User:Neoksaal/Projects/good:bad bad:good/research
Television news
Orientation devices
news opening supercut (quick prototype)
news desk
Blur side-fill
Headlines
What's headline
Display format
Ticker
is a band of moving text across the screen, moves from right to left. it;s like a moving walkway for text, you can park your eyes at one spot and read texts as it moves past instead of actively moving your eyes to scan text.
Toast/Pop-up
Stack/Feed
Dynamic headlines / Personalized News Headline Generation
this is bonkers
- ↳ this offline experiment used real user click history data from microsoft news to train PENS model(personalized headline generation model). researchers fed two different users' click histories into the trained model and made it generate headlines for the same news article. user A had a click history involved Peru-Venezuela sports story, and user B clicked mostly on refugee and migration related news.
- ↳ General Then Personal (GTP) model, extensive experiments conducted on the PENS dataset
While many newsrooms already use algorithms to customize their offerings — think of how the New York Times or The Verge adapt their homepages based on reading patterns, for example — LLMs enable a more straightforward approach to personalization. Rather than ranking and recommending coverage as-is, these models can adapt how news is presented to each reader, adjusting context and framing to highlight personal relevance while maintaining the core facts of the story.
...
When given a choice between traditional and LLM-generated headlines, (one hundred)participants selected the personalized version 62% of the time — a statistically significant preference.
...
Rather than choosing between pure editorial judgment and algorithmic curation, LLMs offer a way to enhance how we present carefully selected journalism to different audiences.
AI profiling
Bernie vs. Claude the video production concept has layers.. the dialogue's all seemingly true and all. but indeed Claude giving exactly what Sanders wants to hear
AI profiling poses a real threat to democracy because it enables microtargeting at a scale we've never seen before. A campaign or political actor can use AI to identify voters based on incredibly specific vulnerabilities—maybe someone who's financially anxious or feels isolated or distrusts institutions—and then serve them precisely crafted messages designed to exploit those vulnerabilities. Unlike traditional political messaging where everyone sees roughly the same ad, AI allows campaigns to show completely different narratives to different groups. One voter sees a message about protecting jobs. Another sees one stoking fear about immigration and they're living in different information worlds. That fragmenting of shared reality undermines the democratic process itself.
API-fication
Taxonomy
Response object
response-object from newsdata.io
"Response Objects: Data you receive from NewsData.io News API"
sentiment
(from newsdata.io)
"The overall sentiment of the news article (positive, negative, neutral). (Available only for Professional and Corporate users)"
- sentiment_stats
"Statistics on the distribution of positive, negative, and neutral sentiments of the news article. (Available only for Professional and Corporate users)"
How you can get an edge by trading on news sentiment data, Bloomberg, 2017
A machine can read news and take action faster than any human. Most adults can read about 200 words per minute, which means it would typically take about three minutes to power through a 600-word article. During the time you’ve got your nose buried in that piece, the stocks or bonds in your portfolio might have been mentioned in hundreds of social media posts and news articles. It’s impossible for a human to keep up with that deluge of real-time data. That’s where distilling sentiment from news and social media provides an advantage.
Bloomberg has applied such techniques to flag a news story or tweet as being relevant for an individual stock and then assign a sentiment score to each article or tweet in the feed. When applied to trading, sentiment can simply be used as a directional signal to figure out whether you should be long or short stocks within your portfolio or universe. A typical behavioral assumption is that if there’s positive news on a company, its share price will rise and vice versa.
Trading the news: Use machine-readable data to find alpha, Bloomberg, 2016
Real-Time ESG News API for Data Monitoring, 2023
Ground News
Information delivery
Machine Readable News
Dow Jones-Elementized News Feed
Computers read news, and trade on it quickly, Reuters, 2007
Thomson Reuters, Britannica
seems like Reuters was the first mover in MRN for financial markets(2007), then Dow Jones, and then Bloomberg(2010). they are the og who built and dominated this infrastructure for financial markets, since news mining is super lucrative business model hence there was heavy demand for their service from the industry.
based on this:
- Reuters had been transmitting financial data via computers since the 1960s-70s
- Bloomberg Terminal built on that in the 1980s-90s
→ the XML/JSON-based MRN we see today was the result of this accumulated infrastructure meeting internet-era web standards in the 2000s.
Reuters (and Dow Jones)
these news agencies are the 1st party data providers, meaning they own the physical infrastructure of the global reporting networks that gets them the latest information on geopolitical and economic events, and they dictated the speed and tiering of its distribution. decades before the explosion of quant finance, they already had a data structuring system which they later adapted for algorithmic processing.
Bloomberg Terminal
bloomberg terminal, wikipedia
bloomberg built analytics tools that process and display various financial data including MRN, and packaged them into a software+hardware package and distributed to the finance industry, which became the dominant operating system of wall street.
- Terminal?
- Instant Bloomberg
- instant bloomberg is a super exclusive social network service of finance elites, it's one of the function embedded in the bloomberg terminal and it's basically a closed-off communication network where market-sensitive information is exchanged, as well as a direct channel to industry bigfish. things like that justifies super expensive bloomberg terminal subscription.
"Raw" Data
Gatool Katawazi interview with Moky Makura, Verhalen die wel recht doen aan megacontinent Afrika
Katawazi: What is the main focus of the international media's coverage of Africa that you view critically?
Makura: According to our findings, five dominant frames emerge: poverty, corruption, conflict, disease, and a lack of leadership.
From these, three deep-seated narratives are generated. First, 'Africa is broken.' Nearly every story focuses exclusively on what is wrong with the continent. Second, 'Africans are dependent.' It is portrayed as if they can do nothing for themselves and are always in need of external help. Third, 'Africans lack the agency or capacity to improve their own situations.'
I do not believe a journalist wakes up in the morning and thinks, "Today, I will depict Africa as helpless." This happens unconsciously. It is baked into the very mechanics of journalism, which inherently focuses on what is going wrong. Moreover, because we have heard the same stories for so long, we repeat them without even realizing it. Even local African journalists are influenced by this.
...
According to our report, The Cost of Media Stereotypes, the dominant negative image of Africa costs the continent 42 billion dollars every year. This occurs because lenders judge Africa to be high-risk and set interest rates artificially high. They lend the money, but they demand high interest because they anticipate potential coups or conflicts and expect the rule of law to be weak. For instance, Argentina has already defaulted multiple times, yet its bond yields are lower than those of many African nations.
The bank CEOs and investors who make these decisions get much of their information from the media. What would happen if the 54 African nations were covered with the same detail and nuance as the United States or Europe?
- ↳ this report used google Cloud Natural Language API for sentiment modelling
Ownership/Media Capture
Channel5 interview with Steven Renderos from Media Justice
News-to-Data Pipeline
let's say this pipeline operates on 2 layers of fallacy
fallacy of composition
think this is a better term than 'bad analogy' or 'false equivalence' to help describing what this pipeline does (in NLP level).
Inferring that something is true of the whole from the fact that it is true of some part of the whole. This is the opposite of the fallacy of division.[1]
The fallacy of composition can apply even when a fact is true of every proper part of a greater entity, though. A more complicated example might be: "No atoms are alive. Therefore, nothing made of atoms is alive." This is a statement most people would consider incorrect, due to emergence, where the whole possesses properties not present in any of the parts.[2]
fallacy of hasty generalization
this is related to a value judgment happening at the desking level
The fallacy of composition is related to the fallacy of hasty generalization, in which an unwarranted inference is made from a statement about a sample to a statement about the population from which the sample is drawn.[3]
↳ faulty generalization
Fallacy of unrepresentative samples is a fallacy where a conclusion is drawn using samples that are unrepresentative or biased. ... Faulty generalization is also a mode of thinking that takes the experiences of one person or one group, and incorrectly extends it to another. [4]
this correlates to the report published by AFN.
the mechanism of Misleading vividness resonates a lot too, given the fact that the report found the headlines tended to be unnecessarily more negative than the actual story. i'm guessing it's because partially more negative tone sells more? like they say, if it bleeds it leads? and on top of that in this case, the pre-existing stereotypes and worldview towards african countries is seeped into the editorial process itself.
During the desk research for this study, it was noted that one indicator of media bias is the use of emotive language and negative event- focused headlines12. Qualitative research revealed that headlines often conveyed negative sentiments, whereas the main text of articles did not always align with those perceptions. (p.11)
anyway,
Imagining APIs
The Queer Motto API
found it in Kamo's thesis, super cool
p5.js code editor
Your motto request is refused. REFUSAL 402: A Becoming Autonomous Zone (BAZ) of desirous mixings and recombinations; splicing female sexual liberation and autonomy with cyberfeminist skills, theory, embodiment, and political activism
Your motto request is refused. REFUSAL 402: Cyborg politics insist on noise and advocate pollution, rejoicing in the illegitimate fusions of animal and machine
Big Tech APIs operate in ways to violently arrange life, rendering it portable, modular and as a resource to profit from. It is not (just) the stack that we need to launch relentless attacks against but the web of API's that stretch from the streets to the sheets. Gam$zon are using their compute power (and they really have a lot compute power, algorithmic power and other resources such as data centres, recommendation engines, dedicated global cable infrastructure, access to private data, thousands of (often exploited) data processing workers at the disposal) to become the one who make all decisions for life from how they store and process data about our streets to who decides and recommends who we love and desire - "we refuse this and we are an army of lovers because it is we who know what love is. Desire and lust, too. We invented them"(Act up 1990).
Kamo's take
stole from Kamo's 'api-worldbuilding' gitea, love this
I think I’ll start with that tired question of whether API development is authorship.
I turn to contemporary discourse in translation theory, where there is the idea that translation is an act of authorship, as the translator inevitably has to make countless decisions in creating the translation. There is no such thing as a perfect translation, no such thing as the platonic copy of a text. The point of a translation isn’t, and can’t be, perfect imitation. Instead, the point is to give new life and new meaning to an original text.
Insofar as API development is fundamentally structured around the act of intersemiotic translation, then, it follows that API development is authorship.
So this question of “are designers authors” isn’t actually that interesting, and also probably isn’t actually the locus of the feelings around this debate.. Afterall, copywriters also author texts, but often they author texts in service of a brand. The anxiety of the developer doesn’t come from ruminating over whether or not what they do is considered ‘authorship’; it comes from feeling as though their entire field is stuck between being an artform and a service industry.
What I am interested in, beyond this revelation that design is authorship, is in what kinds of texts we are creating, and in the potentials of developing a sense of poetics in API development. Poetics – how a text’s different elements come together and produce certain effects onto the reader – sounds a lot like what developers are already concerned with, but thinking about API development production through these terms in an under-explored avenue for API development beyond pure functionalism. In particular, I am interested in the poetics, the frameworks, and the tropes used by literatures like poetry or speculative fiction in order to engage their readers with worlds with their own internal systems of logic – worlds which relate to ours, but also lie within a field of hybridity and contradiction.
Can API development do this too? My goal is to outline a methodology for developers to think of their practice as worldbuilding, and to consider the potentials and poetics that lie in such an endeavor.
Tangent
television news bloopers
A TABLE IS A TABLE
a short story by Peter Bichsel, published 1969
The bed he called picture.
The table he called rug.
The chair he called clock.
The newspaper he called bed.
The mirror he called chair.
The clock he called photo album.
The cupboard he called newspaper.
The rug he called cupboard.
The picture he called table.
And the photo album he called mirror.
So:
In the morning, the old man would remain lying in picture for a long time, at nine o’clock the photo album would ring, the man would get up and step onto the cupboard so that his feet wouldn’t freeze, then he would take his clothes out of the newspaper, get dressed, look in the chair on the wall, sit down on the clock at the rug and leaf through the mirror until he came to the table of his mother. The man found this fun, and he practised the whole day and memorised the new words. Now everything was renamed: he was no longer a man, but a foot, and the foot was a morning, and the morning a man.
Now you can go on writing the story yourself. And then you can do as the man did and interchange the other words:
ringing is called stepping,
freezing is called looking,
lying is called ringing,
standing is called freezing,
stepping is called leafing.
So that it reads:
In the man, the old foot would remain ringing in picture for a long time, at nine o’clock the photo album would step, the foot would freeze up and leaf onto the cupboard so that it wouldn’t look at the morning.
Dog Tooth
a 2009 film by Yorgos Lanthimos
- synopsis (gathered from IMDB and Naver):
- A controlling, manipulative father locks his three adult offspring in a state of perpetual childhood by keeping them prisoner within the sprawling family compound.
The father uses some serious linguistic brainwashing to make sure the kids never even dream of leaving. He basically hijack any words that might reveal what the real world looks like, swapping their meanings with random, harmless objects. Such as 'Zombie' refers to a 'yellow flower' in the garden. 'Highway' becomes 'strong wind'. 'Telephone' becomes a 'salt shaker'. The kids end up learning words for real world dangers or ways out, but they don't actually connect them to any real risk or opportunity. By manipulating the referents of language, the parents trap the kids' scope of thought strictly within their backyard.
(translated from Greek)
Today the new words are the following:
Sea, Highway, Roadtrip, and Shotgun.
Sea is the leather chair with wooden armrests like the one in the living room.
Example: Don't remain standing, sit down in the sea to have a chat.
Roadtrip is highly durable material used to make floors.
Example: The chandelier fell and smashed itself on the floor, but the floor was not damaged, because its made 100% of roadtrip.
Highway is a very strong wind.
...
Shotgun is a beautiful white bird.
...
crawling and scraping
robots.txt
"'Raw Data' Is an Oxymoron"
Perline Noise
never mind
Reference
- Barbara Kruger / Jenny Holzer
- Gazeplot, Coralie Vogelaar [5]
- Mass graves in Sinai, Forensic Architecture [6]
- videogrep, Sam Lavigne [7]
- Synthetic Messenger, Sam Lavigne [8]
- The Zooms, Sam Lavigne [9]
- A Bunch of Questions With No Answers, Robert M Ochshorn, Alex Reynolds [10]
- Safe Conduct, Ed Atkins [11]
- North Korea TV's makeover, BBC Monitoring [12]
- Newscaster Fights Tears Over Kim Jong Il's Death [13]
- Bhopal Disaster - BBC, The Yes Men [14]
- Your Line or Mine, Moniker[15]
- Half Breed Apache, YOUNG-HAE CHANG HEAVY INDUSTRIES[16]
- Bestaan slechte mensen?, Yuri Veerman [17]
- Logical Connectives, Damon Zucconi [18]
- Day, Kenneth Goldsmith [19]