BC FOR LATER READ
Saved by @Mollyycolllinss

This is what happens when you train neural networks largely on tone and its stylistic relics. They pick up formal features of arguments (not so much fallacies as tics) that have almost nothing to do with semantic content (focus on connotation over implication).

 This is a secular problem in the discipline. It's got nothing to do with the Analytic/Continental split in the anglophone world. They've both got the same ramifying signal/noise problem, it's just that the styles (tics and connotations) are different in each pedagogical context.
And this is before we start talking about tone policing and topic policing, which are both rife and essentially make the peer review journal system completely unfit for purpose, populated as it is by a random sampling of pedants selecting for syntactic noise over semantic signal.
We've allowed a system of self-reinforcing and ratcheting filters to evolve that effectively *fuzzes* our contribution to the growth of human knowledge (https://t.co/VmW15pGt7J), because it selects for properties only loosely related to those we claim to want. Let that sink in.
This is literally the opposite of what a filter is supposed to do: extract signal from noise, syntactic compression that preserves semantic content. Instead we are awash in syntactic artifacts optimised for minimal criticisable content and maximal pedantic posturing.
When you suspend the assumption that anyone who has been selected by the system *must* understand what they're talking about and/or have something to say, taking the maxim 'fake it till you make it' as a methodological frame, it all looks like competition over formal prowess.
Yet here 'formal' no longer means 'logical', but something closer to 'rhetorical' and 'bureaucratic'. Once you see this, you see that it's exactly the same effect that debating societies and MBAs have had on political and industrial management in the neoliberal era.
Once mighty cognitive institutions (e.g., universities, political parties, and industrial corporations) have gradually devolved ways that are essentially synergetic, as the stupidity in one provides leverage on the growth of stupidity in another. This is the cunning of stupidity.
And so we return to the (social) epistemology of ignorance, but with a couple new points of reference: the anthropologically inspired theories of James C. Scott (cf. https://t.co/kOJPRNzDLj) and @davidgraeber (cf. https://t.co/CMQoBVd44o).
What we find in their work is an account of the socio-epistemic dynamics through which power breeds stupidity and stupidity breeds power, the genesis of stubborn knots around which active ignorance crystallises, generating complex defences that impede the flow of knowledge.
Here's what I mean by synergy between the devolution of cognitive institutions from various domains: the establishment of active ignorance in one place provides points around which alliances form, and through which leverage is exerted to establish it in other places.
The cunning of stupidity climbs the cognitive cliff face by finding purchase wherever it can within a given set of institutions and without. Bureaucratisation and managerialism are secular trends that have spread through the ratcheting the petty personal powers of the ignorant.
The real story of the dissolution and dissipation of the counter-culture of the 60s in the US and Europe is the story of its failure to generate new administrative norms to accompany the ethical, political, and aesthetic ones it had such success with: no persistent counter-power.
Instead we got the slow integration of the acceptable remnants of those norms into a different administrative culture, not simply that which had preceded the 60s (e.g., the party form), but which was born in the 60s, in the form of general purpose management and business schools.
The result was not simply the constitution of a new class of *managers* distinct from industrial workers, whose culture is projected downward through org charts in a way that aligns the interests of *bureaucrats*, but the universalisation of HR as the blueprint for its power.
I think most people on the left appreciate the point that media consolidation created an unprecedentedly homogeneous epistemic landscape in which the formation of political counter-consensus was effectively suppressed. I don't think they see the analogy in the dominance of HR.
I should show them the online forms I've had to fill in to complete academic job applications, whose mind-boggling irrelevance is nothing so much as a trap to discourage those who cannot play the relevant formal game. They build the filters, the obstacles, and the ignorance.
When you arbitrarily empower people who have no idea what it is they're exerting power over, the best intentions in the world will not stop the systemic proliferation of these knots of power/incompetence. The old banality of evil has a new face (https://t.co/XAZxtOFHLP).
I'm neither an anarchist nor a Quaker, but their critiques of these institutional forms, rooted in the autocatalytic growth of power (both centralised/distributed and personal/impersonal), and the serious alternatives they have tried to build are incredibly instructive here.
To return to my story about administrative culture, I think this lets us understand the split between the socially liberal left and the economically progressive left that became formalised in the 90s, and which haunts every attempt to push us out of this autocatalytic nightmare.
Every liberal concession to the counter-culture of the 60s and the liberation movements that have persisted since then has been channeled through the administrative culture of which HR is the nexus. This produces some positive results, but they're always systematically skewed.
For instance, it's great that hiring processes have been evolving in a direction that actively tries to compensate for demographic bias, but when this is merely one aspect of broken filter that otherwise removes the most useful signals, it's a depressingly pyrrhic victory.
Similarly, it's great that the technical language of liberation movements and the deep thinkers who to bring their struggles to self-consciousness disperses into the wider world, but when its principal use is to play out petty micropolitical feuds in formal games, it's maddening.
Managerial overreach, bureaucratic metrics, and quasi-academic jargon are different facets of the same secular process of self-reinforcing power/ignorance. My hope is that academics' perennial hatred of management might form the basis of a teachable moment regarding our own sins.
That's the end of my thoughts on this topic for now, but as usual I will refer you to some more of my extant work on these themes:
1. My recent thread on communicative interaction, dynamic fallacies, and the (synthetic a priori) maxim 'fake it till you make it': https://t.co/2k2QHRYaC7
2. A couple recent threads responding to @cstross, discussing the broken education system and the relation between the plague of terrible metrics and general purpose management: https://t.co/SW2keOnZcv
3. Another recent thread on the Foucault Wars, which lays out the basics of my theory of jargon as pathological technical vocabulary: https://t.co/ZEu1i39QaC
4. A discussion with @Chican3ry about the difficulties of reading philosophy and how to navigate the canon, which includes thoughts on Sellars idea that 'the history of philosophy is the language of philosophy' and a brief critique of 'Sokal Overfitting': https://t.co/NLvAfU0aQT
5. A very old, but theoretically fresh post on 'The Systemic Problems of Contemporary Academia' with an eye to philosophy in particular. This contains my first attempt to think through these issues in information-theoretic terms: https://t.co/sUvKjlf85M
6. Some more recent threads that begin to think through some solutions, by narrowing the problem down to the role of public philosophy:

a) Video Essays as a Philosophical Medium: https://t.co/a6epfzXYLS

b) Enlightenment and Opportunism: https://t.co/V8DJcgf9jm
7. 'The Going Price of Power' (https://t.co/aa2rcbOc40): An outline of the broader economic narrative I'm trying to construct, which ends with a discussion of managerialism and the secular trends I'm complaining about.
Finally, I'm going to reference a couple recent pieces by @OlufemiOTaiwo and @lastpositivist that have been jostling around in my brain as I'm articulating these thoughts:
1. https://t.co/K9z8XUxSIu
2. https://t.co/e76UYs5Zkq
3. https://t.co/ThvjoWStHQ
4. https://t.co/5izUq7ufea
And that's the morning thread! Additional thanks to @deonteleologist for occasioning some of these thoughts.

More from pete wolfendale

pete wolfendale
pete wolfendale
@deontologistics
@paulkreinerhere @schulzb589 I accept that I haven't given a proof of this impossibility, but I believe the concerns I've laid out could probably be formalised in a way that would lead to one, though it'd involve something like a combinatoric diagonalisation similar to Russell's paradox.

@schulzb589 To state my issue in a more philosophical way, I think the desire to combinatorially totalise mathematical syntax is an effect of something like a transcendental illusion, what I sometimes call the Mythos of Logos.

@schulzb589 It's the Leibnizian desire to obviate creative dialogical interaction (dialectical reason) by providing a deductive system (monological reason) that would allow us to simply 'shut up and calculate'; the desire to obviate the pragmatic autonomy of discovery by automating it.

@schulzb589 It doesn't matter how many times this desire is betrayed by its own formal tools: Russell, Godel, Turing, Chaitin, etc.; it doesn't matter that these betrayals are what drive the externalisation of thought by producing new tools to automate computation. It always comes back.

@schulzb589 The desire to brute force reality, be it mathematical or empirical, is always lying in wait, whispering seemingly 'reasonable' lies about what is and isn't rationally possible. This is as close as I come to endorsing Heidegger and Adorno's worries about technoscience.
FOR LATER READ
pete wolfendale
pete wolfendale
@deontologistics
So, here’s a way of reframing this question: which societies enabled coexistence and collaboration between people with divergent social styles, rather than imposing a dominant social style? Such social pluralism is very important indeed.


I suspect that the vast majority of the answers to the original question will fall foul of the tendency to project ideal social arrangements that reflect our own style of social understanding and engagement, and that this will lead them to talk past one another.

Consider the perspective of someone far away from you on in the neurological map, who doesn’t overlap with your socially calibrated genetic resources for social intelligence: the social heaven of an autist introvert may be the social hell of a bipolar extrovert, and vice versa.

I’ve had many good conversations about this with people in different parts of the map who overlap with me in different ways (h/t @tjohnlinward, @dynamic_proxy, @maradydd, @mojozozoe, @UnclePhobic) whose personal heavens I would like to visit, but maybe not live in full time.

We get to see glimpses of these heavens not merely in the past, but in the present, and abstract their geometries, both in spatial/architectural terms (https://t.co/aTcRgtJOVJ) and in temporal/dynamic terms (). The physical/computational platforms around us configure our agency.
SOCIETY
Advertisement
pete wolfendale
pete wolfendale
@deontologistics
My morning thought. I think what's most incompatible about the way I think and the journal article format as a means of capturing and validating thought is that I have a completely different sense of the relation between tentativeness, rigor, and informatic compression.

The characteristic Pete thought is: wait a minute, this whole area is dominated by an assumption that no one seems to be questioning, and I've got two options to express that: i) outline the logic of the issue in a quick and compressed way, ii) write a small book with references.

The discipline seems to want something in between these poles every single time, and this makes me extremely anxious because I feel (with good reason) like any partially referential engagement with the issue will get instantly torpedoed by anyone outside its referential remit.

To repeat something I've said before: I'm actually quite good at being *concise*, I just find it very hard to be *brief*, because most of the thoughts I have don't permit brevity. Consider what your (referential) reactions are to this brief piece:

One of the main genres of publication I've actually excelled in over the years is dictionary articles, which I take a lot of joy in writing, and include some of the pieces of which I'm most proud. Consider this one on the crux of Meillassoux's project:
FOR LATER READ
pete wolfendale
pete wolfendale
@deontologistics
I think one way of looking at the economic pathologies of our society, without using words like 'capitalism' and 'rentiership', is to talk about inequities in the distribution of risk. Some have more taste for risk than others, sure, but risk is not distributed on this basis.

In the UK, so much of our political troubles come from the growth in the landlord class, who if nothing else, wield disproportionate power relative to their social worth: at one and the same time claim to be martyrs to the risk of investment, and demand these risks be minimised.

Of course, it's possible to make claims like this: e.g., nurses have every right to claim that they shoulder not merely more risk than most, but far more risk than is strictly necessary. The pandemic has made this brutally obvious to anyone who has been paying attention.

However one chooses to tally up compensation for this social role, be it financial, social, or some other privilege, it's pretty obvious that every last 'essential worker' is shouldering more societal risk than most non-essential workers compensated far better than they.

The way this trade off gets (implicitly) articulated as 'pay vs. purpose' is one of the enduring insights of @davidgraeber's Bullshit Jobs. Yet there are many essential workers who don't even get the illusion of purpose in compensation for personal risk (e.g., warehouse staff).
ECONOMY
pete wolfendale
pete wolfendale
@deontologistics
I’m beginning to think that the left has been too obsessed with GDP as a broken economic metric governing the global shitshow, when inflation is more insidious. Deflation may be a more radical demand than degrowth. Burst bubbles. Deflate the price of fossil fuel infrastructure.

It’s amazing just how much political economy is still determined by the threat of stagflation. This has created a weird metonymy in which as long as you’re fight inflation (preserving the interests of (older) savers) you’re also fighting stagnation (supporting (younger) earners).

To some extent the residual fear of deflationary spirals lies behind this, waiting in the wings for any suggestion that stagnation must be fought with deflation. This is entirely understandable, as uncontrolled deflation hits the poorest hardest.

The assumption that conjoins these fears into an engine of unbridled economic stagnation is that there’s a unitary market effectively mediating between the markets in different sectors, be this unity micro-economic or macro-economic in flavour. But this just isn’t how it works.

There is so much more logistic friction in the price system than the popular story of just in time production suggests, not least in the concrete details of differentiated trade between economic actors from local businesses to nation states.
ECONOMY

More from For later read

Boz Tchividjian
Boz Tchividjian
@BozT
The evils committed by Ravi Zacharias were enabled by a toxic organizational leadership that is evidenced by how these same leaders responded when confronted about the widespread and ongoing abuses being perpetrated by the namesake of the institution. 1/29

Please take time to read this piece by @DavideAFrench and the 26-page letter written by a first-hand witness. Here are just some of the excerpts: 2/29

"At one point, Ravi suggested to me that the Task Force “issue a statement that they [the Thompsons] are hooligans, they are scoundrels, they are thugs, they are serial extortioners.” If we didn’t do that, he told me, he would ask his attorney to do so on his behalf." 3/29

"...I distinctly recall Ravi & Sanj mocking the Thompsons’ marriage & scornfully predicting how soon Brad & Lori Anne would get divorced. “I predict their marriage won’t survive the year,” Ravi asserted with laughter. “A year?” replied Sanj. “I wouldn’t give it six months!” 4/29

"In 2017, in the early stages of his RICO against the Thompsons, Ravi would often invoke Kelly [his attorney], making statements such as “my attorney said we are going to crush them.”" 5/29
FOR LATER READ
David Boxenhorn
David Boxenhorn
@davidboxenhorn
1. The death of Silicon Valley, a thread

How did Silicon Valley die? It was killed by the internet. I will explain.

Yesterday, my friend IRL asked me "Where are good old days when techies were


2. In the "good old days" Silicon Valley was about understanding technology. Silicon, to be precise. These were people who had to understand quantum mechanics, who had to build the near-miraculous devices that we now take for granted, and they had to work

3. Now, I love libertarians, and I share much of their political philosophy. But you have to be socially naive to believe that it has a chance in a real society. In those days, Silicon Valley was not a real society. It was populated by people who understood quantum mechanics

4. Then came the microcomputer revolution. It was created by people who understood how to build computers. One borderline case was Steve Jobs. People claimed that Jobs was surrounded by a "reality distortion field" - that's how good he was at understanding people, not things

5. Still, the heroes of Silicon Valley were the engineers. The people who knew how to build things. Steve Jobs, for all his understanding of people, also had quite a good understanding of technology. He had a libertarian vibe, and so did Silicon Valley
FOR LATER READ
Advertisement
Sahil Bloom
Sahil Bloom
@SahilBloom
Survivorship Bias 101

History is written by the victors. But if we exclusively focus on these successes, we allow survivorship bias to distort our understanding of the world.

But what is "survivorship bias" and how does it work?

Here's Survivorship Bias 101!

👇👇👇


1/ First, a few definitions.

Survivorship bias is the logical error of concentrating on survivors (successes) and ignoring casualties (failures).

When we do so, we miss the true "base rates" of survival (the actual probability of success) and arrive at flawed conclusions.

2/ When we completely ignore failures, we lose our ability to correctly identify the differences between successes and failures.

Put simply, exclusively focusing on successes may actually inhibit our ability to identify (and replicate) the actions that led to such success.


3/ Cicero wrote on the topic over 2,000 years ago.

An atheist named Diagoras is shown portraits of people who prayed and were saved from death at sea as proof of God's existence.

Diagoras replies, "I see those who were saved, but where are those painted who were shipwrecked?"


4/ Cicero cuts right to the point - we cannot conclude that prayer led to being saved from the sea if we ignore those who prayed and then drowned.

Survivorship bias can lead to a deeply-flawed understanding of cause-and-effect relationships.

Let's look at some examples.
FOR LATER READ
David Roberts
David Roberts
@drvolts
It's a slow, snowy afternoon. How about a 🧵 on "rationalists" & "rationalism"? I want to talk about what real rationalism might look like & what people devoted to rationalism might actually behave like--and how that differs from what you often see from self-styled rationalists.

Caveat up front: this is not about Slate Star Codex, which I haven't read regularly enough to have an opinion on, or about the recent controversy between SSC & the NYT. I'm more interested in the *idea* of rationalism (vs. the practice).

The core premise of rationalism, as I understand it, is simply that humans' cognitive machinery was not designed by evolution to find truth, but rather to advance the interests & welfare of the collective -- a Venn overlap, but not the same thing. Therefore ...

... reasoning well, both individually & as groups, requires special effort. It requires consciously trying to set aside the parochialism, biases, & other distortions that typically characterize our thinking -- resisting, to some extent, the dynamics of group membership/welfare.

That's all fine. I agree entirely that we should make special effort to reason well & to guard against the typical failures of reasoning to which we humans are prone. I believe it's the best (maybe only) way forward as a species. So far so good.
FOR LATER READ
Ooana Trien
Ooana Trien
@ooana
I shared this on my FB page and asked, can ya really blame him?

I was half kidding. I also assumed someone would think of what I did pretty quickly and waiting for the comment to mention what I assumed was obvious.

The timing. I was sure someone else had thought of it.


But no one did. 20+ comments in people discussed the morality or bad sense or libertarian perspectives. Someone even said I’m thinking about doing that. No one said what I thought was obvious. Have you thought of it? Is it obvious to you?

Here’s a clue...recognize it?


How about this?


The author discusses it with Mike Wallace in 1958
FOR LATER READ

You May Also Like

Ottoman History Podcast
Ottoman History Podcas...
@OttomanHistory
Here is a special thread introducing our series on “The Making of the Islamic World.” It offers 10 episodes about the transformation of Islamicate societies between the 7th and 17th centuries. You can find the series page here: https://t.co/7mL9eWmMmz


This is Chris Gratien posting. I created this series during Fall 2020 for my students as asynchronous material supported by readings and discussion. This is going to be a long thread, but it will be a great shortcut for those looking to make similar use of these materials.

There are two items worth pointing out: 1. This is an overview of the political, social, and cultural history of Islamicate societies. It does not deal with the origins of Islam or many theological questions. It does discuss Islam’s spread and the transformation of Islamic law.

2. This is the first part of a two-part sequence that takes the history of the Islamic world into the present. It emphasizes developments and themes that have particular relevance for later periods.

This is really a multivocal narrative of Islamic history. We feature over 20 scholars from almost as many institutions. Many of those who worked on it, myself included, study the early modern and modern era but have teaching obligations that take us further back in history.
RELIGION
SK
SK
@sk_diary1
Sri Sanjeevi Rayar Temple located in Ayyangar Kulam,near Kanchipuram,TN

#Thread

Ancient but beautiful temple dedicated to Hanuman. This temple is a legendary one. According to the Sthala Puranam of Sanjeevi Rayar temple,it is associated with an episode from Ramayana.


Legend story -Hanuman was flying from Himalayas to Lanka bearing Sanjeevi hill to revive Lakshmana during battle between Rama& Ravana, a piece of the Sanjeevi Parvat fell at this place.Many believe this legendary story & even today that no one here dies of any kind of poisoning.

From historical background of the temple it is known that temple was originally built by a great scholar &administrator Kotikanyadanam Lakshmikumara Tatacharya. Temple itself&village were named following name of great scholar Tatacharya, who was then popularly known as Ayyangar.

He was the royal preceptor of Vijayanagara emperor Venkata Raya who had his capital at Chandragiri in AP.Upon returning from a pilgrimage,robbers tried to attack Tatacharya at this place. Being a great devotee of Hanuman,he instantly composed a poem in Sanskrit praising Hanuman.


The bandits immediately ran away &Tatacharya decided to build a temple dedicated to Anjaneya at that very place.This Sanskrit poem is called Hanumad Vimshati, has been inscribed in Grantha script inside this temple.The huge temple tank called Tatasamudram, in honor of Tatacharya.
ALL
Advertisement
Roger Grosse
Roger Grosse
@RogerGrosse
Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ


The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.
MACHINE LEARNING
Laurie Voss
Laurie Voss
@seldo
The 2018 State of JavaScript survey is out. They got 20,000 responses and have some delicious, delicious data. I'mma thread in some highlights:

First up: JavaScript flavors. There's TypeScript at 47% adoption, a tad higher than npm's own survey results (which said 46%). npm's survey is coming up again soon and will ask about TypeScript in a lot more detail.


I really, *really* like SoJ's "would not use again" question, which lets people who've abandoned a tech self-identify. This is noticeable in the graph above with Flow users -- 41% of people who've used Flow say they wouldn't use it again.

Now JavaScript frameworks. It's been a full year since npm's survey, so these numbers differ from ours in the directions I expected:
React 65% (vs. 60%)
Vue 29% (vs. 24%)
Ember 5% (vs 4%, I was expecting a bigger rise)

But there's a shocker in here: Angular.


npm's survey had Angular at 40% last year and SoJ has it at either:
- 58% (if you include those who don't want to use it again)
- 24% (if you count only those who like it)
Since npm's question didn't ask if they intend to *continue* using it I think that might explain this.
TECH
Erik Torenberg
Erik Torenberg
@eriktorenberg
1/ Someone emailed me asking how to break into VC, so I wanted to answer on Twitter where others could see and contribute to the conversation.

2/ “Being a VC” can mean a lot of different things, so it’s worth asking:

What actual activities do you want to do?

- Deep market analysis?
- Be in the flow of information and people?
- Make deals?
- Work closely w/ founders over time (e.g take board seats?)
- Manage capital?

3/ It’s worth specifying what type of VC you might like to become — as there are different archetypes. E.g.

- Benchmark (Lead series A/B - couple investments a year)
- First Round (Lead seed rounds, partner w/ a few companies a year)
- SV Angel (Make lots of seed investments)

4/ Continued:

Expa - Incubate companies

YC / Village Global - Build a platform to help entrepreneurs at scale

Do you want to join a firm or start one? There’s a lot to consider.

Different paths will require different skillsets & sets of experiences.

5/ Since the person who wrote the email is a young person trying to break into VC by joining a firm (and who doesn’t want to start a company), I’ll tailor this tweet storm to that goal. There’s some overlap.
STARTUPS