Haha that answer it gave you isn't actually correct: it CAN run arbitrary shell commands but you have to convince it to use Python's http://
subprocess.run(…, shell=True) You can even get it to run PHP or Deno or Lua if you know what you're doing https://
til.simonwillison.net/llms/code-inte
rpreter-expansions
…
SAFETY
-
LLM Code Execution Vulnerabilities: Shell Commands and Arbitrary Code
By
–
-
Beware of LLM Misinformation: Critical Thinking Essential
By
–
liars and clowns anybody who cannot see that shouldn't use an LLM because they could believe any crap they're told!!!
-
Safe AI Label Contradicted by Palantir Partnership Concerns
By
–
yup or the fact that they label themselves "Safe AI" and then go work with palantir on some sketchy stuff !!!
-
Model Manipulation Risks: Quantization to Safety Degradation
By
–
not your Weights means eventually not the same Model btw they can > quantize it
> distill it
> hot-swap to a cheaper/weaker checkpoint
> make the model manipulative
> fine-tune it in ways that break safety or depth
> drop its IQ
> run experiments on you and/or your data
> -
Red lines crossed: escalation and direct action required
By
–
Toutes les lignes rouges sont dépassées. Désormais il va falloir aller au contact.
-
Enterprise AI: Why the 5% Accuracy Gap Matters Most
By
–
LLMs are brilliant ~95% of the time. But in enterprise, it’s the other 5% that matters. Even small inaccuracies can carry big risks. Our Co-Founder & Co-CEO, Professor @yshoham explains why accuracy is the blocker and why the path forward is system-level AI. Full
-
Existential concerns about advanced AI survival and mortality
By
–
Pas certain qu'il soit encore en vie à cet instant… https://t.co/E1Dwx44JUT
— Olivier Rimmel 🕊️ (@OlivierRimmel) 9 septembre 2025Pas certain qu'il soit encore en vie à cet instant…
-
AI Agents Privacy Threats: What Companies Must Do Now
By
–
NEW — In The Economist, discussing the privacy perils of AI agents and what AI companies and operating systems need to do–NOW–to protect Signal and much else! https://
economist.com/by-invitation/
2025/09/09/ai-agents-are-coming-for-your-privacy-warns-meredith-whittaker?giftId=859d3a46-a913-4ee0-a3d5-060b774c2501&utm_campaign=gifted_article
… -
Phase 5: AI as a System for Founders’ Leadership
By
–
9/ Phase 5: Leadership AI isn’t a feature. It’s a system. Founders must:
• Think in compounding systems
• Train teams in evals, safety, costs
• Sell AI internally with ROI, not hype