It's officially an American company owned by Eric Yuan who has American citizenship.

Of course it's common knowledge that many silicon valley companies are financed and owned indirectly by China, much of which flows through Chinese Americans loyal to China, but what exactly can be done? They are officially US citizens opening American companies

They would have said Trump is staging a military coup at the election box

Politics is for the most part theater and distraction.

New capital mar a lago

Eng here too.

Couldn't agree more. I find engineers tend to be simple and straightforward people, who like their work to speak for itself. Truth is not coerced, not dependent on preference - it simply exists, and one works with it.

And why can they do this?

There's a point where (information) technological dominance surpasses military dominance. Seems like the technopolies feel they are at at that point. Things like owning people's minds from cradle to grave .. the power to grant access to both mundane but necessary network infrastructure (monetary, markets, etc.) - and ability augmentation (neuralink) - will continue to grow in significance.

Governments have traditionally enforced their will at (threat of) gunpoint. I believe we are seeing the tech companies posture and say to governments "look at us, we can control the population better than you can. We have more effective means at enforcing our will than you."

I think a lot of the support of the breakdown of traditional law and order by big tech is them proving out and enforcing that position. They are saying, look - we can destroy what exists and rebuild it in our image.

Some members of the gov will want to fight back (Trump et al), and others will prefer to ally and aid the technocracy with the military backing of the gov (Obama et al).

Obviously this is also a reason big tech pushed so hard for the deep state uniparty this election - it has clearly taken the position of technocratic ally.

Remote work and autonomous vehicles may do just this

Do you have a source for this? Would like to share around

We need to understand the odds of collision here.

I remember a popular statistic in school - in a class of 30 people, there's something like a > 50% chance 2 people share a birthday.

Following a similar principle, given hundreds of precints and just 100-200 votes per precinct, I think it makes sense that some would "collide" with the same numbers.

Perhaps somebody more statistically inclined can run this analysis.

Vancouver Airport has belt and road and other Chinese propaganda advertising all over it. Also, the security were speaking Chinese at the metal detector/luggage scan. Im 100% sure every major institution in Canada is crawling with pro-ccp spies/informants/influencers. The Chinese act like they own the country.

The analysis is improper. You need to do a trend analysis of the final results vs baselines like Shiva. The original post here was flawed due to unreliability of timestamps data + precision rounding errors.

So the "performance amongst own party" quantity is the performance in relation to proportion of own-party-registered voters?

Why wouldn't they flip down ballot also? That would cover the trail much better I would think

Also comparing timestamped data has issues. Even in an honest system you can have "bouncy" data like this in subsequent API calls due to variable read/write/caching behaviours in the underlying system. You could probably mine it for trends, but you couldn't compare subsequent API calls without a thorough understanding of the underlying systems. I strongly agree that analysis needs to be disregarded

To play devil's advocate, that's going to disenfranchise so many voters. If I voted for Trump, and then my polling station fucked up so bad that my vote can't be counted - in the end, I'm the one being disenfranchised. How would this be reconciled?

Surely Twitter will slap this with a disputed tag any time now...

All electronic systems used in the voting process should at the very least be 100% open source, with multiple levels of auditing to confirm that the published source & system do in fact match, with harsh penalties for any mismatch.

Tabulating votes isn't so magical and proprietary that we have to pay hundreds of millions of dollars for closed source solutions. Imagine if we let private companies handle ballot counting, with no oversight and no poll watchers. That's basically what we're doing with these electronic tabulation systems.

I work with data systems and there are perfectly valid reasons you could have this kind of variability. The underlying assumption of the analysis is that the data provider provides strongly consistent reads on each subsequent update when generating the JSON dump. This is not a safe assumption to make, so unfortunately the analysis isn't valid imo. My message is getting buried unfortunately.

Shivas analysis on the other hand holds significant weight.

I just hope trump and his team and allies don't run with the json analysis without a proper understanding of the potential systems at play

If your Json dump is generated from many single api requests, and your api requests are hitting different caching servers on the provider side, then it could be like that.

This is hypothetical and I'm playing devils advocate, but I want to cover our bases

To play devils advocate, if states all have different data systems. Some would tend to expose more strongly consistent data than others.

Thats a good point and makes sense for a single read cache. I'm thinking of a system with multiple read/write caching servers, and the caching servers aggregate incoming update requests before sending a single update query to the db. You could then have a situation where one server has a pending update with mostly trump votes, while another has a pending update on mostly biden votes. They both won't know about the other's votes until they've sent their respective update queries to the db and have refreshed. If you read data from one or the other, it would look like vote switching.

All that said this is very hypothetical.. I wish we had some more info on the underlying systems at play here

Caching. 2 different caching servers, each had pulled results from a central DB at slightly different times. The first request hit server #1, the second request hit server #2

This is the most plausible "legit" explanation I think

What the analysis should do is compare results from 1 request, to results 10 minutes later. No cache should refresh that slowly. If there is a pattern of vote loss across that time frame, that would be meaningful.

WEIRD NUMBERS HOPPING AROUND LIKE THIS CAN BE EXPLAINED BY CACHING.

In a system with a TON of queries hitting a DB, you'll want to set up caching. If there are enough queries coming in, you'll want MULTIPLE CACHING servers. Depending on when each server refreshes its cache, you can have slightly different numbers being pulled across subsequent API requests.

Assuming this is the case, you CAN NOT just look at 1 API request vs the next API request. That only works in a 100% data consistent system (highly unlikely for this use case).

You'd need to assume there are irregularities from one request to the next, and instead establish some evidence of vote switching across many requests, taking into account the variability due to cache ages. I am not a statistician so I don't know how to do this, but I believe this would be a more accurate way of going about this.

Best case for this analysis would be to get information about how this JSON was generated, especially the underlying system the data is being pulled from.

I want to get more info on how the JSON was generated, and how OP calculated lost votes.

I believe the lost votes could be an affect of accessing the data through a caching layer

I want to verify these findings, as this is potentially massive

How was the NYT's JSON generated?

Fluctuating numbers like this, especially the "lost votes", can often be explained by pulling from an old cache, typically in an eventually consistent data system.

On that same point, are the lost votes calculated just from 1 pull to the next, or do you look at all future pulls to ensure the votes weren't just missing from a single pull (typical of accessing an older cache)

Just want to make sure we've covered our bases.