Offtopic, but I have a really difficult time reading articles like this. I don’t know if this reflects a problem with the style or my ability to focus, but I find it really annoying:
> “SANDHOGS,” THEY CALLED THE LABORERS who built the tunnels leading into New York’s Penn Station at the beginning of the last century. Work distorted their humanity, sometimes literally. Resurfacing at the end of each day from their burrows beneath the Hudson and East Rivers, caked in the mud of battle against glacial rock and riprap, many sandhogs succumbed to the bends. Passengers arriving at the modern Penn Station—the luminous Beaux-Arts hangar of old long since razed, its passenger halls squashed underground—might sympathize. Vincent Scully once compared the experience to scuttling into the city like a rat. Zoomorphized, we are joined to the earlier generations.
This goes on for about seven paragraphs before I have any idea what the article about. I understand “setting the scene” but I can’t tell whether or not to care about an article if it meanders about with this flowing exposition before indicating what its central thesis is.
It seems like a popular style in thinkpieces and some areas of journalism. The author makes a semi-relevant title, provacative subtitle, and five - ten paragraphs of “introduction” that throw you right into the thick of a story whose purpose doesn’t seem clear unless you know what the article is about. Rather than capturing my attention with engaging exposition, I find it takes me out of it. But it must work if it’s so uniquitous; presumably their analytics have confirmed this style is engaging.
It's not the style-- it's just not good writing, but it's trying so hard to be. It's the kind of thing that would show up in a college writing workshop and hopefully get workshopped into something more intelligible. As they say, "Show, don't tell." The passage describes a lot but not in a way that helps you actually visualize any of it, thus it's really hard to follow.
"But, I explained to my work colleagues as the Princeton local pulled out from platform eight and late-arriving passengers swished up through the carriages in search of empty seats, both the original Penn Station and its unlovely modern spawn were seen at their creation as great feats of engineering."
I had to highlight between the commas to get through that one.
Just imagine ... that some day in the future journalism will be AI based, and will generate entire articles tailored to your viewing habits based on extensive psych profiling and AB testing to maximize clicks and screen time!
The content need not be true, but at least everyone will be happy with their preferred writing styles....
What if I dont want the content Im most interested in. It'd probably give me the computer scoentost version of tabloid, but I prefer making myself read things that I don't fully grasp.
That actually sounds quite appealing to me. In recent years I noticed that the writing style of a book had a much bigger part in my ability to derive value from it than its content.
If I could read fiction that is written exactly for me I would love it. And as for "non-fiction", I reference check any particularly interesting claim anyway, so I'd be happy to try and use the AI for that too. The way I see it, reading is much more about exercising the brain in thinking about new things than about learning new facts.
The piece is just overwrought at the sentence level, as in the example below. I think it's partially inspired by trying to sound like an old-style important newspaper columnist, and partially David Foster Wallace. DFWs long sentences are very readable though, because they are conversational, so you can understand them perfectly if you read them as though hearing them aloud.
The existence of writing like this is why analytics (and attention) are not good ways of deciding if style and subject are "working". Clearly many people hate it - like Garlic - many people hate Garlic; Garlic fails the attention / analytics test. Pop Tarts pass!
And yet a world of Pop Tarts is sooooooo boring... And no one makes heart stoppingly good fish stew using Pop Tarts.
This fella may not have written the best piece of the week, we may not remember this piece tomorrow - but I think that the fact that he's attempting to create something gives him a chance of actually getting there. Looking at a dashboard completely kills that in my opinion.
I should clarify: I don't mind "unfocused" writing like this. I can definitely appreciate a creative take on exposition. But I think the introduction of an article is not the most appropriate place to do it. An upfront paragraph - even a few sentences - explaining what is happening would basically resolve this for me.
I actually like the baffler, but I 100% agree. The New Yorker and LRB are similar with the new yorker being far far worse. Its distracting and takes away from the story for me as well, and I love long form journalism.
I appreciate the irony of assuming that the writing style of an article about how data-obsessive engineering cultures strip away the capacity for creative thought and engagement must have been determined by profit-maximizing analytics, especially when the article in question was written for a nonprofit leftist magazine.
A thought: Don't let some of the (valid) criticism alone dissuade you from reading this.
IMO the author makes some very valid points about fuzzy products and endpoints in the current AI/data/ML/magic craze. These are under-articulated elsewhere, because, well hey there's a lot of money flowing! Who wants to be a killjoy and not "get it" (just like in 1999 ;)?
Two more specific points:
1. The descriptions of the CEO are eerily familiar to me. This guy is almost an archetype. Reminds me of a person I've worked with in that role who was also associated with a similar-ish company. It really paints the con-game side of all this.
2. A deeper point (and worth the read for me) was the author's thinking about how all this didn't fit existing needs and workflows and then has a chilling thought:
"It’s possible that the market for a user-hostile data
system that inaccurately predicts the future and turns its
human operators into automatons exists after all, and is
large."
You can make an argument that this kind of thing has already happened in modern customer service and, with greater negative impact, in healthcare. I.e. where the tail of easy metrics and saleable endpoints ends up wagging the dog of quality.
The problem, besides the condescending tone towards everyone around him, is that he doesn't present an understanding of the actual state of the field of AI and deep learning, and what's worse, he cites bad science essays that will misinform more people about how a brain works.
There's a meme going round about how the best way to refute an argument is to 'steelman' it: present the best arguments of the opposing side before refuting them. He doesn't do that here, which is one of the reasons I found it frustrating.
I agree that the way the venture raising market works today rightfully deserves some fair criticism.
By this rate, looks like we need a "Fucked AI", in the style of "fuckedcompany.com". [1]
These people were eating VC hype money to build Hagbard's FUCKUP from the Illuminatus! Trilogy. [2]
Not sure who I feel more sorry for. The smart employees wasting years of their prime chasing some unattainable pipe dream, the VC's who got suckered into pouring their money into some vaporware precog technology, the author trying to disguise a shit river with meandering prose, or my upcoming pay cut when the AI winter sets in.
[2] First Universal Cybernetic-Kinetic-Ultramicro-Programmer (FUCKUP). FUCKUP predicts trends by collecting and processing information about current developments in politics, economics, the weather, astrology, astronomy, the I-Ching, and technology.
In short, nothing that couldn't be revealed from the briefest skim of headlines from tomorrow morning's WSJ.com. One can stay better informed leaving a Bloomberg TicToc (which is partially machine generated) tab open all day.
My takeaway is that the world of the Jim Shinns is rapidly approaching extinction. Deals done poolside at country club dinner dances. Name game shmoozing. And serendipitous encounters on private islands. What was considered the predominant pathway to immortality in Fitzgerald's day.
Viable alternatives exist now. And any business model solely differentiated by prestige will be subsumed by free or near-free competition.
The Jim Shinns of the world are often LPs in VCs like Edison Partners. And/or they have led a major exit for the VC. Either of these situations gives them ample leverage for a "vanity" Series A/B. Even without the pool or country club.
But I found this Sunday AM read enjoyable, articulate, and largely on-point (overlooking a few minor scientific errors).
The core themes here are about the hubris of a rich CEO/founder, the zaniness of the current AI "market," and their resultant effect on a particular NYC startup.
This is a season of "Silicon Valley" (HBO) done east-coast, hedge fund, Ivy League style.
Outside of the firms owned/operated by the real clever boys, I wouldn't be surprised if this describes the vast majority of "AI" efforts unfolding at dozens/hundreds/thousands of companies. Everybody is getting on the bandwagon and either don't have any clue or find out that their customers don't even want what they are selling at the end of the day.
I'd be shocked if anyone in the industry hasn't worked for or with a Jim. Spot-on.
Readit News
> “SANDHOGS,” THEY CALLED THE LABORERS who built the tunnels leading into New York’s Penn Station at the beginning of the last century. Work distorted their humanity, sometimes literally. Resurfacing at the end of each day from their burrows beneath the Hudson and East Rivers, caked in the mud of battle against glacial rock and riprap, many sandhogs succumbed to the bends. Passengers arriving at the modern Penn Station—the luminous Beaux-Arts hangar of old long since razed, its passenger halls squashed underground—might sympathize. Vincent Scully once compared the experience to scuttling into the city like a rat. Zoomorphized, we are joined to the earlier generations.
This goes on for about seven paragraphs before I have any idea what the article about. I understand “setting the scene” but I can’t tell whether or not to care about an article if it meanders about with this flowing exposition before indicating what its central thesis is.
It seems like a popular style in thinkpieces and some areas of journalism. The author makes a semi-relevant title, provacative subtitle, and five - ten paragraphs of “introduction” that throw you right into the thick of a story whose purpose doesn’t seem clear unless you know what the article is about. Rather than capturing my attention with engaging exposition, I find it takes me out of it. But it must work if it’s so uniquitous; presumably their analytics have confirmed this style is engaging.
"But, I explained to my work colleagues as the Princeton local pulled out from platform eight and late-arriving passengers swished up through the carriages in search of empty seats, both the original Penn Station and its unlovely modern spawn were seen at their creation as great feats of engineering."
I had to highlight between the commas to get through that one.
The content need not be true, but at least everyone will be happy with their preferred writing styles....
http://karpathy.github.io/2015/05/21/rnn-effectiveness/
If I could read fiction that is written exactly for me I would love it. And as for "non-fiction", I reference check any particularly interesting claim anyway, so I'd be happy to try and use the AI for that too. The way I see it, reading is much more about exercising the brain in thinking about new things than about learning new facts.
And yet a world of Pop Tarts is sooooooo boring... And no one makes heart stoppingly good fish stew using Pop Tarts.
This fella may not have written the best piece of the week, we may not remember this piece tomorrow - but I think that the fact that he's attempting to create something gives him a chance of actually getting there. Looking at a dashboard completely kills that in my opinion.
Screw the stats! Make what you think is good !
IMO the author makes some very valid points about fuzzy products and endpoints in the current AI/data/ML/magic craze. These are under-articulated elsewhere, because, well hey there's a lot of money flowing! Who wants to be a killjoy and not "get it" (just like in 1999 ;)?
Two more specific points: 1. The descriptions of the CEO are eerily familiar to me. This guy is almost an archetype. Reminds me of a person I've worked with in that role who was also associated with a similar-ish company. It really paints the con-game side of all this.
2. A deeper point (and worth the read for me) was the author's thinking about how all this didn't fit existing needs and workflows and then has a chilling thought: "It’s possible that the market for a user-hostile data system that inaccurately predicts the future and turns its human operators into automatons exists after all, and is large." You can make an argument that this kind of thing has already happened in modern customer service and, with greater negative impact, in healthcare. I.e. where the tail of easy metrics and saleable endpoints ends up wagging the dog of quality.
There's a meme going round about how the best way to refute an argument is to 'steelman' it: present the best arguments of the opposing side before refuting them. He doesn't do that here, which is one of the reasons I found it frustrating.
I agree that the way the venture raising market works today rightfully deserves some fair criticism.
These people were eating VC hype money to build Hagbard's FUCKUP from the Illuminatus! Trilogy. [2]
Not sure who I feel more sorry for. The smart employees wasting years of their prime chasing some unattainable pipe dream, the VC's who got suckered into pouring their money into some vaporware precog technology, the author trying to disguise a shit river with meandering prose, or my upcoming pay cut when the AI winter sets in.
[1] https://en.wikipedia.org/wiki/Fucked_Company
[2] First Universal Cybernetic-Kinetic-Ultramicro-Programmer (FUCKUP). FUCKUP predicts trends by collecting and processing information about current developments in politics, economics, the weather, astrology, astronomy, the I-Ching, and technology.
Some of PreData's recent "insights":
"China Trade War Fears Still Running High"
"Mall Blaze Sparks Outrage Across Russia"
In short, nothing that couldn't be revealed from the briefest skim of headlines from tomorrow morning's WSJ.com. One can stay better informed leaving a Bloomberg TicToc (which is partially machine generated) tab open all day.
My takeaway is that the world of the Jim Shinns is rapidly approaching extinction. Deals done poolside at country club dinner dances. Name game shmoozing. And serendipitous encounters on private islands. What was considered the predominant pathway to immortality in Fitzgerald's day.
Viable alternatives exist now. And any business model solely differentiated by prestige will be subsumed by free or near-free competition.
> Three months later, Predata secured a second round of venture capital funding.
People like Jim Shinn will always find a way. At least that's the argument the author seems to be making.
That's a really embarrassing mistake.
But I found this Sunday AM read enjoyable, articulate, and largely on-point (overlooking a few minor scientific errors).
The core themes here are about the hubris of a rich CEO/founder, the zaniness of the current AI "market," and their resultant effect on a particular NYC startup.
This is a season of "Silicon Valley" (HBO) done east-coast, hedge fund, Ivy League style.
I'd be shocked if anyone in the industry hasn't worked for or with a Jim. Spot-on.
https://en.m.wikipedia.org/wiki/Psychohistory_(fictional)