Not quite. The paper uses its own N-gram rules with positive/negative/invariant weights as a rudimentary attention, and these rules are distilled from the model itself.
This, as I found out from this repo [0] linked in the Twitter thread in the documentation (which for some reason they didn't just link to directly), seems to be a regular Markov chain of context, if it even builds a stochastic matrix. See algorithm below.
Current prompt
"Article: (CNN)French striker Bafetimbi Gomis, who has a history of [...]
Summary: French stri"
Prompt lookup algorithm
1. Get last few tokens from prompt -"French stri"
2. Search for "French stri" in prompt
3. Match found - return next k tokens after match as candidate completion -"ker Bafetimbi Gomis, who has"
Candidate tokens
"ker Bafetimbi Gomis, who has"
Broadly, in the US, the Federal Wiretap Act of 1968 still applies. You're going to have to convince a judge otherwise.
Yes, perhaps broad dragnet type of might be scoffed down by some judges (outside of Patriot act FISA judges ofc)
I would warn you about the general E2E encryption and encrypted at rest claims. They are in-fact correct, but perhaps misleading? At some point, for most, the data does get decrypted server-side - cue the famous ":-)"
yes, but you are very much mission driven in the military, especially in a small unit. For me, it was the bad mission set and not the sleep deprivation that personally drove me out.
My understanding: Like you mentioned the main difference is that digital cash is an abstract type. With our current cash we can very easily convert to/from an int currency to physical cash.
With digital cash it might be more difficult, potentially intrusive or costly to convert to physical assets (look at crypto). If I take $20 out of my bank account and buy donuts, it's going to be very hard to the government to trace that. That will not be the case for digital cash. They will know the exact wallet that came from.
I also believe the need for the Fed for a CBDC is forced onto them because of the inevitability of a negative interest rate on savings accounts (which carries negative consequences). Indirectly, CBDC solves that for them. This person explains it practically: youtu.be/UP0e9MGjyD4
Inversely I'm kinda confused how Porsche (owned by VW) seems to get very good reliability scores[1]. Down to 9/26 from last years 4/26. Compared to VW's 24/26 that's quite a large delta.
I always chalked up reliability as a function of new tech introduction + focus on reliability. I figured Porsche gets decent numbers because of how relatively low-tech they are and the ridiculous prices they charge for optioning-in that tech (basic things like ACC or memory seats).
Is there something else organizationally, process or product wise that Porsche does differently than the VW group? I do know that they have a separate factory w/ seperate engineering departments (:cough VW engineers design Lamborghini's).
Curious if this is just marketing from Porsche. Makes no sense why VW wouldn't adopt the system for similar effects. I'm not sure I buy the reasoning but I'll keep a open mind.
It's fair to say GM did copy it well in the Saturn experiment, but failed to bring it fully to the whole company. VW can't help themselves from making complicated things. Honda is onboard, but Nissan still thinks in terms of supply side and will churn out garbage to meet next quarter targets.
That being said, Tesla went back to early Ford for the full top-down flow system - and it's working for them. So while I love the TPS system and use it in my company, I'm willing to keep on learning.
> Inversely I'm kinda confused how Porsche (owned by VW) seems to get very good reliability scores[1]. Down to 9/26 from last years 4/26. Compared to VW's 24/26 that's quite a large delta.
Reliability was always a selling point for Porsches relative to their competitor. You could buy a 308 and be constantly fixing it, or a 911 and (so long as you didn't put it in a ditch) you'd have an everyday driver.
The fact that it's dropping suggests that they're feeling the effects of VW's ownership, crapping their designs up.
Porsche has a bifurcated product offering... their sports lineup is extremely reliable, but their SUVs and more mainstream products are certainly not (these also tend to have brand new equipment which doesn't filter to the 911 for years)
Porsche sports also spend a ton of time racing, at various levels, which I'd imagine back permeates into the corporate culture and increases reliability.
There's a lot of weaknesses that show up a lot quicker when you're holding an engine at redline.
Dont know about that score, is that self reported?
There is plenty wrong with Porsches, with their glued cooling line recalls, where glue degrades and lets go with age, and on some models (Macan?) this means cooling fluid spewing directly at your front brake disks and tires. Then you have electronics routinely going haywire. At least you dont have to worry about IMS bearings anymore.
I know engineers who worked for Porsche, they are doing things independently from VW. The only thing they took from VW was the cayenne chassis and it's diesel engine. They were absolutely pissed by the diesel gate as they were unaware of it and quickly removed diesel from their lineup.
They share some functions like brake SW or ACC, but they do their own engine and chassis.
Otherwise they source their stuff alone, and due to the low sales number they struggle to find willing suppliers.
I have no evidence to support this, but I'd theorize that some of it has to do with enthusiast owners (porsche) keeping up with recommended maintenance intervals, while budget owners (vw) and ignorant owners (audi) skip maintenance, causing nightmares down the line.
As a personal anecdote: my 2003 VW (1.8T GTI) had 330,000km when I sold it to a friend (it continues to run strong, but not sure the KMs). I did all the work myself and I did it on time. Biggest repair was AC Compressor, the rest was just routine maintenance and a couple $50 sensors. If a CEL came on I addressed it as soon as possible. At the same time, I had a friend with a similar year Jetta TDI which lit on fire around the 290k mark. Every time I was in her car the CEL was on, and I remember driving it home on a road trip once where you'd hold the pedal to the floor the whole way just to keep it at the highway speed limit.
In retrospect I would have done the same: Force the nursing homes to accept their residents back if the residents have no other options.
However, I would have throw in much more aid to the ones that were forced. Cash + PPE + National Guard. If I couldn't provide help, then I wouldn't force. That would mean that it would be up to the hospitals to figure it out and find a home for them or until they were covid-free.
Obviously I'm playing backseat Covid response governor here. Hindsight is 20/20.
If you have to send residents back, you must create COVID and non-COVID facilities to ensure isolation. You also shouldn't send infected persons there just because you can.
Where could those people going back to nursing home go? I'm assuming that if another facility or their family could have taken them they would have opted for that. Reminder that likely these people needed 24/7 skilled nursing help.
> Either properly supply your nursing homes with PPE or bring the nursing home staff to the hospitals.
The PPE that was acutely in short supply and unavailable at the beginning of the pandemic? Even doctors and nurses didn't have adequate PPE back then, and many of them got sick and died because of it.
> He knowingly sent people with covid into nursing homes against the will of the people running them...
Don't mean to be aggressive but where else would you like for the people whose domicile is in the nursing home to go? These are not people in independent living facilities, they likely need 24/7 skilled care.
Allowing the people to go back to the nursing home (aka their home) seemed like the right thing to do at the time. It unfortunately didn't play out so well and I am glad I wasn't the one making that decision.
I think it was a tough decision. The nursing homes could have contained the spread but clearly were incapable of doing so.
> Then stonewalled the Feds asking for information about those deaths.
No excuse for that, just lay it all out. The stonewalling was intentional and possibly illegal, he should answer for that.
There was a hospital ship he could have used, it only had 179 patients over 3 weeks. There were hospitals made that never reached full capacity. Hell, setup a tent. You're saying the only option was to infect and kill 15,000 people?
Hospitals did in fact need the beds (nursing home beds >>> hospital beds). Additionally they are ill-equipped for the skilled help these people needed. People suffer from dementia/alzheimers and hospitals only option is to handcuff people to their beds (which is inhumane long-term).
I don't want to play politics and I do think more could have been done here). Like: quickly granting nursing homes aid in terms of PPE + cash + national guard help. I'm just trying to change your mind that the situation was much more nuanced that the NYTimes is advocating for.
The hospital ships weren't for covid patients, but for others who couldn't get a bed due to the hospitals being over capacity. The military specifically was testing everyone coming to make sure they didn't have covid because it can spread way faster in the confines of a ship.
That was true initially, but they changed tack about a week later. The ship reduced capacity from 1000 beds to 500 specifically so that they could accomodate covid patients, because there weren't enough other patients to matter. In addition, the Javits center field hospital never reached more than about 5% of its available capacity. Meanwhile the nursing home down the street from me was begging the city to let them send their covid positive residents to to the ship, or to the javits, were told that those were only for hospital overflow, and they should keep the residents in place. more than 60 of them died, the highest death toll of any nursing home in the state.
The point is those ships were under capacity the whole time.
Send more non-covid patients to them, keep elderly patients in the hospitals. New York never reached full capacity overall.
If you MUST send covid patients to nursing homes, make sure they are prepared with PPE.
Other states did so, Cuomo and the 5 other governors should have known to do so. We already had outbreaks in a nursing home that everyone was aware of.
BPA/BPE/BPS-free thermal paper is harder to find than one might think.
1: https://en.wikipedia.org/wiki/Bisphenol_A