Core thesis: One coherent narrative of the past 50 years of human civilization progress:

• We went from a materially scarce society to a materially abundant one; and
• from an informationally scarce society to an informationally abundant one.
• The challenge of this decade is attention scarcity. Many of the day-to-day used tools and content in social spaces are actively misaligned, and operating on the basis of soaking up attention for monetization. And because the financial engine underlying this is unbounded (unlike people's honest effort, which is not), on a sufficiently long timespan (such as where we are circa ~2020), this engine eats up everything.

To unpack the problem a bit:

• "…Tools and tricks that casinos and psychologists and others have used for decades, have been sharpened and personalised and put in your pocket leashes around your brain to tug at you and -yeah I know how dramatic that sounds but nonetheless- it is the explicit goal of many of these systems." "it’s easy to point at the worst of them but.. What’s on my mind a lot is I’m almost more worried about what they’ve done to the whole landscape of attention around them because they’ve become so good at capturing your attention. They’ve upped an evolutionary competition from everything else in the world.
• Everything that survives on human attention all entertainment and media has to evolve better ways at capturing that attention or risk starvation or replacement by a more fit species. The intent is not malicious it’s just blind evolution, which is far more concerning because there’s almost nothing that can be done.
• You’re trying to walk a narrow path through a jungle full of things calling to you and with competition comes specialization: even if you know how to avoid the top predators something less obviously dangerous may get you. The tension traps don’t have to be flashy and fast: some work better with a kind of repetitive dullness because they hook into your mind exactly, or symbiotic species form a little loop that you just circle endlessly going from one to the other to another and back again.
• I worry this is creating an attentionally deficit society that you constantly have a brain tugged at and pulled at is to end up with a brain so used to having its time and attention captured and consumed that even if you step away briefly you’re left with a mind unable to be on its own to be idle to reflect.” Source: CPG Grey: Walk with me

• gwern: "Underuse of System II particularly manifests as over-exploitation/under-exploration, where large potential improvements are foregone because of a lack of a habit or other systematic factor which would trigger exploration. [..] One way to measure under-exploration is noting instances where exogenous randomization or destruction of the status quo option leads to permanent changes or net efficiency gains after the shock is removed, indicating learning or that the status quo was suboptimal all along."
• Like with drugs, the issue with these platforms is that they make you feel okay with being bored; while boredom is a trigger for needing to do something more; and these platforms, like a sponge, soak up that extra energy, and redirect it towards hate.

Taxonomy of attention problems:

Software scaffolds:

• youtube & wall/feeds selecting for content which maximizes the "engagement" / amount of time you're spending there; at extremes, this happens by driving people to more extremist content, but the algo tends to find it's way around attention barriers
• software scaffolds optimizing for drama-maximization. Worst offender here used to be Tumblr, which allowed people to pick-quote specific parts from an overall article, without backlinks, or context present. This created a strong evolutionary pressure for the most anger-inducing pieces to spread, which could then further mutate, and become a super-germ
• To a lesser extent, Twitter does the same

Culture scaffolds:

• controversy igniting thought-germs evolution. "If there is some argument that splits the population, and lasts forever, you're looking at a super-successful pair of symbiotic anger germs, that have reached ecological stability" (CGP Grey, https://www.youtube.com/watch?v=rE3j_RHkqJc ) . This is known as the "toxoplasma of rage".
• people taking offense & offense taking being rewarded, which feeds more anxiety in target group: people "sit around looking to be offended to raise their own social value amonst their online peers."
• These things can happen naturally, but more often than not are deliberately led by a "lapidation entrepreneur" : " Someone makes a mistake, or a foolish or offensive comment, and lapidators come out in force, often in a state of frenzy. Usually they are led by lapidation entrepreneurs, who have their own agenda. They might be concerned with self-promotion. They might be concerned with promoting a cause or with defeating an opponent, for whom the lapidation victim is taken to stand, or can be made to stand. They may want to make the occasion for lapidation stand for the opponent, so that the opponent, or the cause for which they stand, is that occasion." [ Lapidation and Apology by Cass R. Sunstein ]

Solutions:

• The general meta-problem with any short-term hacks, is the strong incentive for the memes, and the agents in the system to hack around any published solutions, and so they aren't reflection-complete: the more people know about them, the less they work. Solutions below therefore fall into categories of: either non-scaleable, hard to reach, or have an incentive that is more aligned with the user
• For gathering data on how aligned your infospace is with your own values, KPIs:
• "A key metric of mental health is how many of your thoughts are new thoughts (sentences you haven't spoken inside your mind before) vs old thoughts (words that have run through your mind many times)." (by EY)
• a good qualitative exercise, is to regularly ask yourself how a piece of content makes you feel. If your first intuitive feeling is anger, you might just be looking at a particularly fit piece of toxoplasma. 
• a good quantitative exercise is taking snapshots of your monitor every 30 seconds (autoscreen) for a day, or two, then reviewing where you're spending your time on, and deciding whether you're okay with such allocation of your time

• Things which currently work: small communities, where there is a barrier of entry, for enduring connections (iterative, infinite game, as opposed to transient one-shots prominent on large social networks); reputational capital, and knowing eachother for a long periods of time
• A leverage point in avoiding toxoplasma, is the bridge people: people who are being rewarded for taking offense, and therefore select for the worst possible behavior of the outgroup. These people act as stressors, specifically triggering ideations of worst-case-scenarios. The fix here is removing these people from your feeds/circles of influence.
• participating in non-US communities. There are two angles here: the business models of many US corps can't be made economical in non-US territories, and therefore they don't target these; and for US-specific culture pieces, anything which does show up goes through the great filter of geography -had to be deemed useful enough for other countries to include it in their communities.
• Self-hosted systems for anything valuable. Dropbox -> owncloud, filters below
• filters, filters everywhere
• One of the most promising solution is coming from MIT media labs: https://gobo.social/ which allows you to (transparently) filter your twitter / facebook feeds based on preferences. Imperfect, but directionally correct
• For Twitter, I wrote my own filter which uses ML magic to classify individual tweets according to pre-labelled tags.
• Software filters, a lot of them. Specifically:
• Against advertisement (ublock origin, and it's many extensions)
• Against being enrolled into targeting systems (again, ublock to filter out client-side JS tech)
• Tampermonkey for overriding site's defaults (eg opting out from dark design patterns)
• Site-specific filters: HN keyword filter, reddit enhancement suite, fluffbusting purity <- great thing about HN keyword filter, and RES is their transparent filtering: ie you see when something has been omitted, and therefore can build trust as to whether they keyword filters are aligned with where you want to be.
• For specific keyword list, see eg https://www.reddit.com/r/BestOfNoPolitics/wiki/filters
• Mental filters, selecting for reliable/useful information sources:
• The best source on this is Crash Course: Navigating Digital Information , highly recommended
• When evaluating a site/social account:
• Who is behind this information? Friends/person/company? & Why are they sharing it? What are the incentives in that content being read?
• What is the evidence for their claims? Is it backed up with evidence? Is that evidence from a reliable source?
• search for opposite -is there any evidence *against* it?
• Lateral reading: Lateral reading: you have to leave the site to understand the site. Read content side-by-side with what others have to say about it. What do other sources say about the organization and its claims?
• [ref: The Facts about Fact Checking @ CC]
• Being very selective for the communities I participate in. At the very basic level, still-working communities have either very strong moderation, some barrier of entry (payment, knowing about the community in the first place, initial onboarding interview), or more likely all of the above
• Aggressively filtering out people who take offense, replicate the toxoplasma of rage, or are otherwise being animated by memes. Specific exercise for this: open fb feed, read until you find a thing which does not improves your life, click on owner (author / group) -click “friend” -> unfollow. Iterate until sanity is restored.

Finally, a bit of activism, which when scaled up to large number of people can actually correct for some of this:

• For persons, the constructive way to nudge into proper behavior is sending them a friendly pm on how to make the conversations more productive: https://devonzuegel.com/post/the-silence-is-deafening
• Corporations sending messages specifically to you are subject to the CAN-SPAM act, and can be Cease&desist'd

Further reading:

• Tools for keeping focused @ lesswrong
• Superstimuli, the other large attention-soak: The Banality of Catgirls @ Bailey podcast