When astronauts return from space, they are carried out of their capsule because the full effect of Earth’s gravity, felt instantaneously upon landing, is strong enough to break bones. After months of floating in zero-G, your muscles atrophy, bone density drops, fluids redistribute, and your balance and sense of spacial orientation recalibrate. It takes time for the body to readjust to what used to be normal. Heartbreak feels a lot like coming back down to Earth because falling in love is akin to taking flight: Known to many by his Latin name only, Cupid (desire), the Greeks called the god of love Eros and often portrayed him as a handsome young man with rosy skin and large, lovely wings. The Greeks visualized this feeling of love by adding “pt” to “eros”, forming “pteros”, meaning “wing”: We take flight when we surrender to romance, letting it take us from over here to over there. Longing for someone is an invitation for us to travel from the status quo to a new world; perhaps, the reason why eroticism can be thrilling, embarrassing, or repulsive to talk about is because it offers us adventure — one of novelty and danger.

All of these promises are very Trumpian: Big, bold ideas that are (in many cases) widely supported by the vast majority of Americans, but are also obviously vague, in some cases not realistic, and very "strongman" in their attitude (like promising to punish protesters and conduct mass deportations).

Being broadly effective, but somewhat inefficient, is better than being narrowly efficient, but less effective.

Targeting can increase the scale of effects, but this study suggests that the cheaper approach of not targeting so specifically, might actually deliver a greater financial outcome

As Wiberg’s findings point out, the problem with targeting towards conversion optimisation is you are effectively advertising to many people who were already going to buy you.

If I only sell to IT decision-makers, for example, I need some targeting, as I just can’t afford to talk to random consumers. I must pay for some targeting in my media buy, in order to reach a relatively niche audience.  Targeting is no longer a nice to do, but a must have. The interesting question then becomes not should I target, but how can I target effectively?

What they found was any form of second or third-party data led segmenting and targeting of advertising does not outperform a random sample when it comes to accuracy of reaching the actual target.

Contextual ads massively outperform even first party data

We can improve the quality of our targeting much better by just buying ads that appear in the right context, than we can by using my massive first party database to drive the buy, and it’s way cheaper to do that. Putting ads in contextually relevant places beats any form of targeting to individual characteristics. Even using your own data.

The secret to effective, immediate action-based advertising, is perhaps not so much about finding the right people with the right personas and serving them a tailored customised message. It’s to be in the right places. The places where they are already engaging with your category, and then use advertising to make buying easier from that place

Even hard, sales-driving advertising isn’t the tough guy we want it to be. Advertising mostly works when it makes things easier, much more often than when it tries to persuade or invoke a reluctant action.

Thinking about advertising as an ease-making mechanism is much more likely to set us on the right path

If your ad is in the right place, you automatically get the right people, and you also get them at the right time; when they are actually more interested in what you have to sell. You also spend much less to be there than crunching all that data

Traces of Things (2018) is a video installation and series of thirty digital prints that explore what happens when history is remembered and re-remembered. Past moments in time are re-lived through the eyes of an artificial intelligence model, trained on images Ridler sourced from public and private Maltese archives, to create its own depiction of what it thinks should be included in an archive of Maltese photography. The process of how an AI recreates realities through a process of deliberating and deeming what is important echoes the selective and subjective human process of repeatedly recreating memories each time they are recalled.

Every time we remember something we are also actively recreating it. Traces of Things, a video installation and a series of thirty digital prints, explores this loop - remembering and revision - by passing through moments of history through an artificial intelligence model trained on material from a variety of public and private Maltese archives. At what point do the images change from one thing to another? At what point do they break down into nothingness?

I took photographs that showed historic Malta from a variety of sources, some primary, some second hand, some public, some private,  to create my own dataset of what the island has looked like. There are similar issues with using archives to the issues that exist with datasets: what we have deemed important enough to count and quantify means that what is recorded is never simply “what happened” and can only show sometimes a very narrow or very incomplete view

Traces of Things shows how quickly meaning can break down if only a narrow dataset exists. Human memory works by filling in the blanks, creating essentially confabulations, a type of memory error where a person creates fabricated, misinterpreted, or distorted information, often found with dementia patients. In this piece memories are mixed with inventions; inventions are modelled on memories. There is a term used often in computer science and machine learning called “overfitting” which is used when a model cannot create new imagery but constantly remembers just one thing, the link to dementia again coming through.

current technology still has the elements of transformation each time something is recalled, or played, or copied, that become encoded into it. These moments are compelling: the creation of a copy where things start to slowly transform.  In Traces of Things, boats turn into houses, houses into mountains, mountains into harbours. This power to metamorphose without real control is something that within an art context is more closely associated with work that deals with biology or nature, than the digital, which tends to be all smooth and clean. The style that comes out is ruined, decaying and decomposed - something antithetical to a certain  digital art. But at the same time, to my mind, beautiful. The link, then, to the biological processes - the neuroscience - that have inspired much of the research into artificial intelligence as memories and matter are constantly recalled and revised.

Instead of altering the fundamental fabric of reality, maybe it is used to create better versions of features we have used for decades. This would not necessarily be a bad outcome. I have used this example before, but the evolution of object removal tools in photo editing software is illustrative. There is no longer a need to spend hours cloning part of an image over another area and gently massaging it to look seamless. The more advanced tools we have today allow an experienced photographer to make an image they are happy with in less time, and lower barriers for newer photographers.

You’re also not learning anything this way. Part of what makes art special is that it’s difficult to make, even with all the tools right in front of you. It takes practice, it takes skill, and every time you do it, you expand on that skill. […] Generative A.I. is only about the end product, but it won’t teach you anything about the process it would take to get there.

I feel lucky that I enjoy cooking, but there are certainly days when it is a struggle. It would seem more appealing to type a prompt and make a meal appear using the ingredients I have on hand, if that were possible. But I think I would be worse off if I did. The times I have cooked while already exhausted have increased my capacity for what I can do under pressure, and lowered my self-imposed barriers. These meals have improved my ability to cook more elaborate dishes when I have more time and energy, just as those more complicated meals also make me a better cook.

I am wary of using an example like cooking because it implies a whole set of correlative arguments which are unkind and judgemental toward people who do not or cannot cook. I do not want to provide kindling for these positions.

Plenty of writing is not particularly artistic, but the mental muscle exercised by trying to get ideas into legible words is also useful when you are trying to produce works with more personality. This is true for programming, and for visual design, and for coordinating an outfit — any number of things which are sometimes individually expressive, and other times utilitarian.

This boundary only exists in these expressive forms. Nobody, really, mourns the replacement of cheques with instant transfers. We do not get better at paying our bills no matter which form they take. But we do get better at all of the things above by practicing them even when we do not want to, and when we get little creative satisfaction from the result.

The swarms of online surveillers typically only know how to detect clearly stated opinions, and the less linguistic jouissance the writer of these opinions displays in writing them, the easier job the surveillers will have of it. Another way of saying this is that those who read in order to find new targets of denunciation are so far along now in their convergent evolution with AI, that the best way to protect yourself from them is to conceal your writing under a shroud of irreducibly human style

Such camouflage was harder to wear within the 280-word limit on Twitter, which of course meant that the most fitting and obvious way to avoid the Maoists was to retreat into insincere shitposting — arguably the first truly new genre of artistic or literary endeavor in the 21st century, which perhaps will turn out to have been as explosive and revolutionary as, say, jazz was in the 20th.

Our master shitposter has perfectly mirrored the breakdown of sense that characterizes our era — dril’s body of work looks like our moment no less than, say, an Otto Dix painting looks like World War I

The fundamental question in Friday’s ruling boiled down to: ‘who decides,’ courts or agencies? The conservative majority’s answer — courts — affects everything from clean air to drug safety to student loans, the broad landscape of government regulation. And that power matters more than ever now that Trump, who had appointed 28 percent of federal judges by the time he left office, has the prospect of naming more in a second term.”

A synthesizer produces music very differently than an acoustic instrument. It produces music at the lowest level of abstraction, as mathematical models of sound waves.

Once we started understanding writing as a mathematical object, our vocabulary for talking about ideas expanded in depth and precision.

An idea is composed of concepts in a vector space of features, and a vector space is a kind of marvelous mathematical object that we can write theorems and prove things about and deeply and fundamentally understand.

Synthesizers enabled entirely new sounds and genres of music, like electronic pop and techno. These new sounds were easier to discover and share because new sounds didn’t require designing entirely new instruments. The synthesizer organizes the space of sound into a tangible human interface, and as we discover new sounds, we could share it with others as numbers and digital files, as the mathematical objects they’ve always been.

Because synthesizers are electronic, unlike traditional instruments, we can attach arbitrary human interfaces to it. This dramatically expands the design space of how humans can interact with music. Synthesizers can be connected to keyboards, sequencers, drum machines, touchscreens for continuous control, displays for visual feedback, and of course, software interfaces for automation and endlessly dynamic user interfaces. With this, we freed the production of music from any particular physical form.

Recently, we’ve seen neural networks learn detailed mathematical models of language that seem to make sense to humans. And with a breakthrough in mathematical understanding of a medium, come new tools that enable new creative forms and allow us to tackle new problems.

Heatmaps can be particularly useful for analyzing large corpora or very long documents, making it easier to pinpoint areas of interest or relevance at a glance.

If we apply the same idea to the experience of reading long-form writing, it may look like this. Imagine opening a story on your phone and swiping in from the scrollbar edge to reveal a vertical spectrogram, each “frequency” of the spectrogram representing the prominence of different concepts like sentiment or narrative tension varying over time. Scrubbing over a particular feature “column” could expand it to tell you what the feature is, and which part of the text that feature most correlates with.

What would a semantic diff view for text look like? Perhaps when I edit text, I’d be able to hover over a control for a particular style or concept feature like “Narrative voice” or “Figurative language”, and my highlighted passage would fan out the options like playing cards in a deck to reveal other “adjacent” sentences I could choose instead. Or, if that involves too much reading, each word could simply be highlighted to indicate whether that word would be more or less likely to appear in a sentence that was more “narrative” or more “figurative” — a kind of highlight-based indicator for the direction of a semantic edit.

Browsing through these icons felt as if we were inventing a new kind of word, or a new notation for visual concepts mediated by neural networks. This could allow us to communicate about abstract concepts and patterns found in the wild that may not correspond to any word in our dictionary today.

What visual and sensory tricks can we use to coax our visual-perceptual systems to understand and manipulate objects in higher dimensions? One way to solve this problem may involve inventing new notation, whether as literal iconic representations of visual ideas or as some more abstract system of symbols.

Photographers buy and sell filters, and cinematographers share and download LUTs to emulate specific color grading styles. If we squint, we can also imagine software developers and their package repositories like NPM to be something similar — a global, shared resource of abstractions anyone can download and incorporate into their work instantly. No such thing exists for thinking and writing. As we figure out ways to extract elements of writing style from language models, we may be able to build a similar kind of shared library for linguistic features anyone can download and apply to their thinking and writing. A catalogue of narrative voice, speaking tone, or flavor of figurative language sampled from the wild or hand-engineered from raw neural network features and shared for everyone else to use.

We’re starting to see something like this already. Today, when users interact with conversational language models like ChatGPT, they may instruct, “Explain this to me like Richard Feynman.” In that interaction, they’re invoking some style the model has learned during its training. Users today may share these prompts, which we can think of as “writing filters”, with their friends and coworkers. This kind of an interaction becomes much more powerful in the space of interpretable features, because features can be combined together much more cleanly than textual instructions in prompts.

Deleting sections of a bill holding companies accountable for making data-driven decisions that could lead to discrimination in housing, employment, health care, and the like spurred a strong response from civil society organizations including the NAACP, the Japanese American Citizens League, the Autistic Self Advocacy Network, and Asian Americans Advancing Justice, among dozens of others.

In a letter this week to E&C Democrats, obtained by WIRED, the groups wrote: “Privacy rights and civil rights are no longer separate concepts—they are inextricably bound together and must be protected. Abuse of our data is no longer limited to targeted advertising or data breaches. Instead, our data are used in decisions about who gets a mortgage, who gets into which schools, and who gets hired—and who does not.”

these provisions contained generous “pro-business” caveats. For instance, users would be able to opt out of algorithmic decisionmaking only if doing so wasn’t “prohibitively costly” or “demonstrably impracticable due to technological limitations.” Similarly, companies could have limited the public’s knowledge about the results of any audits by simply hiring an independent assessor to complete the task rather than doing so internally.

I wanted to take care of everyone’s feelings. If I made them feel good, I would rewarded with their affection. For a long time, socializing involved playing a weird form of Mad-Libs: I wanted to say whatever you wanted to hear. I wanted to be assertive, but also understanding and reasonable and thoughtful.

I really took what I learned and ran with it. I wanted to master what I was bad at and made other people happy. I realized that it was: bad to talk too much about yourself good to show interest in other people’s hobbies, problems, and interests important to pay attention to body language my job to make sure that whatever social situation we were in was a delightful experience for everyone involved

More than anything I felt like I had been airlifted into a surreal parallel universe, in which everyone is wealthy and on vacation and having beautiful children who go on field trips to aquaria. The inbox in question belongs to Jane Appleseed, and one wonders whether Jane knows her private life is being used to sell hardware and promises.

Layers upon layers of references are stacked together in a single post, while the posts themselves fly by faster than ever in our feeds. To someone who isn’t “chronically online” a few dislocated images or words may trigger a flash of recognition – a member of the royal family, a beloved cartoon character – but their relationship with each other is impossible to unpick. Add the absurdist language of online culture and the impenetrable algorithms that decide what we see in our feeds, and it seems like all hope is lost when it comes to making sense of the internet.

Forget words! Don’t think! In today’s digitally-mediated landscape, there’s no need for knowledge or understanding, just information. Scroll the feed and you’ll find countless video clips and posts advocating this smooth-brained agenda: lobotomy chic, sludge content, silly girl summer.

“With memes, images are converging more on the linguistic, becoming flattened into something more like symbols/hieroglyphs/words,” says writer Olivia Kan-Sperling, who specialises in programming language critique. For the meme-fluent, the form isn’t important, but rather the message it carries. “A meme is lower-resolution in terms of its aesthetic affordances than a normal pic because you barely have to look at it to know what it’s ‘doing’,” she expands. “For the literate, its full meaning unfolds at a glance.” To understand this way of “speaking writing posting” means we must embrace the malleability of language, the ambiguities and interpretations – and free it from ‘real-world’ rules.

Hey guys, I just got an order in from Sephora – here’s everything that I got. Get ready with me for a boat day in Miami. Come and spend the day with me – starting off with coffee. TikTok influencers engage in a high-pitched and breathless way of speaking that over-emphasises keywords in a youthful, singsong cadence. For the Attention Economy, it’s the sort of algorithm-friendly repetition that’s quantified by clicks and likes, monetised by engagement for short attention spans. “Now, we have to speak machine with machines that were trained on humans,” says Basar, who refers to this algorithm-led style as promptcore.

As algorithms digest our online behaviour into data, we resemble a swarm, a hivemind. We are beginning to think and speak like machines, in UI-friendly keywords and emoji-pilled phrases.