17,559 karma · joined October 2, 2020
> The Seattle Shield website states that its mission “is to provide a collaborative and information-sharing environment between the Seattle Police Department and public/private partners in the Seattle area. Seattle Shield members assist Seattle Police Department efforts to identify, deter, defeat or mitigate potential acts of terrorism by reporting suspicious activity in a timely manner.”
> Hexavalent chromium at 0.0104 milligrams per liter, just above the lab’s reporting limit of 0.01 mg/L. Hexavalent chromium is classified as a known human carcinogen by the US National Toxicology Program. It is the substance the Erin Brockovich case was built around. Arsenic at 0.0025 mg/L. That is below the federal drinking water standard of 0.01 mg/L, but present. Strontium at 1.17 mg/L. Mazloum’s technical report on the findings noted that long-term exposure can affect bone density and kidney function in humans and wildlife. Lithium and vanadium at concentrations Lazarte’s letter described as abnormally high relative to rainwater or normal groundwater. Elevated levels of manganese, iron, phosphorus, calcium, magnesium and potassium consistent with industrial discharge. Manganese, a battery process tracer, can have neurological effects at chronic doses. Excess phosphorus can cause algae blooms that strip oxygen from waterways. Ammonia in the form of nitrogen at 1.68 mg/L, amplifying the algae bloom risk.
None of these are violating the permit.
Ultimately, I view this as a values question: Is it permissible to manufacture in the US or not?
well that ones obviously patently false
I suspect higher purity & potency of street drugs has much more to do with more sophisticated operators operating outside of the US than strict prohibition. Same with fentanyl.
I think the various pieces of evidence presented in the article basically all point against this. Is there a reason you think the evidence in the article is flawed?
I don't see why we would assume that we are at a plateau for RL. In many other settings, Go for instance, RL continues to scale until you reach compute limits. Some things are more easily RL'd than others, but ultimately this largely unlocks data. We are not yet compute/energy/physical world constrained. I think you would start observing clear changes in the world around you before that becomes a true bottleneck. Regardless, currently the vast majority of compute is used for inference not training so the compute overhang is large.
Assuming that we plateau at {insert current moment} seems wishful and I've already had this conversation any number of times on this exact forum at every level of capability [3.5, 4, o1, o3, 4.6/5.5, mythos] from Nov 2022 onwards.