2 points, 0 comments on Hacker News
Source: [Hacker News](https://zenodo.org/records/22235396)
📰 Read Full Story
This is an aggregated headline summary. For the complete report, visit the original publisher.
Continue Reading at Hacker News ↗More Headlines
Show HN: Wife: AI that lets you pick the best Card for each Purchase
Hi Hackernews, being an avid maximizer of rewards/gifts/credit cards, I decided to build a product that enables me to select the best card for each purchase/transaction.
Anthropic launches Enterprise Frontier Safeguards for Claude
Developing story — details emerging. Check the source link for the latest updates.
Anthropic upgrades Claude with new Fable 5.1 model, details here
In June, Anthropic first released Fable 5, its Mythos-class AI model for Claude. Three months later, Anthropic is enhancing Claude with its new Fable 5. 1 model upgrade.
Claude Fable 5.1 and Claude Mythos 5.1
What's new in Claude Fable 5. 1 – https://platform. claude.
I Built an AI That Rewrites Its Own Prompts — Its Safety Gate Rejected Every Single Edit
AgentSelfEdit is an open-source sidecar that rewrites its own system prompt from execution feedback. It A/B tests edits and promotes only statistically-proven winners. Code: github.
Show HN: Semantic Overlays – an NX bit for LLM prompt injection (live demo)
I've built a new method for steering LLMs called Semantic Overlays, small trained adapters on a frozen model which change how it perceives a piece of its context. The most readily applicable usage is to mitigate prompt injection, and it lets us take a very-injectable Qwen-3. 5-9B to SOTA scores ...