I mentioned Canada and France. Could also list other countries like Austria and Netherlands that had similar issues. Hell, even Brazil, who had been improving economically and especially for its poorest people consistently since the 90s, and was taken over by a far right government that rose through social media, exactly the way Haidt describes.
My point is that what we are witnessing here is better explained by the sudden access to social media rather than socioeconomic deterioration. By the way, we agree that the working class in the US (and UK) have also experienced a relative decline in the past decades.
France, Canada, Britain and many other countries in the West that have everything you listed are also showing the same signs of division, all emerging in the same timeframe described by Haidt.
That's not an appropriate comparison. You are talking about bad policies and political decisions, while Haidt is focusing on internal division and institutional erosion. The issue here is whether you will be free to see and declare the next war as a mistake, or whether the public space and political arena will be dominated by authoritarianism, fueled by the processes described in the article.
Yes. At its core, the issue with social media is that it amplifies "engagement" at the cost of anything else. It is impossible to breed socially productive ideas, in the allegory you describe, if your only focus is infectiousness. What you are left with are literal mental viruses that take up our collective mind space, leaving nothing in return.
If anything, enforcing identification would be very damaging to these companies. Their user numbers are absurdly inflated by bots and fakes, and most of their engagement is also likely not coming from human activity. The author is not proposing a solution for people in oppressive regimes - those are already restrained in their access to internet and social media -, but for America.
The internet can remain anonymous at large, but we could require any company with an audience larger than, say, 10k people to operate within the parameters Haidt describes. In other words, you can be completely anonymous if using small communities, but if you are operating within a platform so large that it has the power to influence the outcome of elections, we need to regulate whether or not you are actually human. And yes, we do need to make you accountable if you break laws in these platforms. I do value anonymity immensely, yet I see no other way around this hurdle. The solution you describe - simply banning Twitter, Facebook, etc. - is more damaging in my opinion.
Honest question, do you know a serious source I can read about how this rollout is being executed? I am only familiar with European energy policy, and grid storage currently is a pipe dream, more efficient construction is very slowly being rolled out and so far doesn't make a dent in the problem, and the renewable mix being expanded is inevitably intermittent (mostly solar and wind, with the rest being negligible), as hydro is a very small component of the mix so far and in the medium term future. The only question about our ability to build nuclear is cost-efficiency and regulations.
I get what you mean, but do you see how the politics around otherwise sterile scientific decisions influence how scientists will decide what to test, and what gets ultimately funded and reported? For example, HCQ is and has always been a shot in the dark, based on some wonky theory about transmission rates in Africa. That it became a major talking point in some parts of the world before any tests were conducted meant that a) a lot of funding would inevitably go towards testing it, which is likely suboptimal and b) scientists who do not support the politicians who talked about HCQ in the first place may be inclined to review the evidence in a more negative way (and vice versa). Another level of distortion is added by media, as different organizations may report only results that confirm the biases of their clientele. "Just do good science" is an aspirational goal, not the reality we see.
Yeah, I think he picked some bad examples to illustrate an otherwise good point. For instance, the issue with Ioannidis is not really that people thought he was supporting Trump, it was that he made wildly controversial points on flawed, limited data. And even with what we know now, he was wrong. The HCQ stuff was hardly ever controversial from a technical standpoint, it has always been a stupid shot in the dark.
On the other hand, the trends he highlights are indeed dangerous. We have become way more partisan and dogmatic in the last few years, even when it comes to scientific research and communication. And there is a growing tendency to fossilize an imagined consensus into scientific truth. As we all descend into tribalism, it seems that our interpretation and communication of basic scientific facts is succumbing to it as well.
Yeah I was very interested about that point in particular. I think synthetic data is one of the ideas, but I got the sense that he also means helping to identify what makes a data set good, even if small. It looks like Andrew Ng is developing a platform for automatically detecting whether a dataset is suitable and, if not, what are the steps to improve it. A sort of automated ML consultant, allowing you to sell capabilities much cheaper than if you needed to consult an actual expert.
Exactly. It bothers me immensely that we cannot have an adult conversation about topics such as these, because people who only have a superficial understanding try to pick a certain phrase or argument as proof that you are on "the other side". It's actually part of the problem of why we cannot have reasonable public opinion a the moment, as journalists and intellectuals have also adopted this way of thinking.
Chomsky is absolutely right. We are sleepwalking into disaster and all I see is a deterioration of public discourse, where we fail to acknowledge how big the threat of extinction currently is.
The issue is not as straightforward as you describe. The promise was made and several actors confirm it [1]. That it was not put in a treaty does not contradict what Chomsky said. NATO expansion is a huge part of the current conflict, and it is a pity that it cannot be discussed without smears like yours.
Many others have commented on possible biases here. That's true, but I also find this a really interesting observation based on a huge dataset. Once you can correlate it to actual job performance - performance reviews, whether you were fired or not, maybe even salary -, it might be one of the best ways to see if technical interviews actually measure anything significant.
There is no consistent definition of race in humans. Certainly no scientifically valid one. We are all playing word games, politically motivated ones as demonstrated by this article. That's not to say racism doesn't exist of course.
This is a very misleading headline. The data actually shows that the level of neutralizing antibodies for Omicron in both vaccinated and non-vaccinated with previous Covid exposure is similar (panel D), while not been exposed or vaccinated is much worse. So, yes, it does make a difference if you are unvaccinated and have already caught Covid, almost the same level of protection as if you were vaccinated. Plus, this is an in vitro study done with 59 people, so we absolutely should not draw strong conclusions from it.
You are right about this claim, but the only thing I could find to substantiate it is a 2013 article in which leaked documents suggests German federal police wanted to use it [1]. It is still incorrect to suggest this is a German state spyware, although of course you are right about the ambiguity here.
I didn't know about this company before, but a cursory search suggests they were only based in Germany, not used by the German government. Title is misleading, especially as it is not the original title of the article, and Germany has in fact prosecuted them.
That said, good riddance. We desperately need accountability for malware coming from state actors.
This is fascinating. I feel that we are still in the infancy of the field, however. These observations are analogous to naturalists of the past describing an animal's behavior, but we need to get to the point where more accurate estimates are made (ie, how often does it do each thing, how accurate it is after 100+ tries, etc). Every day we see a new observation showing wha GPTs can do, we also need a good way to make these observations systematic.
Create a massively adopted product ->
Start breaking your product to milk profitability ->
Ignore user input and feedback, override their stated preferences with yours
The Microsoft playbook, currently in play at Google.
> The NukkAI challenge required the human champions to play 800 consecutive deals divided into 80 sets of 10. It did not involve the initial bidding component of the game during which players arrive at a contract that they must then meet by playing their cards.
Each champion played their own and their “dummy” partner’s cards against a pair of opponents. These opponents were the best robot champions in the world to date – robots that have won many robot competitions but that are universally acknowledged to be nowhere near as good as expert human players.
Seems like two pretty hefty caveats. Why not pit their best AI against humans directly? And can someone with a better understanding of the game explain how important the initial bidding round is?
The economic incentive to fix your problem, or many of the other areas in which people perceive AI has shortcomings, is very small. The economic liability for failure is also astronomically different in both cases.