California's Safe Harbor level for lead is 0.5 µg/day. The CDC's safe level is 10 µg/day, and was 25 µg/day from 1985 to 1991. 12−25 times 0.5 is 6−12.5 µg, which is basically within the CDC's safe level, and was only found in two samples. (Also, as Soylent's own reply pointed out, they tested version 1.5, and 2.0 has a different recipe with even lower—but still safe—levels.)
As You Sow has also found lead and cadmium levels above California's Safe Harbor threshold in 26 chocolate products, including Ghirardelli, Hershey, Mars, Trader Joe's, and Whole Foods...
With SumatraPDF 3.0 on Windows 8.1 x64, the links in the PDF version do not show up. With Adobe Reader 11 on Windows 7 x86, they look fine. On the other hand, SumatraPDF can also handle the MOBI and EPUB versions.
...When confronting something which may be either a windmill or an evil giant, what question should you be asking? There are some who ask, "If we do nothing, and that is an evil giant, can we afford to be wrong?" These people consider themselves to be brave and vigilant. Some ask, "If we attack it wrongly, can we afford to pay to replace a windmill?" These people consider themselves cautious and pragmatic. Still others ask, "With the cost of being wrong so high in either case, shouldn't we always definitively answer the 'windmill vs.
I took the survey. The BSRI reminds me of the MBTI, though, in that the questions are vague and I would probably give different answers depending on various factors, like what time it is or whom I've interacted with recently.
Made an account here to comment that I filled out the survey, and to make future participation more likely.
Tay doesn't tell us much about deliberate Un-Friendliness. But Tay does tell us that a well-intentioned effort to make an innocent, harmless AI can go wrong for unexpected reasons. Even for reasons that, in hindsight, are obvious.
Are you sure that superintelligent AIs would have a "correct ontology/semantics"? They would have to have a useful one, in order to achieve their goals, but both philosophers and scientists have had incorrect conceptualizations that nevertheless matched the real world closely enough to be productive. And for an un-Friendly AI, "productive" translates to "using your atoms for its own purposes."