By 'architecture of the internet', the authors mean the nature of social media feeds.

> Those algorithmic biases have demonstrable behavioral consequences.

The algos optimize for engagement, which can roughly translate into the people drive the algos, as they would stop watching or visiting or commenting, if it was not something they wanted to engage in.

So in some ways, is this not democracy to the max?

I wonder if articles like these don’t like the outcomes, or the reflection of society that the algos create. And thus attack them, because they would rather curate and limit conversation and expressions on the internet they don’t like or agree with.

Why not:

(1) directly fund studies and reproductions of studies (promising ahead of time to publish the results, even if negative) targeting the exact issues they're concerned about

(2) writing and publishing extensively to show people the results and help them arrive at a correct interpretation of the data

(3) make a public commitment ahead of time to change opinion based on what the data says, and not to overstate underdetermined theses

... instead of spending money trying to control the political narrative?

That would simply be science doing science -- which has always threatened the establishment because it's accountable to reality, not authority.

Science rightly done never claims authority, just reports on what the data says. Truth is powerful enough on its own.

I couldn't read this article because Science.org left Bot Fight Mode or Super Bot Fight Mode enabled in their Cloudflare settings, causing me to be blocked by a “security verification”. If you use Cloudflare, disable bot stop modes by going to dash.cloudflare.com and selecting your domain and then clicking on “Security” and then clicking on “Settings” and then using the buttons to disable Bot Fight Mode or Super Bot Fight Mode.

So do the limitations (and requirements) of hardware and operating systems. And corporations and billionaires financing and supporting antidemocratic systems and politicians.

Modern smartphones could easily be meshnet nodes, but they don't really support P2P networking.

See: FireChat, Bitchat (removed from the Chinese app store), Airdrop (Apple limited its functionality in China)

Mainstream narrative-shapers concerned that they are losing control of the narrative. Film at 11.

Any story about threats by the Internet to democracy that revolve around Twitter has to account for the fact that only a minute portion of the electorate ever looks at Twitter.

TL; DR: Develop and deploy algorithms that downrank or deprioritize anti-democratic, extremist, or polarizing content.

Just call your opponents anti-democratic, extremist or polarising and here you go. Democracy!

Every single one of these "internet is a threat to our democracy" takes is really about a few things, none of which is a threat to democracy.

1) Hand-wringing about information disintermediation: previously, institutional gatekeepers filtered information and interpreted it for the public. Now, the public sees raw information and forms its own judgements.

2) Social media has cut revenue streams for the sorts of organizations that bleat non-stop about how social media is a thread.

3) Weakening of ability of the institutional class to censor defectors and promulgators of inconvenient facts, which disaffected former censors call "disinformation".

Far from being a threat to "democracy", the internet is the best thing that's ever happened to it. Social media and the internet more broadly have enabled an unprecedented increase in breadth and depth of public participation in the marketplace of ideas. Those who don't like the result never liked democracy.

It's exhausting, this ceaseless cacophony of high-minded bullshit. I'm sick and tired of hearing people exclaim that the internet is a danger to "democracy" when, really, the problem is that the internet produces democratic outcomes they don't like.

It's not a panacea or a magic fix for human nature, but one of the root causes of this is that the underlying architecture of the HTTP(S) Web is just inadequate. The world needs (technically viable and widely-used systems of) content-addressable storage: inherently achivable, mirrorable and recoverable, properly supporting intermittent connections, providing the stability which is the necessary (though not sufficient) base for building things like annotations and back-linking. That certainly can't force people not to choose the laziest and stupid options, but it really can't hurt if at least the underlying technology doesn't make doing anything but the laziest and stupidest thing inherently hard, esoteric and unrewarding. Instead we've created TV on the computer from the visionary Doug Engelbart manifesto Don't Create TV on the Computer. Worse, some people still seem to be trying to pat themselves on the back for the supposed pragmatism and savviness of those decisions, even while at the same time using their other hand to wave a fist at the Big Tech incumbents, content farms and grifters which they gave a structural advantage to. There aren't many things which should be a higher priority, and which are a bigger blocker of general improvement, than the continuing lack of widely adopted and widely adoptable content-addressable storage. Need to do something big about that, folks, and promptly.