HNHacker News
TopNewBestAskShowJobs

quinndupont

726 karma · joined June 9, 2010

Web: iqdupont.com Twitter: @quinndupont

[ my public key: https://keybase.io/quinndupont; my proof: https://keybase.io/quinndupont/sigs/1KgERye-FogNCHQls2Bapg2znpUmIi1zbg3HlH4LIGo ]

submissionscomments
quinndupont··on Labor market impacts of AI: A new measure and early evidence
AI is coming for jobs—but the real risk isn’t where most people are looking.

The leading AI exposure indices (Anthropic, Eloundou et al.) focus on which jobs get automated. They treat low exposure as “safe.”

But the least exposed workers—cooks, roofers, dishwashers, construction laborers—are often in the worst jobs: low pay, high physical toll, short career spans, and little upward mobility. Safe from AI, but not from burnout or injury.

I built JQADI (Job Quality-Adjusted Displacement Index) to combine AI exposure with job quality. It surfaces three kinds of risk:

High AI exposure → classic displacement risk Low AI, low quality → “trapped” workers in grinding, unsustainable jobs Moderate AI, low quality → partial automation strips cognitive work and leaves physical drudgery (the “task residual” effect)

Findings: 83.5M workers are in low-AI, low-quality jobs. Customer service reps, data entry keyers, and medical records specialists sit at the intersection of high exposure and poor quality. Meanwhile, chief executives and lawyers are both low-exposure and high-quality.

The index uses ONET, BLS, and Anthropic exposure data. Code and methodology are open source. LINK https://github.com/quinndupont/JQADI

quinndupont··on Show HN: JobApplicator (tailored job applications in minutes)
This isn’t ChatGPT-style slop. There’s some secret sauce and it takes your process from 30-60 min per application to just minutes. It’s been through many recruiters for testing.
quinndupont··on Claude's Cycles [pdf]
Interesting to see the mathematical solution space get optimized away. On account of “there’s no accounting for taste” this actually makes me hopeful that creative workers have durable skills that can’t be optimized, which I can’t say about mathematics and computer science.
quinndupont··on Scientists discover beer bottle in the Mariana Trench (2024)
SO MANY ADVERTISEMENTS. Tis a shame everything has to be fluffed up and sold
quinndupont··on Socialist ends by market means: A history
To what degree is marginalism dependent on price as the active force? How can marginal utility be realized without all goods having a market price? What happens when goods are difficult or impossible to price?
quinndupont··on Vanity activities
Vanity is contextual. Everything is vanity from some ultimate perspective.
quinndupont··on Welcome, the entire land – "Hello, world" in hieroglyphics (2009)
Representing the alphabetic orthography for Hello World is obviously impossible in a written script lacking letters, revealing the folly implicit in the exercise.
quinndupont··on Neural audio codecs: how to get audio into LLMs
There’s a long history of attempts at artificial speech that take this approach, recreating mouth parts and vibrating air. They are all pretty silly, like this work, which fails to understand how writing isn’t just a derivative of speech.
quinndupont··on Neural audio codecs: how to get audio into LLMs
Y’all need to learn about the history and development of spoken language and writing. Writing isn’t just a copy or derivation of writing. LLMs work because of the conceptual characteristics of writing (consider the distinctions between ideographic, logographic, alphabetical…). What a sloppy mess!

Read some Wittgenstein and Goodman, but especially Derrida who calls this logocentrism.

quinndupont··on Why do LLMs freak out over the seahorse emoji?
pictograms ≠ alphabetic symbols
quinndupont··on x402 — An open protocol for internet-native payments
Corporate capture of payment rails, masquerading as open payments.
quinndupont··on DeskHog, an open-source developer toy
The hardware is UGLY.
quinndupont··on Alumni Labs: public good funding for blockchain diplomas
Thanks for the feedback! Making sure the issuer is legit is basically an oracle challenge that needs off chain coordination.
quinndupont··on Will Digital Technologies Kill the Nation State?
This short article offers wild opinions with no understanding of history or any meaningful analysis. The authors also clearly have an ideological axe to grind (I mean, come on… the CATO Institute?)

If there were any facts presented I’d say they lie, but since they’re just talking about a made up fairy tale, it’s only a silly piece of fiction. I feel stupid leaving a comment warning others.

quinndupont··on Ask HN: How do I easily catalog a couple thousand physical books?
I use Libib as well. It has a blazing fast barcode scanner, though manual entry is a bit cumbersome. They seem like a good company, though longevity might be a concern. There is an easy export option, anyways.
quinndupont··on Spices Have Made, and Unmade, Empires
Is nobody going to mention how beautifully it was written? Surprising erudition and style.
quinndupont··on The Human Cost of Amazon’s Fast, Free Shipping
Seems to me that the real story here isn’t Amazon but the heartbreaking frequency of vehicular crashes. Always fun to pick on Amazon though.
quinndupont··on A generic beer that was too cheap for Saskatchewan
Agreed. While I find minimum price thresholds and sin taxes personally annoying, there's lots of good research associating ease of access and low prices with alcoholism, which has high costs in terms of healthcare and individual suffering.
quinndupont··on Elixir: A Mini-Documentary 2018
Naive question: are these sorts of advertising/documentary things common in open source programming languages? Open source software more generally?

I watched the documentary and I found it to be compelling and well produced while also informative.

quinndupont··on AlphaSmart
Not to turn HN into ebay, but if anyone wants to buy my Freewrite (modern version of AlphaSmart, as the developers point out below), hit me up: iqdupont.com
quinndupont··on Voynich manuscript: the solution?
That's the curious thing: to an academic there's no such thing as "extremely boring Latin", especially when it comes to the Voynich. Why we're only given a tantalizing line is odd. Perhaps there is no single "universal" solution? Perhaps this is just a guess that happens to roughly work for a single line, but nothing else?
quinndupont··on Voynich manuscript: the solution?
Agreed. It's a super strange article (sure, written like an academic, but even most don't bury the lede that badly!)

From the article it isn't clear if all or just a portion of the text is decipherable using the implied logic (ligatures of abbreviations of medicinal items).

quinndupont··on Bitcoin Energy Consumption Index
Bitcoin wants to be money, but it really isn't. It fails the traditional account (originally from Jevons); see Kubát 2015 [1]; Mandjee 2015 [2]. More obviously, it is far too volatile to function like money in a normal sense, see Golumbia 2017 [3]

[1] http://www.sciencedirect.com/science/article/pii/S2212567115... [2] https://litigation-essentials.lexisnexis.com/webcd/app?actio... [3] https://motherboard.vice.com/en_us/article/xwwv83/cryptocurr...

quinndupont··on The History of Alice and Bob
If I recall the gist of the Rogaway article, from memory, (again, it's an important enough article that it really does merit serious attention):

he argues 1) cryptographers (and computer scientists in general) should be more political (good!), 2) cryptography needs a new framing (good!), 3) privacy and exception from government search is an unalloyed right (not so good), 4) better crypto will solve privacy issues (not so good).

In sum, the article does a lot of good work, and more than anything, it contains some important and refreshing rethinking of the field of crypto. This is all very important. Nonetheless, it ends up taking for granted a number of political positions that should, I think, also be contested (Rogaway takes the first step!). More crypto does not equal a better world.

quinndupont··on The History of Alice and Bob
Author here: I've never enjoyed the basic premise of the Rogaway article, but that would require its own engagement to properly discuss. That said, I had forgotten about his engagement with Alice and Bob, so I ought to include it. His argument against cutesiness, however, misses the deeper political point, and unfortunately comes off as politically superficial.
quinndupont··on After 3072 hours of manipulating BGP, a Nyancat was drawn on this RIPE interface
Oodles of "actual" net art from back in the 1990s and early 2000s. Jodi is especially famous/good: http://www.net-art.org/jodi
quinndupont··on The blockchain paradox: Why DLTs may do little to transform the economy
This argument certainly has some validity (see also Ensmenger's project to understand the environmental role of computing in general: http://homes.soic.indiana.edu/nensmeng/enviro-compute/dirtyb...).

However, it only goes so far, since other consensus mechanisms could have lower energy requirements. Proof of stake, e.g., (if it ever gets off the ground) could change the game. Similarly, if the economics of proof of work were re-jigged, it isn't impossible that the energy required for mining could be dramatically lowered.

But, as it stands, Bitcoin, in particular, has a shamefully large environmental footprint.

quinndupont··on The blockchain paradox: Why DLTs may do little to transform the economy
If there's a new argument to be had here (I'm doubtful, it sounds like a re-hashing of governance issues that have played out over the last year), it's this claim:

"This rule-making is what we refer to as governance."

This conception of governance is the premise for the conclusion, what Lehdonvirta calls a "paradox":

"And this leads me to my final point, a provocation: once you address the problem of governance, you no longer need blockchain; you can just as well use conventional technology that assumes a trusted central party to enforce the rules, because you’re already trusting somebody (or some organization/process) to make the rules. I call this blockchain’s ‘governance paradox’: once you master it, you no longer need it."

My own analysis draws on a more expansive definition (from Introna; Saurwein et al.): http://iqdupont.com/assets/documents/DUPONT-2017-Preprint-Al...

This expansiveness is necessary, as Lehdonvirta recognizes. Contracts are not just enforcement, they also encapsulate the processes of making ("rule-making"), or better, negotiating contracts.

So, are we authorized to go from a rule-making notion of governance to the conclusion that once this is addressed blockchain technologies will be redundant? Lehdonvirta doesn't give us any analysis of this move, but I don't think so. Rather, we need to recognize the contested forms, which are the essence of governance, and then create a socio-technical structure around them. Consider: double-entry bookkeeping is just as much a technology for governance as blockchain technology (and, of course, they share a historical lineage). Does double-entry bookkeeping go away because we "address the problem"? Hardly. Rather, we need to recognize the ways that it has developed, and by whom, and for what use. Just like blockchain technologies.

quinndupont··on The blockchain paradox: Why DLTs may do little to transform the economy
This is an excellent reply.

That said, we'd need to agree on definitions of "state" here. Some kinds of money surely existed before anything recognizable as a "state", but then again, such notions of money are likely as meaningless as this kind of unrecognizable state.

quinndupont··on Cryptoeconomics 101
I have a short section on cryptoeconomics in my upcoming book, which I'll post below (still developing, and still in draft). Unlike practically every other description of the "field", I don't take it for granted and don't uncritically adopt it. (Feedback always welcome!)

<snip> The technological imperative of smart contacts and blockchain technology presupposes rational contracting, which leads to particular moral and epistemological positions. Early rationalizations of the moral and epistemological position of cryptographically-assured smart contracts can be found in Reagle (1996) and Szabo (1997). Since that time, several authors have further developed the concept, leading to what is today sometimes called “cryptoeconomics.” This emerging field deploys the technological affordances of (public key) cryptography: specifically, the ability to authenticate and verify parties; ensure non-repudiation of data; and in some cases, selectively reveal identities (as with blind signatures) (Chaum, 1982). Leveraging these cryptographic affordances, authors use the closely-related fields of rational choice theory and game theory to construct protocols and arrangements for creating social behaviours. In an early speech, Buterin suggests that, with cryptoeconomics, “as soon as you have a decentralized consensus framework that controls an internal asset and there is this valuation equilibrium where people care about it… the cryptosystem has the capability to have real world consequences…” (Vitalik Buterin: Cryptoeconomic Protocols In the Context of Wider Society, 2014). For instance, Buterin considers a number of rational choice strategies, including the prisoner’s dilemma, Nash equilibria, hawk/dove games, markets, reputation, and an odd characterization of legal courts. Following Nick Szabo’s earlier discussion of smart contracts, Buterin focuses much of his discussion of cryptoeconomics on Schelling points.

Schelling points, or focal points, are a game theory concept developed by Thomas Schelling in his book The Strategy of Conflict (1960). The strategy allows two or more parties to reach agreement in an environment with little information and no possibility of communication. The key insight of Schelling’s theory is that coordination is easier when people’s attention is drawn to prominent or focal points. Schelling offers the example of two people in New York who need to meet in an undetermined location without having prior communication. According to Schelling, on game theory assumptions, each party will likely pick Grand Central Station because in a world in which many possibilities are equally likely, humans look for patterns or unusual focal points of reference. The assumption made by each party is that if I think Grand Central Station is a good meeting spot, the other person may think the same.

Similarly, in his descriptions of cryptoeconomics, Buterin offers the example of picking matching numbers from a list without coordination or communication (Vitalik Buterin: Cryptoeconomic Protocols In the Context of Wider Society, 2014). People will, according to the theory, look for something—anything—that makes one number stand out—perhaps an even number, or a number with many zeros, or a well-known lucky number. Thus, given the following numbers: 2349 65 22 100 932, both parties will likely choose “100.” In Western, decimal-based numeracy, 100 “stands out” from the rest of the numbers, and therefore, lacking any other way to coordinate, the rational choice is to assume a common set of beliefs with the person you are trying to coordinate with. Of course, these choices are culture and context specific. With numbers, other patterns might be significant (such as “lucky” numbers), and the same goes with meeting in New York: as times, people, and places change there might be other possible focal points—Times Square, One World Trade Center, Brooklyn Bridge, or the New York Public Library. In these scenarios, all available options might be “weak” focal points. Consider the following scenario where there might be only weak focal points: negotiations between a workers’ union and a business break down and cannot make progress. On the theory of Schelling points, this might be due to lack of strong focal points, in the sense of shared interests and values. Until some stronger focal point can be found, coordination will not be effective. On the other hand, some focal points are very “hard”—so hard that they curtail negotiation. For example, in most of the Western world, the price tag on a retail good is not usually an invitation to start haggling (Szabo, 1997). The stated price is a hard focal point that communicates the context and sometimes the parameters of the contract.

While the use of Schelling points is quite powerful and can be used to carefully navigate social contexts, it is not without issue. At its heart, like all rational choice and game theory strategies, a rational (or maximizing) actor is usually assumed, along with some kind of contractual framework or rationality. Fundamentally, cryptoeconomics fails to fully recognize that actual societies are also the result of war, exploitation, racism, and patriarchy (Held, 1987, p. 113). Patriarchy is an especially important value to consider for cryptoeconomics, since complete freedom and equality—the ability to sincerely engage in contractual negotiations—is a uniquely male capacity in most modern contexts. Rational choice or game theory assumptions lead to a shallow view of human sociality, which is likely to produce bad ethics and/or lead to significant errors in use and development (as with The DAO in 2016, see Chapter 8). One alternative to this kind of rationality, suggests Virginia Held, is to focus on “relations of concern and caring and empathy and trust” rather than a fiction about ideal contracts (Held, 1987, p. 125). Such a suggestion does not mean that we should dismiss the role of contracts in society (“smart” or otherwise), rather it implies that contracts themselves are embedded in social contexts.

← PreviousPage 2 of 13Next →