions of where AI is heading. Only one of
them is yours to keep.
This is the story of GKH's "clanker_t1000," AMD's Strix Halo, the new
Linux AI policy, Anthropic's Mythos vs HERMES.md, Zig's anti-AI ban —
all on one map, with the variable that explains who's right.
— Codacus
═══════════════════════════════════════════════
🕒 CHAPTERS
═══════════════════════════════════════════════
0:00 Intro
0:36 Who is Greg Kroah-Hartman
1:17 Clanker T1000 + the SMB bug
2:31 Why the hardware matters (AMD Strix Halo)
3:27 Two more bugs (Nouveau, IO_uring)
4:14 Why these bugs mattered
4:44 The killer quote — accountability
5:40 Subscribe (one-second plug)
5:51 The Zig contrast — opposite direction
6:34 Anthropic Mythos — cloud AI's win
7:13 The HERMES.md kicker — same vendor, two faces
8:14 The take — who's accountable for the output
8:57 What this means for you
9:25 Final thought
═══════════════════════════════════════════════
🔗 LINKS & SOURCES
• Linux kernel AI policy (official):
docs.kernel.org/process/coding-assistants.html
• Phoronix — "Clanker T1000" coverage:
phoronix.com/news/Clanker-T1000-AMD-Ryzen-AI-Max
• Tom's Hardware — Linux's #2 hunts bugs locally:
tomshardware.com/software/linux/linux-kernels-second-in-command-uses-framework-desktop-to-hunt-bugs-with-local-ai
• Anthropic Mythos / Project Glasswing announcement:
red.anthropic.com/2026/mythos-preview
• Simon Willison — Zig's anti-AI policy:
simonwillison.net/2026/Apr/30/zig-anti-ai
• HERMES.md billing flaw — Consumer Rights Wiki:
https://consumerrights.wiki/w/Anthropic_Claude_Code_HERMES.md_billing_flaw
• HN thread on the $200 charge:
news.ycombinator.com/item?id=47952722
═══════════════════════════════════════════════
⚙️ HARDWARE in this video:
AMD Ryzen AI Max+ 395 (Strix Halo) on Framework Desktop
16 Zen 5 cores · 40 RDNA 3.5 CUs · 128GB unified LPDDR5X
═══════════════════════════════════════════════
🟢 If you're tired of being the QA team for AI you didn't ask for,
subscribe — I'm doing more videos like this. The fight isn't AI
vs no-AI. It's who's accountable for the output.
#linuxkernel #localai #vulnerability #cybersecurity #StrixHalo #aisecurity #opensource #opensourceai
