Redistributed

All Articles

Concepts, patterns, and warnings — everything in the knowledge base.174 articles

Explainers(18)

View section →

Deep dives into AI systems, written for humans.

AI in Healthcare: Who Benefits from the Diagnostic Revolution

A clinic in rural Tamil Nadu has no radiologist. The nearest one is four hours away. A patient arrives with a cough that won't quit and a chest X-ray taken on aging equipment. An AI system reads the scan in seconds, flags a suspicious mass, and recommends follow-up. In a well-funded Boston hospital, the same AI system runs alongside a team of specialists who barely need it. Both sites use the same algorithm. The outcomes couldn't be more different.

explainerhealthcareaccess

Synthetic Data and the Ouroboros Problem

AI labs have a resource problem. The models keep getting bigger, but the planet's supply of high-quality human-generated text is finite. The entire indexable internet has already been scraped — books, articles, forums, code repositories, academic papers, social media posts, government filings. The well is running dry. So the industry has landed on an elegant-sounding solution: have AI generate its own training data. The trouble is, this solution contains its own undoing.

explainersynthetic-datatraining

World Models vs. Large Language Models: Two Visions of Machine Intelligence

There is a quiet civil war in AI research, and the stakes are not just technical. On one side: large language models, the architecture behind ChatGPT, Claude, and Gemini, which have consumed hundreds of billions of dollars in compute and crowned a handful of companies as the gatekeepers of machine intelligence. On the other: world models, a fundamentally different approach that aims to learn how reality works — not by predicting the next word, but by building internal simulations of the world.

explainerworld-modelsllms

Common Pitfalls(9)

View section →

One misconception at a time, dismantled carefully.

No, AI Doesn't Learn Like a Brain

The "neural" in neural network was a marketing decision from the 1940s. As Gary Marcus argues in his critical appraisal, the resemblance to biological brains is superficial at best (Marcus, 2018). The analogy has been productive in some ways — it inspired architectures that genuinely work — but it misleads in consequential ways when taken literally. Understanding where the metaphor breaks down helps us reason more clearly about what these systems can and cannot do.

pitfalllearningneuroscience

The Scaling Laws Plateau

Since 2020, the AI industry has operated on a seductive premise: just make it bigger. More parameters, more data, more compute, better results. The scaling laws published by Kaplan et al. at OpenAI seemed to confirm this with mathematical precision — model performance improves as a smooth, predictable function of scale (Kaplan et al., 2020). Two years later, DeepMind's Chinchilla paper refined the ratios, arguing that training data should scale proportionally with parameters (Hoffmann et al., 2022). The message to the industry was clear: spend more, get more.

pitfallscalingmodels

Paper Reviews(15)

View section →

Accessible summaries of significant AI papers.

Documentation Reviews(7)

View section →

Critical reads of AI product docs, TOS, and model cards.

Policy & Governance(22)

View section →

AI regulation, standards, and institutional choices.

AI and Antitrust: Why Competition Law Struggles with Foundation Models

A handful of companies control frontier AI development. OpenAI, Google DeepMind, Anthropic, Meta — and increasingly few others — determine what foundation models exist, what they can do, and on what terms the rest of the world gets to use them. This concentration is not incidental. It follows from the economics of the technology: training frontier models requires billions in compute, proprietary datasets at scale, and engineering talent that a few employers have locked up.

policyantitrustcompetition

Artificial Dissuasion: Can Nuclear Deterrence Logic Apply to AI?

The Cold War produced the most consequential piece of game theory in human history: the doctrine of mutually assured destruction. Two superpowers, each capable of annihilating the other, reached a stable equilibrium in which neither used its weapons — precisely because the weapons existed. Deterrence worked not through goodwill but through mathematics. The logic was brutal and elegant: any first strike would trigger a retaliatory second strike, so no rational actor would strike first.

policyai-safetydeterrence

Institutional Decoherence: When Technology Outruns Governance

The complaint that "regulation can't keep up with technology" has become a cliché. Like most clichés, it gestures toward something real while obscuring the mechanism. The problem is not that governance is slow and technology is fast. The problem is that different parts of society — political systems, legal frameworks, cultural norms, economic structures, international institutions — adapt to technological change at different rates, creating gaps between them. Those gaps become exploitable. And the actors best positioned to exploit them are the ones who least need the protection that governance provides.

policygovernanceinstitutional-decoherence

The Academic Brain Drain: Who Pays to Train AI's Best Minds?

The pipeline works like this. A student enters a public university. Taxpayers subsidize the tuition. Federal grants fund the lab. A professor — herself trained at public expense — supervises the dissertation. Five to seven years later, a newly minted AI PhD walks out the door and into a corporate lab offering four times the salary, a hundred times the compute, and a signing bonus that exceeds the annual budget of the research group that trained them.

policyacademiabrain-drain

Systems & Access(19)

View section →

The infrastructure, economics, and standards that shape AI access.

AI and Education: Tutoring, Testing, and Who Gets Left Out

In 1984, the education researcher Benjamin Bloom published a finding so striking it became known simply as "the 2 sigma problem." Students who received one-on-one tutoring performed two standard deviations better than students in conventional classrooms. That's the difference between an average student and one in the 98th percentile. The problem was cost: you can't give every student a private tutor. Bloom's paper was essentially a challenge — find a way to deliver tutoring-level instruction at classroom scale.

infrastructureeducationaccess

The Environmental Cost of AI

Every time you ask an AI model a question, a server somewhere draws electricity, generates heat, and consumes water to cool itself back down. Multiply that by hundreds of millions of daily queries, add the months-long training runs that created the model in the first place, and factor in the mining, manufacturing, and shipping of the hardware — and you begin to see the outline of an environmental footprint the industry would rather not discuss.

infrastructureenvironmentenergy

Labor & Work(12)

View section →

AI and the future of work — displacement, augmentation, and power.

AI and the Creative Industries: A Redistribution Case Study

If you want to understand how AI redistributes value, look at the creative industries. Not because artists are uniquely sympathetic — though they are — but because the mechanism is unusually visible. Human creators make work. AI companies scrape that work to train models. Those models generate competing output at near-zero marginal cost. The original creators lose income. The AI companies capture the surplus.

laborcreative-industriesart

Community(3)

View section →

Contributors, conversations, and open questions.

Projects(16)

View section →

Tools, collectives, and initiatives advancing AI redistribution.

Other(53)

View section →

Uncategorized articles.

The Copyright Status of AI-Generated Art: Still Unresolved

Can AI-generated images, text, music, and other creative works be copyrighted? Almost three years into the generative AI boom, the answer remains unsettled — and the uncertainty itself is reshaping who captures value from creative work. This question sits alongside the training data copyright fight (see [[AI Copyright Rulings Are Shaping the Training Data Economy]]) as the other half of a legal pincer that will define the economics of [[AI Creative Industries]].

the-wirecopyrightart

AMI Labs: LeCun's Billion-Dollar Bet Against LLMs

Yann LeCun, the Turing Award winner who spent a decade building Meta's Fundamental AI Research lab (FAIR), has launched AMI Labs — Advanced Machine Intelligence — with a $1.03 billion seed round at a $3.5 billion pre-money valuation. It is Europe's largest seed round ever. The company, headquartered in Paris with offices planned in New York, Montreal, and Singapore, is building "world models" based on LeCun's Joint Embedding Predictive Architecture (JEPA) — a fundamentally different approach from the large language models that dominate the industry.

the-wireworld-modelsopen-source

Big Tech's AI Layoffs: Spending Billions, Cutting Thousands

Between 2022 and 2025, the five largest technology companies in the United States eliminated more than 100,000 jobs. Google cut over 12,000 in January 2023, then continued trimming through 2024 and into 2025, restructuring entire divisions to funnel resources toward AI. Meta laid off 21,000 workers across two rounds in late 2022 and early 2023, with Mark Zuckerberg declaring 2023 the "year of efficiency." Amazon cut 27,000 positions over roughly the same period. Microsoft eliminated more than 10,000 roles in January 2023. Smaller but significant rounds followed at each company — a few hundred here, a reorganized division there — a steady background hum of displacement that rarely made front pages after the initial shock wore off.

the-wirelaborbig-tech

DeepSeek and the Price of Open AI

In late December 2024, a Chinese AI lab called DeepSeek released V3, a large language model that matched or exceeded the performance of models costing ten to fifty times more to train. The company claimed a training cost of approximately $5.6 million — a figure that, even if it understates the full investment, represents a fraction of the hundreds of millions spent by OpenAI, Google, and Anthropic on their frontier systems. A month later, DeepSeek released R1, a reasoning model competitive with OpenAI's o1, and published it with open weights. By January 27, 2025, the DeepSeek app had surpassed ChatGPT as the most downloaded free app on the US App Store.

the-wireopen-sourcechina

Nvidia's $3 Trillion Question

In June 2024, Nvidia briefly surpassed Microsoft and Apple to become the most valuable company on Earth. Its market capitalization crossed $3 trillion — a figure that would have seemed hallucinatory five years earlier, when the company was primarily known for making graphics cards for gamers. By the end of fiscal year 2024, Nvidia's data center revenue had surged to over $47.5 billion, up from roughly $15 billion the prior year. The company now commands an estimated 80-90% market share in AI training and inference chips.

the-wireinfrastructurehardware

OpenAI's For-Profit Conversion: The Mission Pivot

OpenAI was founded in December 2015 as a nonprofit research lab. The founding charter was explicit: develop artificial general intelligence that is "safe and beneficial to humanity." (For a close reading of how that mission has been operationalized, see [[OpenAI's Terms of Service]].) Early backers — Elon Musk, Sam Altman, Peter Thiel, Reid Hoffman, and others — contributed over $1 billion. The tax-exempt structure was the point. AGI was too important, too dangerous, too consequential to be governed by shareholder returns.

the-wiregovernanceopenai

Say No to Suno: Artist Coalition Targets AI Music's Royalty Drain

A coalition of artist-rights organizations published an open letter on February 23 urging the music community to reject Suno, the generative AI music platform. Signatories include the Music Artists Coalition, the European Composer and Songwriter Alliance, the Artist Rights Alliance, and prominent figures like songwriter Helienne Lindvall and artist David Lowery. The letter calls Suno a "brazen smash and grab" — unauthorized AI trained on human artists' work, generating output at industrial scale.

the-wiremusiclabor