3 ms·

Hiding information and general untrustworthiness harms credibility, which is a chicken that's going to come home to roost the next time there's a public health

Hiding information and general untrustworthiness harms credibility, which is a chicken that's going to come home to roost the next time there's a public health crisis. The CDC blew a lot of its credibility with the (arguably noble) early lie about masks not working. Pharma companies probably don't have that much to lose. Now it's the FDA's turn. "We must hide legally-required-to-be-available information because some people may misuse it" is a path with unlimited abuse potential, best avoided.

Robotbeat 5y ago
Look, again, I’m not suggesting any legally required information should be hidden. But if you think the end result of releasing massive amount of even totally-above-board information is going to be increased public trust, then you must have sleepwalked the last 6 years. There’s a good case to be made that releasing this is the best course. But let’s not pretend that public trust won’t be eroded because it takes an order of magnitude more work to disprove bulls### than to make it up (or twist it) in the first place. And it isn’t any particularly information I’m talking about, simply the vast quantity involved. The vaccine becoming political is probably the worst thing to happen to humanity in the last 2-3 years. Worse probably than COVID itself. EDIT: just thought of one possible solution or at least mitigation to the “Bull### asymmetry principle” when combined with massive data dumps: Limit the amount of material that can be requested. Any specific dates or whatever could be requested within the stack. Or if a representative sample is desired, the entire dump is sectioned in pieces and each section is securely hashed. The one doing the request can then request, say, up to 100-1000 sections and each such section can then have its hash compared. That ensures the random selection is random to the satisfaction of the requesting party. And it limits the amount of bulls### that needs to be countered to some reasonable level. Otherwise, any attacker will be able to overwhelm the public with endless claims from a massive stash of requested documents. This doesn’t limit the amount of work for scrubbing PII, though, since you’d have to scrub it before hashing or the hashing wouldn’t be recreatable. But at least you’d limit the amount of unique bulls### that’d need to be countered.
- Karunamon 5y ago
  The FDA could've had the policy that the data is scrubbed and made available as its ingested. And, quite honestly, I see this complaint of theirs as one made in bad faith. Once they have no other choice, they'll comply in a lot faster than 3/4 of a fucking century. Really, I see FOIA requests as failures of government; as in they represent the government being poked to be transparent when transparency should be the default.

Robotbeat 5y ago

Look, again, I’m not suggesting any legally required information should be hidden. But if you think the end result of releasing massive amount of even totally-above-board information is going to be increased public trust, then you must have sleepwalked the last 6 years. There’s a good case to be made that releasing this is the best course. But let’s not pretend that public trust won’t be eroded because it takes an order of magnitude more work to disprove bulls### than to make it up (or twist it) in the first place. And it isn’t any particularly information I’m talking about, simply the vast quantity involved. The vaccine becoming political is probably the worst thing to happen to humanity in the last 2-3 years. Worse probably than COVID itself. EDIT: just thought of one possible solution or at least mitigation to the “Bull### asymmetry principle” when combined with massive data dumps: Limit the amount of material that can be requested. Any specific dates or whatever could be requested within the stack. Or if a representative sample is desired, the entire dump is sectioned in pieces and each section is securely hashed. The one doing the request can then request, say, up to 100-1000 sections and each such section can then have its hash compared. That ensures the random selection is random to the satisfaction of the requesting party. And it limits the amount of bulls### that needs to be countered to some reasonable level. Otherwise, any attacker will be able to overwhelm the public with endless claims from a massive stash of requested documents. This doesn’t limit the amount of work for scrubbing PII, though, since you’d have to scrub it before hashing or the hashing wouldn’t be recreatable. But at least you’d limit the amount of unique bulls### that’d need to be countered.

Karunamon 5y ago
The FDA could've had the policy that the data is scrubbed and made available as its ingested. And, quite honestly, I see this complaint of theirs as one made in bad faith. Once they have no other choice, they'll comply in a lot faster than 3/4 of a fucking century. Really, I see FOIA requests as failures of government; as in they represent the government being poked to be transparent when transparency should be the default.