You searched for:


Trolley Problems and AI

by John Holbo on July 15, 2023

More AI madness! Couple of months ago there was a weird Daily Beast piece. It’s bad, but in a goofy way, causing me to say at the time ‘not today, Hal!’

But now I’m collecting op-ed-ish short writings about AI for use as models of good and bad and just plain weird writing and thinking, to teach undergrads how hard it is to write and think, so they can do better. And this one stands out as distinctively bad-weird. First the headline is goofy: “ChatGPT May Be Able to Convince You Killing a Person Is OK.” Think about that. But it’s unfair to blame the author, maybe. But read the rest. Go ahead. I’ll wait. What do you think? It’s funny that the author just assumes you should NEVER let yourself be influenced by output from Chat-GPT. Like: if Chat-GPT told you to not jump off a bridge, would you jump off a bridge? There is this failure to allow as we can, like, check claims as to whether they make sense? A bit mysterious how we do this yet we do. And ethics is a super common area in which to do this thing: so it only makes sense that you could get Chat-GPT to generate ethical claims and then people could read them and, if they make sense, you can believe them due to that. Never mind that the thing generating the prospective sensible claims is just a statistics-based mindless shoggoth.

If a shoggoth is talking trolley sense about OK killing, believe it!

Anyway, I thought it was funny. [click to continue…]

A trolley problem

by John Q on July 26, 2017

I’ve generally been dubious about trolley problems and similar thought experiments in ethics. However, it’s just occurred to me that an idea I’ve tried to express in the economistic terms of opportunity cost, without convincing anybody, might be more persuasive as a trolley problem. So, let’s start with the standard problem where the train is about to kill ten people, but can be diverted onto a side track where it will kill only one.

In my version, however, there is a second train, loaded with vital medical supplies, which is about to crash. The loss of the supplies will lead to hundreds of deaths. You can prevent the crash, and save the supplies, by diverting the train to an alternative route (not killing anybody), but you don’t have time to deal with both trains. Do you divert the first train, the second train, or neither?

Hopefully, most respondents will choose the second train.

Now suppose that the first train has been hijacked by an evil gangster and his henchmen, who will be killed if you divert it, but will otherwise get away with the crime. As well as the gangsters, the single innocent person will die, but the ten people the gangster was going to kill will live.

The impending crash of the second train isn’t caused by anybody in particular. The region it serves is poor and no one paid for track maintenance. If the train doesn’t get through, hundreds of sick people will die, as sick poor people always have, and nobody much will notice.

Does that change your decision?

[click to continue…]

A non-violent, unfunny trolley problem

by John Q on December 3, 2013

So, you’re a controller for a municipal trolley system with a perfect safety record. You’ve just been alerted that one of your tracks, serving a community of 5000 people has suffered unexpected damage which could cause a trolley car accident involving fatalities among philosophers. You have no budget allowance for this, so the only way of fixing the line is to abandon planned maintenance of another line, serving 1000 people, which would then have to be closed until more funds become available. Presumably, in these circumstances, most people will decide to fix the more important line.

Now, we change the situation. You no longer control the funds for the other line, which are within the jurisdiction of your colleague the Fat Controller, so named for obvious reasons. If you draw management attention to his obesity problem, HR will force him to take leave until he can get his weight within acceptable limits. You will then be given temporary control over his line and the associated budget, which you can divert to fixing the more important line.

What should you do?

[click to continue…]

Irish Steam Trolley

by John Holbo on June 2, 2011

The National Library of Ireland is on Flickr, contributing public domain photos to the Commons. If anyone is looking for an image for the cover of their new book on trolley problems – and is too shy to ask to use mine – this might be the ticket.

Steam Tram, Antrim

Trolley Problem Follow-Up

by John Holbo on October 7, 2010

My 9-year old daughter was rather curious about my Trolley cartoons, so we made a podcast. Her art criticism is spot-on, no question, and she makes some pretty strong moral claims. What do you think? (Don’t worry. I have her permission to post this. I interviewed my 6-year old, as well, but she declined to give consent to publish her philosophical work at this time.)

Trolley Problems

by John Holbo on September 25, 2010

Evergreen philosophy topic! Sure to inspire much friendly discussion!

I don’t usually lecture about the stuff myself, but this semester I decided to, so I cartooned up some images for the PPT slides. So the first thing I have to say is that if anyone has a use for ’em, I’ve released ’em under a CC license. [click to continue…]

Trolley Problems

by Brian on August 30, 2003

A staple of intro philosophy courses is the ethics of runaway trolleys. There’s probably an interesting sociological study as to why this is so, but rather than delve into that I thought I’d share a new-sounding version of the trolley problem due to Carolina Sartorio posted on Philosophy from the (617).

[click to continue…]

On Shallow Ponds and Effective Altruism

by Eric Schliesser on April 5, 2024

In the wake of the sentencing of SBF last week there were two mighty takedowns of effective altruism: one (here) “The Deaths of Effective Altruism Sam Bankman-Fried is finally facing punishment. Let’s also put his ruinous philosophy on trialby Leif Wenar (Stanford) at Wired; the other and better written, “Neo-Utilitarians Are Utter Philistines” by Justin Smith-Ruiu (Paris) at his (here) Substack (Hinternet). In response Richard Pettigrew (Bristol) wrote a rather sensible criticism, “Leif Wenar’s criticisms of effective altruism” at his blog (Richard’s Substack). On of our very own,  Chris Bertram, shared it it on social media with a note that Wenar’s piece was shared widely without “a sober assessment of the merits of his arguments.”

Now, I am an avid reader of Pettigrew’s blog because more than anyone today, he makes on-going debates in formal epistemology and decision-theory available to wide audiences in a relatively fair and relatively accessible fashion. And like the very best blogs, he also shows the salience of different debates within some specialist area to other areas of philosophy (and the sciences/life). I also find Pettigrew rather judicious generally.

[click to continue…]

1. Ideology

Silicon Valley’s ideology is this: Libertarianism for me. Feudalism for thee.

In more detail:

• Surveillance, manipulation and coercion; at first, just for profit, later by necessity, and ultimately for the hell of it

• Disruption and capture, not competition; monopoly or at least duopoly in each industry it envelops.

• Oligarchy to begin with, creeping autocracy for the win. Overseas autocrats the best of friends.

• Pick me or China wins.

• Ever-increasing inequality and the concentration of capital within a small, interconnected group who back each other’s companies and public moves.

• There is no such thing as human rights. There is only identity politics and culture war, which are profit centres.

• Far right white supremacism; libertarianism for white men, forced birth for white women. Eugenics for everyone else.

• A series of bullshit dark utopias designed to drive the hype and private equity cycles, distract and dazzle gullible politicians and policymakers, and convince everyone else that there is no alternative. E.g. crypto-currencies, Facebook’s Metaverse, AI and, of course, Mars.

• Systematic racism and misogyny in the workplace, the destruction of organised labour, the ever-worsening of working conditions, extreme inequality.

• Denigration of human agency and creativity, beginning with writers, artists and musicians. Systematic destruction of their ability to earn a living and suggest alternatives.

• Obsessive optimisation along narrow spectrums; externalisation of risks and costs to others, i.e. workers, ‘data subjects’, the public sector.

• Gutting of independent media, hatred of journalism in particular and accountability in general. Buying out or shutting down all opposition.

• State subsidies and tax dodging. Hollowing out the state. Making private – both in terms of ownership and secrecy – what used to be accountable and universal public services.

• The spoils to the strong, the costs to the weak. Might is right. Winner takes all. The state is an enforcer, not a support. Let the long tail starve.

Silicon Valley ideology is a master-slave mentality, a hierarchical worldview that we all exist in extractive relation to someone stronger, and exploit and despise anyone weaker. Its only relations to other humans are supplication in one direction and subjugation in the other, hence its poster-boys’ constant yoyoing between grandiosity and victimhood. Tech bros like Thiel, Musk and Andreesen are the fluffers in the global authoritarian circle jerk. Putin is the bro they’d be tickled to receive calls from, making them feel they’re on the geopolitical insider’s inside track. MBS is the bro they envy but tell each other scary stories about. Like most of them, MBS inherited his head start in life. He has all the money, all the power, a nice bit of geo-engineering on the side, and he dismembers uppity journalists without consequence. A mere billionaire like Thiel can only secretively litigate them out of business.
[click to continue…]

Nietzsche and Isohighertype

by John Holbo on July 16, 2023

It’s Sunday. It’s quiet. I’ll just clear the decks of my philosophy faux-infographics jokes. It’s not just trolleys. Some months back I considered it seriously:

“Nietzsche’s key design insight: complex, esoteric ideas, appreciable only by the few — perhaps only by the One! — can be conveyed via simple, conventionalised iconography, suitable for delivering simple, readily understandable ideas to the many.

Naturally, as this insight made no sense, no one had any notion what Nietzsche was on about …”

But that’s no reason why it didn’t happen! [click to continue…]

Kevin Drum points to an obscure, but radical proposal to change the way the US government does benefit cost analysis. The Office of Management and Budget has released draft guidance saying

One practical approach to implementing weights that account for diminishing marginal utility uses a constant-elasticity specification to determine the weights for subgroups defined by annual income. To compute an estimate of the net benefits of a regulation using this approach, you first compute the traditional net benefits for each subgroup. You can then compute a weighted sum of the subgroup-specific net benefits: the weight for each subgroup is the median income for that subgroup divided by the U.S. median income, raised to the power of the elasticity of marginal utility times negative one. OMB has determined that 1.4 is a reasonable estimate of the income elasticity of marginal utility for use in regulatory analyses.

This is pretty obscure, but what it means is that, a project that delivers a dollar of benefits to each of a group of poor people is worth more than a project that delivers a dollar of benefits to each of a group of poor rich people.

A lot more !

Kevin uses a graph to illustrate, showing that an extra dollar for the median household is worth 50 times as much as an extra dollar for a household with an income of $1 million a year. Conversely, an extra dollar for households at the bottom of the income distribution is worth 12 times as much as an extra dollar at the median.

It’s actually simpler to get the intuition of you use an elasticity of 1, which corresponds to logarithmic utility. Then you can sum up the implications by saying that a given percentage increase (or reduction) in income yields the same additional (or reduced) utility no matter who gets it. So, for example, if a policy halved Elon Musk’s income, while doubling the income of a single randomly chosen US household, it would be evaluated as neutral. If the policy doubled the income of two households, it would be beneficial. More generally, you can just add up all the percentage changes in income from the project (included the taxes needed to finance* it). If that sum is positive, the project should be approved.

[click to continue…]

A cage went in search of a bird

by Maria on November 18, 2019

Publishing here my afterword for “2030, A New Vision for Europe”, the manifesto for European Data Protection Supervisor, Giovanni Buttarelli, who died this summer. The manifesto was developed by Christian D’Cunha, who works in the EDPS office, based on his many conversations with Giovanni.

“A cage went in search of a bird”

Franz Kafka certainly knew how to write a story. The eight-word aphorism he jotted down in a notebook a century ago reveals so much about our world today. Surveillance goes in search of subjects. Use-cases go in search of profit. Walled gardens go in search of tame customers. Data-extractive monopolies go in search of whole countries, of democracy itself, to envelop and re-shape, to cage and control. The cage of surveillance technology stalks the world, looking for birds to trap and monetise. And it cannot stop itself. The surveillance cage is the original autonomous vehicle, driven by financial algorithms it doesn’t control. So when we describe our data-driven world as ‘Kafka-esque’, we are speaking a deeper truth than we even guess.

Giovanni knew this. He knew that data is power and that the radical concentration of power in a tiny number of companies is not a technocratic concern for specialists but an existential issue for our species. Giovanni’s manifesto, Privacy 2030: A Vision for Europe, goes far beyond data protection. It connects the dots to show how data-maximisation exploits power asymmetries to drive global inequality. It spells out how relentless data-processing actually drives climate change. Giovanni’s manifesto calls for us to connect the dots in how we respond, to start from the understanding that sociopathic data-extraction and mindless computation are the acts of a machine that needs to be radically reprogrammed.

Running through the manifesto is the insistence that we focus not on Big Tech’s shiny promises to re-make the social contract that states seem so keen to slither out of, but on the child refugee whose iris-scan cages her in a camp for life. It insists we look away from flashy productivity Powerpoints and focus on the low-wage workers trapped in bullying drudgery by revenue-maximising algorithms. The manifesto’s underlying ethics insist on the dignity of people, the idea that we have inherent worth, that we live for ourselves and for those we love, and to do good; and not as data-sources to be monitored, monetised and manipulated.
[click to continue…]

Hack Gaps and Noble Lies

by John Holbo on December 9, 2018

These days we are healthily cynical about the omnipresence of motivated reasoning in cognition and communication. Everyone is working to fool everyone, starting with themselves. (It used to be you had to read Nietzsche to learn this stuff. Ah, those were the days.) [click to continue…]

Clerihew corner (philosophy edition)

by Harry on October 31, 2017

I rediscovered Clerihews at my dad’s house recently, noticing and devouring his copy of The Complete Clerihews. Clerihews were invented by EC (Edward Clerihew) Bentley (author of Trent’s Last Case, one of the greatest detective novels to be written by someone who wrote only one detective novel). [1] They have a strict format: they must be about a person who is reasonably well known, 4 lines, with an AA BB rhyming scheme. They also seem to observe other unstated criteria. They are not cruel, or didactic, and the person about whom they are written must have some sort of substance. Gavin Ewart points out in his introduction to The Collected that they could be used for biting satire but Bentley never does that and he (Ewart) is unaware of them ever being used so. They are gentle. The pay-off would ideally come as something of a surprise, but not be completely irrelevant: Surrealism and whimsy are permitted, indeed encouraged. The name of the subject usually (but not always) comes at the end of the first line, challenging the author to find a good, but non-obvious, rhyme (as you’ll see in a moment he rhymes Plato with potato which is only one part of the genius of my favourite clerihew). There are no rules about rhythm except, as far as I can tell, that if you write several, you should vary the meter. (All Bentley’s clerihews seem to have been illustrated by GK Chesterton: I can’t rival that!).

Here are three of Bentley’s best:

The intrepid Ricardo
With characteristic bravado
Alluded openly to Rent
Wherever he went.

It was rather disconcerting for Hannibal
To be introduced to a cannibal
Who expressed the very highest opinion
Of cold pickled Carthaginian

And my favourite of all:

Although the dialogues of Plato
Do not actually mention the potato
They inculcate strongly that we should
Seek the Absolute Ideal Good.

The unwritten strictures rule out certain subjects. It seems morally dodgy to be whimsical about President Trump, for example (Bentley does one about Goebbels and Hitler, and it doesn’t feel right at all, especially when you learn it was written in 1939); and just difficult (though not impossible, and I don’t think dodgy) to be whimsical or surreal about Jacob Rees-Mogg (how can you be surreal about a person who declares himself, correctly, an embodiment of the surreal?). Bentley tends toward historical subjects – kings, presidents, philosophers, musicians, literary figures – with just a smattering of his contemporaries, most of whom are no longer household names, represented (Captain Wedgewood-Benn features – with luck there will always be a Benn about whom to write clerihews).

So, I’ve been trying to write some. They’re hard!

[click to continue…]

Walton’s Republic

by John Holbo on January 28, 2016

“It was the most real thing that had ever happened.”
– Jo Walton, The Just City

Thanks to Jo Walton for writing an SF novel in which people, including a pair of gods, try to realize Plato’s Republic. (I’ve only read the first Thessaly novel, The Just City. So if what follows is premature? That sort of thing happens.)

This is an experimental novel. Succeed or fail, you learn from an experiment. But even well-constructed experiments can be failures. That’s the risk.

Logically such a thing should exist. A novelization of Plato’s Republic, I mean. How can no one have written this already? But can such a damn thing be written ? Surely it will fail as a novel, somewhat, at some point. But how? Only one way to find out. [click to continue…]