> “ Can you think of a time in tech when a new method…”
1. Apple/Google exposure notifications during the latest pandemic.
2. The Signal Protocol.
3. Differential Privacy (DWork et al) deployed by the US Census and Apple/Google.
4. Examine David Chaum’s body of work.
5. PIR motion sensors.
6. Elder care monitors.
7. Cities selecting lidar over cameras in traffic applications (motivated by privacy).
8. Retail foot-traffic counters.
9. Stickfigure/silhouette masking in Japanese elder care recordings.
10. Those traffic signs that read, “your speed is X.”
11. On-air or active-session indicators in recording studios.
Jeez it turns out that when you go looking for examples of designed-for-privacy monitors they are EVERYWHERE. Sounds to me like it is a completely reasonable expectation.
There's considerable cynicism in these comments and I have to suspect a puritan bent in most of them. No, no one is watching you. No, these cameras are not here to catch you doing bad things.
The purpose is very clear: reduce roadway mortality by reducing distracted driving. Frankly, I want OTHER PEOPLE to have this system in their car, while they are driving around me. Half of the cars I've driven over the past 3 or so years feature this system, and it is a good one: only occasionally invoked, and effective at observing driver attention. I'm very glad that my family members drive cars with this system.
Now to the privacy concern – yes, that is absolutely a concern. While generally my solution is to have no cameras at all, surely there are other solutions. Car makers have an incredible ability to spy on their customers, and rather than making that impossible by dumbing down the cars, I'd hope lawmakers and public advocates can devise solutions that allow both safe cars and secure data practices. This is not a case where safety in one domain needs to be traded for safety in another.
I'm sure the analysis could be expanded – why don't you give your try at an essay? If you write it, I'll read it.
How much should the essay be expanded? Surely we don't want a 1:1 scale model of reality. We probably should not factor in the conductivity of water and compare against the cost of 1 liter of milk; although if I wanted to attack the integrity of the essay I might lament those factors' absence.
Snark aside, I like that the author chose to compare Switzerland, Germany, and the US. It elegantly illustrated the similarities and differences.
The author is Swiss – your ad hominem attach fails twice. And again, I'll direct you to original point of my above comment and a point made in the essay: if you want to be effective you'll need to look past ideology to see the actually working dynamics of the system. There is no doubt that the Swiss internet service provision system is superior to the US and Germany – no doubt. The essay explains why, an essay you are welcome to read.
Look, as mentioned in the prior comment there are varying spreads of performance across geography (including in Switzerland), and while you might be experiencing one expression of that variance (good internet connection in Utah) you need to look beyond that anecdote to trace the bigger picture and observe performance/cost/reliability beyond your narrow environment.
> "That's not a capitalism problem"
It literally is a capitalism problem, as clarified by the essay; capitalism requires competition to work, and the essay eloquently describes how to accomplish competition in internet service provision.
(Also, what are you talking about, California with "shit internet"?)
I'll briefly reiterate the essay's primary point, for your convenience, which is agnostic to population size: when a natural monopoly exists, the best solution – regardless of ideology – is to make the underlying infrastructure at once singular and public while fostering competition at the level of servicing that infrastructure. That is how Switzerland accomplished its highly functional fiber-optic network, a system that is at once cheaper, more reliable, and more performant than its US and German peers.
This is a perfect example of my earlier comment: you are talking around the subject, cynically dismissing the point of the essay without even addressing it.
I’m seeing a lot of misplaced cynicism in the comments, much of which fails to deal with the subject matter of the article.
The US really does have a capitalism crisis with declining competition — it does not require any form of special intelligence to see that.
Switzerland really does have vastly superior infrastructure — it does not take some stroke of brilliance to see that.
The essay elegantly articulates the why. Even if the anti-public commentariat doesn’t like Switzerland’s strong governance, even if there is a varying spread of speeds/competition or whatever else is being measured, even if one small country is out-performing a big one on many metrics… it doesn’t change the underlying insights of the essay, insights that the US desperately needs to understand.
There really is this bias — it’s crazy. Not just the name “Swiss,” but allusions to the Swiss flag, Swiss typography, other Swiss branding. It almost looks like some of the products are state run.
(Also, the internet connection actually is phenomenally good.)
“Already lost” — Culture doesn’t move towards an end state that is homeostatic, it keeps on evolving. The web will keep on evolving, digital tech and communications will keep on evolving. And as documented in the essay, there has in the past been a role for both identification and proof of personhood, and at present it continues to play a role, and in the future it will continue to play a role; this role shifts and changes, and good designs do emerge.
(I share your cynicism/pessimism regarding the state of the web right now, but as a realist I insist on recognizing continuous change and try to personally overcome the common “end state” fallacy/delusion.)
I think your model of open/closed is incomplete and thus misleading. There are more states to a process than "active" and "inactive," and it is not optimal for the system to simply move processes between those two gross states. The obvious example is non-foreground apps during multitasking. A less obvious example is an app during a background refresh.
"Fully closing" a process is not necessarily cleaner than letting the system allocate intelligently, despite what one's puritanical upbringing might make them believe. (Consider how artists often need a messy space to optimally hold their processes.)
As information flows abundantly, and as information processing flows more abundantly, where will the bottlenecks in the system emerge? It surely won't be in design and production. It probably won't be in chips, infrastructure, and energy (already commodities, increasingly competitive). So... what's the bottleneck? Political will? Human discernment/taste? Raw materials?
I don’t think that comparing LLM’s to the printing press (and radio, film, TV, etc) is an apt analogy, and I don’t think that people have said the same things about the two technologies; the prior technological changes in information dealt with distribution, while this one deals with processing and production.
Recall the notion of a bottleneck, and this distinction will become clear. Those prior technological changes never inverted a bottleneck, and this one does.
Let's imagine that Anthropic/OpenAI fail to manufacture scarcity by villainizing Open Weight models (a sincere probability). What is left for these corporations to prop up their prices, or any margin at all? I expect scaffolding around tool use, supporting bespoke implementation and driving risk down for institutional adoption. (They might even build an insurance tool to protect accountants/lawyers from errors in compounded probabilism!)
A question for economists... It seems plainly clear to me that information and information processing is commodifying (for the first time in human history?). Without the age-old bottlenecks at the top of the value chain, capital will surely flow downwards, right?
Author here. I chose the biology references for two reasons. 1) I love nature 2) It seemed to provide the most distant examples of IR from what most of us are familiar with.
Of course compilers are probably the most obvious example of Intermediate Representation and form the basis for much of the computer science we draw on any time we compose an IR in code. That, to me, made it less compelling for a feature in this brief essay. I like that my example of an IR in Swift was designed, while the biology examples of IR are not deliberately designed but rather landed upon during extremely chaotic evolution. That contrast compelled me to reference biology, instead of the more immediately adjacent body of work around compilers. Of course I’m not the best author… maybe I chose poorly.
Regarding Notion and blocks — I suspect Notion chose the block architecture because it is maximally agnostic (it is kind of an IR in its own right), allowing for rapid development of anything upstream or downstream of those blocks. Maybe one of their early team members was inspired in some way by the field of IR, or otherwise intuited this design.
While I am ideologically aligned with the author (ie, humans matter) and agree with their proposed interventions (progressive taxation, spread the base of capital across society, aggressive antitrust enforcement) and appreciate the thrust of the essay, I’m doubtful of the underlying predicate that human labor will be replaced with AI.
I don’t believe we have seen AI-driven layoffs yet (despite the CEO’s prognostications and suspect justifications). I personally have more to do than ever despite “AI being able to do everything I can do” (‡). I still want to speak to a human at the company I am a client of. And how many AI bloggers even know what a nurse at a hospital does?
I agree with the author’s repeated statement and implication that it is all about the human. The human is the lynchpin. Imagine for a moment this “dead economy,” or instead imagine a virtualized economy that is just incredible with trillions of this and trillions of that, absolute abundance, perfect chemical processing, impeccable design, unlimited resources. How much is that worth without humans?
Further, in many societies humans became irrelevant to labor in its simplest definition decades ago, without becoming economically irrelevant. If the US lost 10% of its able-bodied workforce economists would be primarily concerned with lost consumption, and not production. Apparently this “human labor” is a lot more than its superficial and material interpretation.
What I don’t agree with is the fickleness assigned to the human role. The human role is not a light or arbitrary one. It is the defining characteristic of our societal system that all subsequent characteristics rely upon and are derived from.
(‡ – Despite “AI being able to do everything I can do,” it cannot tie it together, because there is no such thing as agency in LLM’s. Probabilistic processes as n approaches infinity become gobbledygook at best. Deterministic interruptions are a necessity of agency.)