Eximius

Security starts at the developer, you have to be deluded to think otherwise.

NDA, bulletproof'ed laptops, kernel-level-oversight, VPNs are just mitigations.

It's a (large) language model. It's good at language tasks. Helps to have hundreds of Gigs of written "knowledge" in ram. Differing success rates on how that knowledge is connected.

It's autocorrect so turbocharged, it can write math, and a full essay without constantly clicking the buttons on top of the iphone keyboard.

You want to keep a pizza together? Ah yes my amazing concepts of sticking stuff together tells me you should add 1/2 spoons of glue (preferably something strong like gorilla glue).

How to find enjoyment with rock? Ah, you can try making it as a pet, and having a pet rock. Having a pet brings many enjoyments such as walking it.

The goal posts were not moved at any point. It was a discussion of the situation, as it is.

Please look at the paper you refer to: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(10)60175-4/abstract It was only retracted because of "In particular, the claims in the original paper that children were “consecutively referred” and that investigations were “approved” by the local ethics committee have been proven to be false. Therefore we fully retract this paper from the published record." It was retracted due to fraud. I don't think it's in any way wise to blame the possibility of fraud on the peer review process. Just as fraud can happen in any field because some people decide to pathologically lie.

However, besides the fraudulent ethics, the paper is fine, and as always previously reiterated multiple times. All it says are a bunch of maybes. It makes no extraordinary claims, it holds no conclusive proof, just a lot of "this maybe hints to something". The paper is publishable.

The actual scandal was caused by the Wakefield lying profusely in media.

These are two different things: what Wakefield said in media, and what Wakefield said in the paper. You should separate them.

In just the same way you can get away from taxes by lying vehemently... he lost his job and reputation in less than three years.

Since the paper itself was okay, but the data was falsified, obviously it was hard to prove the data was false until other studies not only showed it, but also his reputation was discredited and (presumably) investigations finished.

Incorrect data can happen even to a good paper in good faith due to instrument error.

The paper in question, again, was lots of "maybes" and no direct conclusions. The earth shattering conclusions were reached in press conferences where the guy lied vehemently, and the journalists ate it up like coke.

A sentence made out of fluff. What technology? AMD took x86 and gave it wings, better efficiency, neither is only negligible iterative improvements. Intel failed to use lower nm nodes as a first fail.

"[The paper] admitted that the research did not "prove" an association between the MMR vaccine and autism."

"He was reportedly asked to leave the Royal Free Hospital [around 2001] after refusing a request [presumably around 1999] to validate his 1998 Lancet paper with a controlled study."

You could say it took to long to retract the paper, which was essentially full of data-fudged "maybes". But it supposedly was "science" until it was uncovered as just fraud.

Apart from the data fudging, and intense bullshit and hype-train pushing by the now deregistered "professional" [fraudster].

Sorry, this just shows the resillience of publishing, and the scientific community to fraud and [alleged] corruption.

No lmao.

You'll have to actually reference a published paper for that claim.

1. You completely disregarded the paper.
2. Completely disregarded peer review as a thing without any grounding.
3. Went ad hominem as a hail marry.

Bye.

That is complete unfounded fluff words. No paper would be published if it was biased and as selective as you say. Look at the paper at least briefly and we can discuss.

I think you can download it here: https://www.researchgate.net/publication/240678278_Why_Civil_Resistance_Works_The_Strategic_Logic_of_Nonviolent_Conflict

Of interest maybe would be the indicators of a campaigns success:

The outcomes of these campaigns are identiªed as “success,” “limited success,” or “failure.” To be designated a “success,” the campaign must have met two criteria: (1) its stated objective occurred within a reasonable period of time (two years) from the end of the campaign; and (2) the campaign had to have a discernible effect on the outcome.40 A “limited success” occurs when a campaign obtained signiªcant concessions (e.g., limited autonomy, local power sharing, or a non-electoral leadership change in the case of dictatorship) although the stated objectives were not wholly achieved (i.e., territorial independence or regime change through free and fair elections).41 A campaign is coded a “failure” if it did not meet its objectives or did not obtain signiªcant concessions.42

They werent selectively chosen. " An original, aggregate data set of all known major nonviolent and violent resistance campaigns from 1900 to 2006 is used to test these claims." As well as any researcher who isn't a complete buffoon would only look at statistics that has only a 2-3 sigma chance of only being stochastic noise.

I guess that's a fair example. But logically sounds impossible for such control over the population to be had. If a group went out to the streets to oust the government, you would say at least maybe 45% would join.

There is the semi-usually-known research that suggests 3.5% is enough for non-violent protests to reach changes. https://www.jstor.org/stable/10.7312/chen15682

0.5% is 1 in 200 people, essentially everyone knowing personally one person who is against the government. Maybe it isn't enough.

But also, 0.5% homogenously (instead of country-wide being concentrated in Moscow), would be 600k people peacefully marching in Moscow streets