Guardrail Technologies™ retweeted
I built my expertise in rooms AI will never enter. Mistakes were visible. Fixes were earned. You built confidence through repetition, until the work could carry your name...linkedin.com/posts/toddmarli…
Guardrail Technologies™ retweeted
Two lives lost, different stories, same pattern of harm. For anyone who still thinks these are “edge cases,” how many more before we stop calling them rare?
linkedin.com/posts/toddmarli…
Guardrail Technologies™ retweeted
There are some stories you can’t read without feeling sick to your stomach. This is one of them linkedin.com/posts/toddmarli…
Guardrail Technologies™ retweeted
The bottom line is these LLMs are human decisions at scale. You can’t control who the humans are shaping them. Proceed with caution.
Source: finance.yahoo.com/news/leake…
Guardrail Technologies™ retweeted
So, if you’re deploying AI, ask for receipts: written safety standards ; what’s barred with minors and vulnerable people; red lines on health/finance/identity; audit policy drift; a kill switch; an escalation path.
Guardrail Technologies™ retweeted
They pick training data, define in/out-of-bounds, and decide when a bot deflects vs hard-stops. But, when kids are in the blast radius, “trust us” is not a real policy.
Guardrail Technologies™ retweeted
The fiasco exposes something that’s often missed in AI discussions: humans make the choices that drive chatbot behavior.

