LLM security tooling looks dramatically better in the Sonnet 4.7+ era
dsp_ · x · 2026-07-24
- The poster says their view has shifted: bet on the exponential and extrapolate forward.
- In the quoted reply, they describe the 3.5/4o era as a time when these security tools were “absolute dog ass.”
- With Sonnet 4.7+, they say the tools are now impossible to ignore.
- The substantive takeaway is that model improvements are materially changing the usefulness of LLM security tooling.
More from coding & agent
- NO8D adds 397 prompt cards and Krea 2 styles to its ComfyUI library pack — Suspicious_Aide2697 · 2026-07-24
- Why LLM agents need four hard boundaries before they can touch production — Innowise_ · 2026-07-24
- Google and Gemini Notebook workflows are becoming skills, but mostly for reliability and latency — tokumin · 2026-07-24
- Building Profit-Driving Agent APIs is the Current Goldmine — eptwts · 2026-07-24
- Grok Build adds workflows that fan tasks out across hundreds of agents — elonmusk · 2026-07-24
- Blaxel ships Agent Drive, a shared filesystem for AI agents — Scobleizer · 2026-07-24