Hacker News: edf13https://news.ycombinator.com/user?id=edf13Hacker News RSShttps://hnrss.org/hnrss v2.1.1Mon, 15 Jun 2026 13:29:18 +0000<![CDATA[Mythos Proves AI Safety Can No Longer Live Inside the Model]]>Article URL: https://grith.ai/blog/mythos-ai-safety-cannot-live-inside-the-model?a=0

Comments URL: https://news.ycombinator.com/item?id=48540029

Points: 1

# Comments: 0

]]>Mon, 15 Jun 2026 11:59:58 +0000https://grith.ai/blog/mythos-ai-safety-cannot-live-inside-the-model?a=0edf13https://news.ycombinator.com/item?id=48540029https://news.ycombinator.com/item?id=48540029<![CDATA[Mythos Proves AI Safety Can No Longer Live Inside the Model]]>Article URL: https://grith.ai/blog/mythos-ai-safety-cannot-live-inside-the-model

Comments URL: https://news.ycombinator.com/item?id=48518185

Points: 3

# Comments: 0

]]>Sat, 13 Jun 2026 15:23:25 +0000https://grith.ai/blog/mythos-ai-safety-cannot-live-inside-the-modeledf13https://news.ycombinator.com/item?id=48518185https://news.ycombinator.com/item?id=48518185<![CDATA[The Risk Isn't Rogue AI. It's Plausible AI]]>Article URL: https://grith.ai/blog/real-risk-isnt-rogue-ai-plausible-ai

Comments URL: https://news.ycombinator.com/item?id=48492464

Points: 2

# Comments: 0

]]>Thu, 11 Jun 2026 16:23:28 +0000https://grith.ai/blog/real-risk-isnt-rogue-ai-plausible-aiedf13https://news.ycombinator.com/item?id=48492464https://news.ycombinator.com/item?id=48492464<![CDATA[New comment by edf13 in "1-Bit Bonsai Image 4B Image Generation for Local Devices"]]>Odd… UK visitor and I get:

Website Not Allowed “⁦‪prismml.com‬⁩” is a restricted website.

]]>Mon, 01 Jun 2026 06:42:23 +0000https://news.ycombinator.com/item?id=48353367edf13https://news.ycombinator.com/item?id=48353367https://news.ycombinator.com/item?id=48353367<![CDATA[New comment by edf13 in "AI Makes Adding Features Faster – So Why Not Add Just One More?"]]>Exactly - the flow of AI assistance makes it so much easier to get caught in the one more feature trap!

]]>Mon, 25 May 2026 06:27:33 +0000https://news.ycombinator.com/item?id=48264016edf13https://news.ycombinator.com/item?id=48264016https://news.ycombinator.com/item?id=48264016<![CDATA[AI Makes Adding Features Faster – So Why Not Add Just One More?]]>Article URL: https://grith.ai/blog/just-one-more-feature?nop

Comments URL: https://news.ycombinator.com/item?id=48263974

Points: 3

# Comments: 2

]]>Mon, 25 May 2026 06:20:21 +0000https://grith.ai/blog/just-one-more-feature?nopedf13https://news.ycombinator.com/item?id=48263974https://news.ycombinator.com/item?id=48263974<![CDATA[AI Makes Adding Features Faster – So Why Not Add Just One More?]]>Article URL: https://grith.ai/blog/just-one-more-feature

Comments URL: https://news.ycombinator.com/item?id=48193471

Points: 2

# Comments: 0

]]>Tue, 19 May 2026 14:03:57 +0000https://grith.ai/blog/just-one-more-featureedf13https://news.ycombinator.com/item?id=48193471https://news.ycombinator.com/item?id=48193471<![CDATA[New comment by edf13 in "Vibe Coding Still Needs a Senior Engineer (For Now)"]]>Most of the AI-security discourse (and most of my posts) right now is about prompt injection and agent hijacking. But there are still the move-fast-break-things issues that are exacerbated with agentic coding/vibe coding...

I reviewed a colleague's vibe-coded internal tool last week, found 28 security issues, and none of them were that kind of bug - they were the same classic stuff juniors have always shipped, just produced at much higher throughput.

Wrote it up because the "senior engineer review" step quietly disappeared from a lot of AI-assisted workflows, and the bugs that step used to catch are still there (We are still needed!).

]]>Mon, 11 May 2026 13:03:54 +0000https://news.ycombinator.com/item?id=48094454edf13https://news.ycombinator.com/item?id=48094454https://news.ycombinator.com/item?id=48094454<![CDATA[Vibe Coding Still Needs a Senior Engineer (For Now)]]>Article URL: https://grith.ai/blog/security-bugs-ai-doesnt-write

Comments URL: https://news.ycombinator.com/item?id=48094271

Points: 5

# Comments: 1

]]>Mon, 11 May 2026 12:46:52 +0000https://grith.ai/blog/security-bugs-ai-doesnt-writeedf13https://news.ycombinator.com/item?id=48094271https://news.ycombinator.com/item?id=48094271<![CDATA[New comment by edf13 in "Before GitHub"]]>Ha ha... yes... that brings back memories!

]]>Wed, 29 Apr 2026 08:51:16 +0000https://news.ycombinator.com/item?id=47945774edf13https://news.ycombinator.com/item?id=47945774https://news.ycombinator.com/item?id=47945774<![CDATA[New comment by edf13 in "Before GitHub"]]>My first was this monster:

https://en.wikipedia.org/wiki/Microsoft_Visual_SourceSafe

]]>Wed, 29 Apr 2026 08:33:12 +0000https://news.ycombinator.com/item?id=47945630edf13https://news.ycombinator.com/item?id=47945630https://news.ycombinator.com/item?id=47945630<![CDATA[Five AI Agent Failures in 36 Days. Zero Times the Agent Caught It]]>Article URL: https://grith.ai/blog/36-days-5-ai-agent-security-failures-0-self-detections

Comments URL: https://news.ycombinator.com/item?id=47935638

Points: 3

# Comments: 1

]]>Tue, 28 Apr 2026 15:03:10 +0000https://grith.ai/blog/36-days-5-ai-agent-security-failures-0-self-detectionsedf13https://news.ycombinator.com/item?id=47935638https://news.ycombinator.com/item?id=47935638<![CDATA[New comment by edf13 in "Bitwarden CLI compromised in ongoing Checkmarx supply chain campaign"]]>Manually review the package and override the setting

]]>Thu, 23 Apr 2026 18:25:57 +0000https://news.ycombinator.com/item?id=47879480edf13https://news.ycombinator.com/item?id=47879480https://news.ycombinator.com/item?id=47879480<![CDATA[New comment by edf13 in "Parallel agents in Zed"]]>Anyone know of a similar tool to conductor for Linux?

]]>Thu, 23 Apr 2026 07:07:06 +0000https://news.ycombinator.com/item?id=47872942edf13https://news.ycombinator.com/item?id=47872942https://news.ycombinator.com/item?id=47872942<![CDATA[New comment by edf13 in "CrabTrap: An LLM-as-a-judge HTTP proxy to secure agents in production"]]>Hand written I’m afraid… regular comments on this topic is true - it’s an area I’m very interested in.

]]>Wed, 22 Apr 2026 07:51:20 +0000https://news.ycombinator.com/item?id=47860431edf13https://news.ycombinator.com/item?id=47860431https://news.ycombinator.com/item?id=47860431<![CDATA[New comment by edf13 in "CrabTrap: An LLM-as-a-judge HTTP proxy to secure agents in production"]]>Yes, true ;)

]]>Wed, 22 Apr 2026 07:50:22 +0000https://news.ycombinator.com/item?id=47860415edf13https://news.ycombinator.com/item?id=47860415https://news.ycombinator.com/item?id=47860415<![CDATA[New comment by edf13 in "CrabTrap: An LLM-as-a-judge HTTP proxy to secure agents in production"]]>We are Open Source… code will be published soon (before launch)

]]>Wed, 22 Apr 2026 05:00:53 +0000https://news.ycombinator.com/item?id=47859235edf13https://news.ycombinator.com/item?id=47859235https://news.ycombinator.com/item?id=47859235<![CDATA[New comment by edf13 in "The Vercel Breach Needed Malware. The Next One Needs a Bad Readme"]]>Yes, fair point.

Feedback accepted, thanks!

]]>Tue, 21 Apr 2026 05:44:59 +0000https://news.ycombinator.com/item?id=47845012edf13https://news.ycombinator.com/item?id=47845012https://news.ycombinator.com/item?id=47845012<![CDATA[The Vercel Breach Needed Malware. The Next One Needs a Bad Readme]]>Article URL: https://grith.ai/blog/next-vercel-breach-ai-coding-agent

Comments URL: https://news.ycombinator.com/item?id=47844929

Points: 1

# Comments: 3

]]>Tue, 21 Apr 2026 05:32:34 +0000https://grith.ai/blog/next-vercel-breach-ai-coding-agentedf13https://news.ycombinator.com/item?id=47844929https://news.ycombinator.com/item?id=47844929<![CDATA[New comment by edf13 in "Claude Opus 4.7"]]>Related: https://news.ycombinator.com/item?id=47803847

]]>Fri, 17 Apr 2026 09:02:11 +0000https://news.ycombinator.com/item?id=47803907edf13https://news.ycombinator.com/item?id=47803907https://news.ycombinator.com/item?id=47803907