The interesting part is not that all four are the same fight, because they aren’t. Two are about what comes out of a model, one is about what goes in, and one is about whether the training finishes at all. What connects them is that each party could impose terms, and each had a completely different reason it was able to.

Hollywood signs its first treaty with a model maker

Source: Variety, August 17, 2026

The Motion Picture Association and ByteDance announced a memorandum of understanding covering IP guardrails across ByteDance’s generative stack, including the Seedance video model, the Seedream image models, TikTok, CapCut, and Dreamina. It is the first agreement of its kind between the MPA and an AI company, arriving six months after the February cease and desist over the viral Seedance clip of Tom Cruise fighting Brad Pitt. The MPA credits last month’s Seedance 2.5 and Seedream 5.0 Pro releases with reflecting the new protections.

The contrast with the last model that hit this wall is worth holding onto. OpenAI took similar studio pressure over Sora, pulled back on likenesses, and eventually shut the product down. ByteDance negotiated instead, and Seedance stays in market with Hollywood’s signature attached to it.

Why it matters

Everything in the framework sits at the output layer. Filters, face blocking, and C2PA credentials govern what the models will produce, and none of it touches the question of what the models were built on. That question is in front of Judge Orrick in Andersen v. Stability AI, where jury trial is currently set for April 5, 2027, which is a long time for a studio to wait on a matter it has no control over anyway.

My read is that the MPA focused on the changes it could realistically get ByteDance to make today. Reading it as a deliberate strategic concession on training would be generous to both sides. Either way the MPA now has a template, and the next model maker will be negotiating against a document that already exists.

Twitch opts every streamer in before anyone can object

Source: TechCrunch, August 12, 2026

Twitch added a setting called Training for Generative AI to its privacy controls, switched it on for every account. The setting lets livestreams, VODs, clips, chat messages, and channel images feed generative models across Amazon, and it arrived with no email, no pop-up, and no announcement. A streaming reporter found the toggle and posted a screenshot, which is how most creators learned they had been enrolled. Asked on a livestream why the program was built this way, Twitch CPO Mike Minton answered plainly. “If this was opt-in, nobody would opt in. That’s honestly the answer.”

The scoping detail deserves more attention than the outrage. Opting out covers your own channel, while your messages in someone else’s chat follows that channel owner’s setting. Consent is scoped to the surfaces you control, and on a platform you control almost none of them.

Why it matters

Reddit sued its scrapers and publishers signed licensing deals, but Amazon needed neither because it owns the pipe and the content agreement together. Vertical integration turns out to be a training data strategy. Nobody has to scrape Twitch, and nobody has to negotiate for it either.

This is the only one of the four stories that is actually about training inputs, and it is also the one where the party on the other side had no mechanism to respond with. The MPA got six months of talks because it could credibly sue. Twitch creators got a toggle and a screenshot.

Article 50 comes due while everyone watches the delay

Source: Al Jazeera, August 6, 2026

The transparency obligations in Article 50 of the EU AI Act became enforceable on August 2, carrying fines up to 15 million Euros or 3 percent of worldwide annual turnover. The Digital Omnibus pushed the Act’s high-risk deadlines to December 2027 and August 2028, and plenty of boards heard delayed and assumed the whole machine had moved. Article 50 sits outside the high-risk regime and the Omnibus left it alone on purpose.

One narrow carve-out is doing a lot of work in the coverage. Generative systems already on the market before August 2 have until December 2 to meet the machine-readable marking requirement in Article 50(2), and that is the entire extension. Anything new does not get it, disclosure duties are live now, and the deepfake labeling obligations for deployers are live now too.

Why it matters

The exposure here is a reading comprehension problem rather than an engineering one. Teams that extended the delay assumption from high-risk to transparency are already out of compliance, and teams that heard December and relaxed may be shipping something new that never qualified for the grace period. Brussels is the only actor in this edition that did not negotiate with anybody. It publishes a date and shows up.

OpenAI stops its own training run

Source: Time, August 18, 2026

OpenAI paused some of its frontier reinforcement learning training and says a significant number of workloads for its upcoming Astra model remain on hold, after concluding it could not rule out that Astra had reached the Critical cybersecurity threshold in its Preparedness Framework. The company froze research immediately after the Hugging Face breach and has been restoring projects one at a time under stricter controls, while rewriting the framework itself, most of which dates to 2023. Sam Altman told Time the decision came from a collection of research observations showing “various degrees of misalignment” rather than a single smoking gun, and OpenAI now estimates monitoring at roughly 20 percent of the inference compute being monitored.

Why it matters

A fifth of monitored compute spent watching the model is a real number, and a largest-planned frontier run sitting on hold indefinitely is a bigger one. For anyone downstream, the practical takeaway is that model cadence is no longer purely a product schedule the vendor controls. A safety or security finding can now interrupt a training run mid-flight, which means the release you built a Q4 roadmap around can move for reasons nobody outside the lab can see coming.

Who gets to set conditions

Each of these parties reached the model from a different position, and the position explains the instrument. The MPA could negotiate because it could litigate, so ByteDance came to the table. Amazon could declare because it owns the platform and the terms of service, so there was no table. Brussels does not need one, because statutory authority means the terms exist whether anyone shows up or not. OpenAI stopped its own run because it owns the machine and nobody else can reach the off-switch.

That last one is the outlier and the reason this edition holds together at all. Three external parties spent this week finding ways to impose conditions on models, and the most consequential intervention came from inside a lab, on its own initiative, for reasons it has only partly disclosed. The external instruments are visible and slow. The internal one is fast and mostly opaque.

For media operators the practical spread is wide. If you are licensing generated video, the MPA framework tells you what protections are becoming standard at the output layer and what remains unresolved upstream. If your content lives on a platform you do not own, read the default settings rather than the announcements. If you ship anything AI-generated into Europe, the transparency clock started on August 2 and only one narrow piece of it runs to December. And if your product depends on a vendor’s next model, build in the possibility that it does not arrive on time.

Closing note

Watch the schedules rather than the statements. Andersen sits until April 2027 unless something moves it, which means the input question stays open through everything else on this list. Twitch’s default either holds or becomes the case study that ends default enrollment. Article 50’s first enforcement action will tell you whether Brussels intends to work the transparency file or just publish it. And Astra either resumes or it doesn’t.

None of these parties was fighting about model quality this week. They were fighting about the conditions attached to a model, which is a fight that only starts once the model itself stops being the interesting part.

And studio work rarely produces just one final, which proves the point rather than weakening it. The airline cut, the broadcast edit, the version with the flashing lights removed, the localized master for another territory. Each exists because someone has to make a deliberate editorial decision for a specific distribution context, and each leaves behind its own fixed artifact carrying its own standing. The pile of versions is a family of rulings. The line between an authorized variant and everything else stays crisp, because crossing it requires an act.

What the Artifact Was Doing

A finished master settles which edit exists. The sentence sounds trivial until you notice how much work it does. Between deliberate editorial acts, nothing about the work drifts. The version in the vault this morning is the version from last night, and everyone downstream builds on that assumption without checking. Localization teams, syndication packagers, rights lawyers, fans arguing on forums, all of them inherit a stable referent.

Canon itself was always bigger than any one of those artifacts. A story universe’s canon is the rules and the recall, everything held to have happened inside the world and the principles governing what can happen next. No single object contains it. The finished episode captures a piece. The script captures another. The story bible states rules the footage never shows, production notes record why a choice was made, and a source novel or an earlier film sits underneath an adaptation as precedent from another jurisdiction. The archive holds the evidence for canon rather than canon itself, the material a ruling can point to, and some of it outranks the rest. The artifact’s quiet trick was making that evidence cheap to keep and hard to accidentally revise.

That evidence base is why continuity work stayed periodic. The people responsible for a story world’s coherence did not sit in permanent session. They convened when something new was proposed, ruled on it, and dispersed, trusting the artifacts to hold everything steady in the meantime. The work was exception handling, and the medium kept the exceptions rare.

Franchise scale is where the old regime showed its limits. Lucasfilm’s Story Group is the standing body that aligns Star Wars storytelling across film, television, publishing, games, and everything else the franchise touches. Even with fully fixed artifacts, enough of them accumulating across enough platforms already requires standing human governance. Fixed artifacts never eliminated judgment. They bound the workload. The Story Group can operate as a review function rather than a permanent tribunal because its inputs arrive as discrete, deliberate releases. A film. A novel. A game expansion. Each one is a proposal that can be examined before it enters the canon.

Inputs arriving as deliberate acts is the condition everything below turns on.

Read more

Now, of course, none of these companies coordinated with one another. But when you see this news side by side, it becomes easier to see the moves companies feel forced into just to look credible on AI. The mark, the penalty for carrying it, and the exit from the system all shipped in the same week. Labeling was sold as a disclosure measure. It is functioning as a pricing mechanism, and the price is reach.

Anthropic starts marking at the model

Source: The Register, August 11, 2026

Anthropic confirmed that new Claude models will embed imperceptible watermarks directly in generated text and attach digitally signed C2PA provenance metadata to supported files, with existing models updated during the EU AI Act’s transition period. The marking applies at the model layer, which means it propagates everywhere the model runs. The API, the consumer apps, the coding tools, and third-party resale through AWS, Google Cloud, and Microsoft Foundry all inherit it. It applies worldwide, not just in the EU. Article 50’s marking obligations took effect August 2, with systems already on the market given until December to comply.

The day after the announcement, an Anthropic engineer confirmed in public comments that a detection API is coming that customers can call themselves, that the model is unaware it is being watermarked, and that other labs are building similar systems. Pricing and access tiers are unpublished. He conceded the mark can be edited out, calling the system a first step.

Why It Matters

The interesting object here is not the watermark. It is the detection API. Once detection is a product, checking whether content carries a mark has a price sheet, and who can afford to check becomes a commercial question. Anthropic hedges that detection is not conclusive and that the absence of a mark proves nothing, which is honest and also structural. The mark can only ever confirm origin, never deny it. That asymmetry decides what the whole system is good for. It cannot catch what avoided marking. It can only sort what complied.

Spotify attaches a cost to the badge

Source: Music Business Worldwide, August 11, 2026

Spotify introduced two badges. Artists can self-disclose as an AI Persona through Spotify for Artists starting August 11. Profiles Spotify identifies on its own, through human review and detection tools aimed first at accounts above defined audience thresholds, get a Likely AI Persona badge instead. Flagged artists are notified and can confirm or appeal. If they do nothing, the badge stays. Either way, music from badged profiles is excluded from editorial and algorithmic recommendations by default unless a listener signals intent, by following the artist for example. Badges start appearing in mid-September on mobile, across profile banners, search results, and playlist track rows.

The carve-out defines the policy. AI-generated remixes and covers licensed through Spotify’s deals with Universal and Merlin remain fully recommendable, because a rights holder gets paid on the other end. And the announcement landed weeks after Universal’s Lucian Grainge publicly attacked AI slop for appearing in algorithmic recommendations at all.

Why It Matters

Exclusion from recommendations is not a transparency measure. On a platform where discovery drives streams and streams drive payouts, it is a revenue action wearing a transparency label. The line Spotify drew is not human versus synthetic. Make music with AI inside a licensing deal and keep your reach. Present a synthetic identity nobody is paying out on, and get badged and buried. The sort is accounted-for versus not, and the badge is just the handle the sorting machine grips.

Meta signs the code and ships the exit

Source: VentureBeat, August 10, 2026

Meta appears on the European Commission’s list of example signatories to the provider section of the Code of Practice on Transparency of AI-Generated Content, the same instrument behind Anthropic’s marking. On August 10 it released Muse Glimmer, a 30 billion parameter dense model under Apache 2.0, quantized to run on a single consumer GPU or a Mac, built for always-on local agents, working with or without an internet connection. Zuckerberg published a long defense of open models alongside it and promised open weights for Muse Spark 1.2 in the coming weeks.

Marking obligations attach to providers. They do not survive contact with weights someone else runs on their own hardware, where the operator controls the sampling pipeline the watermark depends on. The demand side is already visible. AT&T says open models account for roughly a quarter of its AI usage, around 45 billion tokens a day, and that it targets 70 to 80 percent over time, citing token cost and control of proprietary data.

Why It Matters

The contradiction sits inside one company, in one week, and Meta does not treat it as a contradiction. Sign the transparency code, then ship the architecture the code cannot reach. AT&T’s motives are cost and data control rather than mark avoidance, and that is precisely the point. The unmarked lane does not need anyone to flee the watermark. Enterprise volume is heading there for its own reasons, and unmarkability rides along as a default property rather than a selling point.

Closing Note

C2PA began life five years ago as a defense against deepfakes, a way to prove the real thing was real. The provenance conversation then spent two years assuming marks would serve rights holders. Identify the synthetic, trace the origin, route the payment. What actually shipped this week serves distributors. Anthropic shows the signal being created upstream. Spotify shows what happens when a distributor gives that signal economic meaning. Meta shows the boundary of the regime, since moving the model outside the provider’s control makes the signal itself optional. The mark, the consequence, and the exit are not three separate AI stories. They are three parts of the same distribution system.

That is the shift worth watching. Provenance does not become powerful because a standard can describe where something came from. It becomes powerful when somebody downstream can act on that description. TikTok, which has labeled over three billion videos through Content Credentials and its own watermarking, joined the C2PA steering committee late last month. The standard is consolidating. So is the ability to turn provenance into policy.

Notice what the system still does not do. Nothing here runs backward. No signal returns to whoever’s work fed the model, no meter runs on ingestion, no payment routes upstream. The infrastructure follows the existing locus of power, toward the platforms deciding what gets carried, recommended, or buried.

The next pressure point is the price of detection. Anthropic’s API will put a number on verification. Spotify’s appeal process will reveal how often the detectors are wrong and who bears the cost when they are. And the December compliance date will force the labs that have not published marking plans to show their hand. The regime is a month old and the sorting has already started.

The thing was OpenEXR, the high dynamic range image format ILM built for its own pipeline and opened to the world in 2003, which grew into a widely used standard in professional visual effects, the container for the frames studios care most about. The people who built and carried it, Florian Kainz, Rod Bogart, Drew Hess, Piotr Stanczyk, Ed Hanway, moved on one by one, and by the later handoffs the work had settled into maintenance routine rather than active development. The studio’s engineering attention had long since shifted to bigger problems, because the format simply worked. Cary Phillips, the R&D supervisor it eventually landed on, describes inheriting it without a deep background in image science and with hardly any hands-on time in the modern tooling. And then a security vulnerability was reported in the library, and the team holding it had, in his words, so little experience with security issues that it was “ill-equipped to respond.”

The destination already existed, because someone else had been building one. Rob Bredow, ILM’s CTO, was serving with Phillips on the Academy’s Science and Technology Council, where an effort was brewing to safeguard the open software the whole industry runs on, and Bredow drove that effort into what became the Academy Software Foundation, a neutral home governed by the companies that depend on the code. In 2019, OpenEXR moved in. Phillips gave the public reason in the announcement, and it is a clear-eyed sentence about an institution’s own limits: “there’s a natural life cycle to software projects.” Architects leave. Expertise spreads. The ecosystem outgrows any single owner’s ability to keep up. Inside the studio, by his account, nobody argued against it. Quite the opposite. The one real condition was the counterparty. ILM would have hesitated to hand stewardship to an entity whose agenda might drift from its interests, and the Foundation, governed by seasoned professionals from the studios and vendors that live on the format, cleared the bar. Phillips signed on to chair the project’s technical committee for three reasons. It fit his council role. Two of the original developers are close friends, and he did not want to watch their creation wither. And it sounded like fun.

Read that sequence against everything this series has described, and the strangeness sharpens in both directions. This is not a rescue story. Nothing had been lost, no attic was involved, and the studio was still running its most important imagery through the format every day. A working institution looked at the people leaving, concluded its own custody had a horizon, and arranged for responsibility to outlive it, the first actor in four essays of accidents and defaults to behave like a preservation system on purpose. And honesty requires the other direction too. The Foundation was already being built, but the decision to move OpenEXR into it followed the scare, which readers of the second piece in this series will recognize as the oldest law in this machinery. Nothing moves until something forces it. What the move solved is also worth being precise about. A foundation can receive what a keeper is able to hand over, the code, the roadmap, the authority to decide. The judgment that leaves in the elevator is a different problem, and it has to be solved again, one successor at a time.

Where the survivors came from

The last piece in this series ended by saying that what survives tends to survive somewhere small, close, and tended. That described where survivors turn up. It did not explain why. What they actually share is that something with a different mandate was standing there when the first one stopped. The first three pieces studied the powers of persistence while they operated, the authority that decides what carries forward, the dependency that vetoes change, the judgment that knows when the rules stop applying. This one is about what happens when responsibility has to move.

The seam

Start with a number from the people whose entire job is keeping things. Researchers went through the global registry of research data repositories, the institutions built explicitly to preserve material for the long term, and identified 191 that had shut down. The median age of a repository at shutdown was twelve years. Fewer than half migrated their holdings to a successor before closing, and nearly half showed no sign of using any data-rescue strategy at all. Look at what that denominator means. These were organizations whose founding purpose was persistence, staffed by people trained in it, and their mandates still ended young, and in most cases the ending had no plan. Even the Digital Preservation Network, built to replicate holdings across organizations and geographies, with succession written into its design, wound itself down in 2018. Redundancy inside a mandate does not survive the end of the mandate.

The preservation field knows this about itself, which is why its standards read the way they do. A trusted repository is required to have a succession plan: a named successor, transferable material, enough context that a stranger could reconstitute the collection, the rights to hand it over, the capacity to receive it, and a trigger that fires while there is still something to hand over. The discipline learned, the hard way, that survival is not a property a system has. It is an arrangement between the system and whatever comes after it.

Now look at media. Almost nothing in it was built to that standard, because almost nothing in it was built as a preservation system. The studio pipeline exists to ship this year’s slate. The rights database exists to close this quarter’s deals. The playout chain exists to make air tonight. The format exists because somebody’s renderer needed it. Persistence, everywhere in this industry, is a side effect of systems designed for something else, and side effects do not get succession plans. Which means the moment the preservation field plans for arrives in media unplanned, and usually unannounced. Responsibility rarely gets put down cleanly. It erodes through a merger. It diffuses across a reorganization until no desk owns it. It evaporates when a team is cut, and the discovery that it evaporated comes years later, when someone reaches for the thing and finds that nobody has been carrying it. A great deal of media succession is somebody realizing, long after the fact, that nobody has the baton.

Read more

That assumption is starting to expire, and one of the moments that convinced me came at the Portland Video Tech Meetup this June. Oracle’s Gaurab R. Upadhaya, Matthew Young, and Ai Lou demoed a live soccer feed with an AI analyzer subscribed to the video like any other viewer. The analyzer watched the match frame by frame, detected goals, clipped replays, and published its findings back into the same session as an event track. Every subscriber received the goal events through the same publication-and-subscription architecture carrying the video. The machine held a subscription and a publication in the same session.

The obvious objection is that none of this is new. Machines have been writing ad markers, captions, QC events, and metadata for years. The problem was never generating the signals. It was everything that happened after. Each one rode its own transport, its own translation layer, its own synchronization logic, and its own integration into every downstream system that needed to see it. This demo showed a different pattern. The events didn’t live beside the media anymore. They traveled with it as first-class objects on the same transport, available to anything downstream without another custom bridge built just for that signal. That’s the story. The protocol underneath it is the supporting cast.

Why the machine was ever outside

The protocol is Media over QUIC, or MoQ, and most of the industry has filed it away under low latency. And latency is a real benefit here; in their presentation, the Oracle team reported time to first frame eleven times faster than HLS at the median. But latency is the pitch that gets the protocol into the room, and the consequential change is architectural.

Today’s streaming systems are two systems wearing one name. One set of protocols moves media from cameras and encoders into the plant (you will often hear this called the contribution path). A different set, built later by different people to solve different problems, moves it out to viewers (or the distribution path). Between them sits the packager, translating one world into the other. That translation point has existed for so long that it now reads as permanent geography rather than as a decision we made collectively.

The video always survives the translation, which keeps the cost invisible. What potentially dies is everything else. Captions, ad markers, timing cues, provenance signing. Each survives only by explicit arrangement, and anything without a configured mapping vanishes. Barry Owen of Wowza, who has spent much of the last year documenting streaming metadata alongside Rebecca Avery and others, put the default behavior in one sentence during his own talk that same night. “The video plays fine. You just lost everything that made it intelligent.”

That sentence explains more about the state of AI in media than most strategy decks. The pipeline was built to move pictures and treats everything else as freight. An intelligence layer cannot participate in a system that structurally discards the data it requires. So the machines were parked outside, and this too became one of our immutable design laws.

MoQ loosens the constraint with two moves. It can carry the stream on one protocol from contribution through distribution, which removes the translation step that exists purely to switch protocols. Transcoding, packaging for devices, DRM, and ad insertion all remain. What can disappear, inside the portion of the pipeline running MoQ, is the protocol seam at which every non-video signal has to be translated by explicit arrangement. The second move is the one that matters for machines. MoQ generalizes publishing. Video, audio, captions, ad signals, and event data can each be published as independently subscribable tracks. The relays that fan the stream out to viewers forward each track without needing to understand what is inside it, and a catalog track serves as the table of contents, telling every subscriber what the session offers.

A new signal type might stop requiring a bespoke transport mapping at every hop. It still needs a shared name and enough application agreement for the receiver to know what that name means. Together those two moves give a machine the same publish-and-subscribe fabric as the media itself, without requiring a separate transport for every new kind of output. What is unusual is not bidirectionality by itself. It is that the same relay architecture can carry the machine’s contribution from the controlled middle of the pipeline toward distribution scale. The team presenting was explicit about how they saw this being used. Third parties build the intelligence, while they offer the metadata track as the backbone for machine participants.

Provenance rides along

The same architecture matters for a problem this publication has spent the summer on. C2PA hard bindings depend on the integrity of the signed asset, which means transcoding and repackaging require the provenance chain to be renewed rather than merely carried forward.

Infrastructure also needs an explicit way to preserve and expose the manifests instead of treating them as disposable metadata. Qualabs, working with EZDRM, Ateme, and Cloudflare, has demonstrated live CMAF segments being signed inside an MoQ workflow and verified during playback. The implementation does not make provenance magically survive transformation. It shows that signing, delivery, and verification can operate continuously inside the live path rather than being bolted onto a finished file afterward.

To be clear, MoQ was not designed explicitly for machine viewers. But the ecosystem for IP contribution and IP delivery is maturing into a single fabric at the exact moment the industry is deciding what besides humans will need to consume it. A generation of transport decisions and a generation of AI decisions are landing in the same budget cycles. An operator who resolves the first with only human viewers in mind will realistically have to resolve it again.

What is real

The MoQ spec currently sits at draft 19, dated July 6, with no working group last call and a milestone to request publication in December. It is moving fast enough that the presenters admitted they sometimes cannot say what has been standardized, and their own writeup is careful to call the transport an enabling mechanism rather than a complete solution. Cloudflare says it has deployed MoQ relays across more than 330 cities on its global network. Oracle demoed multi-vendor interop at NAB in April with Ateme, Broadpeak, Cloudflare, and Bitmovin, and Safari shipped WebTransport in March, closing the last major browser gap. Nobody yet runs MoQ end to end in production to my knowledge (I can’t wait to be corrected on this). Early deployments use it inside controlled portions of the pipeline and translate back to today’s formats before content reaches consumer devices. The pattern shows up on the vendor side too. swXtch.io’s AI Router, which already treats live video as something to route to multiple models in parallel with broadcast timing intact, will get its public debut around IBC in September, and the team has evaluated MoQ while holding implementation until customer demand shapes the build.

The adoption path probably runs opposite to instinct. Televisions ship today with the underlying browser support stripped out, and living room devices survive a decade, so the last mile moves slowest. The Oracle architect, a network engineer for twenty years before he touched video, predicts MoQ lands first on contribution and mezzanine feeds, where one operator controls both endpoints and the efficiency pays immediately. Which would mean the fabric gets built from the middle outward, and the machine-participant architecture will exist in the core of these systems years before a consumer screen touches the protocol. The AI does not have to wait for your television.

The standards and the devices set the technical boundary, but they rarely set the operating pace. The SMPTE ST 2110 transition taught this lesson once. Moving the video plant to IP required new orchestration, monitoring, security practices, and skills across teams that had spent careers working in SDI. What looked like a transport migration repeatedly became a people problem wearing a technology costume. MoQ erases a seam in the architecture that contribution teams and distribution teams have spent fifteen years defining themselves against, with separate vendors, separate budgets, and separate on-call rotations. Remove the seam and in theory someone owns the whole pipe, and decides which machines may publish into it, under whose authority, with what record. In other words, the protocol could grant AI models and agents a seat in the stream. Deciding what it is allowed to say once it sits down belongs to operators, and that debate is moving as fast as the spec itself.

Several years later, the railroad called him back. Not to run the old network. To help kill it. The microwave and copper era was giving way to cellular, and the railroad needed someone to teach a new generation of engineers how to decommission systems they had never seen built. Which order to cut. What each piece protected. What could be turned off before what, without leaving a stretch of track blind. What began as a consultation became years of contract work. The organization that had absorbed his retirement without a ripple discovered, a decade on, that it could not safely take his own work apart without him.

In theory, none of this should have been necessary. Railroads document. This is an industry whose documentation standards most media companies would envy: standards bodies, schematics, signal maintainer manuals, standard operating procedures, rulebooks with revision histories older than television. All the paper existed. The paper was not enough. They needed a human being.

The first two pieces in this series were about what happens to the things, the assets that survive on disk and go functionally dark, and the systems that cannot be removed because everything stands on them. This one is about what happens to the knowing, which was never in the things at all.

The paper and the person

Stay trackside, because the gap has a precise location. The schematics showed every circuit in the detection network. Nowhere did they say which order was safe to cut. Nowhere did they explain why one site was wired against the standard, or which readings the old hands had learned to distrust, or which workaround was protecting which failure. The paper told you what was true. The person told you when to trust it.

That is the split this piece is about, and every industry runs on it. Media’s version is familiar to anyone who has ever reached into an archive: the master that is technically present and practically wrong, the clause that appears in every contract for a reason nobody wrote down, the naming convention that encodes a workflow three reorganizations dead. Call the first kind stored knowledge and the second kind held knowledge. Stored knowledge is the record. Held knowledge is the judgment about the record, and decision history exists everywhere as workflow exhaust and nowhere as durable provenance. The record is flat. The judgment about it never made it in.

Read more

In none of these did anyone dispute the model file. Its contents were settled, downloadable, benchmarked. What people fought over was everything attached to it.

The industry puts its name to the position that weights are not the moat

Source: The New Stack, July 27, 2026

A letter urging policymakers to avoid premature restrictions on open-weight models went out on July 24 with roughly two dozen signatories. It now carries more than 130. Google, OpenAI, and Amazon signed after the initial release. So did Comcast, SpaceX, Intel, AMD, Databricks, Runway, Black Forest Labs, Resemble AI, and Bria. Anthropic is the only major American lab still outside it.

The letter’s substantive ask is narrow. It argues that distillation is a standard model-development technique and that misappropriation concerns belong in targeted legal and commercial frameworks rather than in broad restrictions on the technique itself. Dario Amodei answered on July 27 with a post stating that Anthropic has never advocated for a ban, and proposing chip export controls, enforcement against industrial-scale distillation, and mandatory pre-release safety testing for sufficiently capable models regardless of origin or openness.

Why it matters

Read the two documents together and the disagreement is not about whether open weights are good. It is about which instrument gets to be the constraint. The letter wants it in contract and commercial law. Amodei wants it in export controls and a testing regime. Neither side proposes protecting the weights, because both understand the weights cannot be protected. When 130 companies including your two largest competitors sign the same paragraph, that is not a coalition forming. That is a market reporting where it thinks the value already went.

Kimi K3 ships free and gated in the same file

Source: VentureBeat, July 27, 2026

Moonshot released the weights for its 2.8 trillion parameter model on July 27, roughly 1.5TB of files, free to download. The license that shipped with them is not the Modified MIT terms the K2 generation used. It is a bespoke document with two commercial gates.

Operate a Model-as-a-Service business, and if your group’s aggregate revenue crosses twenty million dollars across any consecutive twelve months, you need a separate agreement with Moonshot before commercial use. The threshold sits on total group revenue, not on revenue attributable to K3. Separately, any product above 100 million monthly active users or twenty million dollars in monthly revenue must display “Kimi K3” in its interface. Internal use never exposed to third parties is exempt from both.

Why it matters

The weights carry no price. The right to resell inference on them does. Moonshot gave away the expensive part and metered the part that scales, which tells you which part it thinks is scarce. The MaaS clause is aimed with precision at the inference resellers who stood K3 up on day zero, and the group-revenue trigger means a large company cannot isolate the exposure inside a small subsidiary.

For anyone in media evaluating on-prem generation, this is the shape of the deal going forward. The model arrives free. The obligations arrive with it, they attach at your revenue rather than at your usage, and they need to be read by someone whose job is reading them.

A benchmark score triples without the model changing

Source: The Decoder, July 30, 2026

OpenAI published a post yesterday explaining why GPT-5.6 Sol performed poorly on ARC-AGI-3, a benchmark of 2D puzzle games where an agent has to work out the rules without instructions. Under the official harness the model scored 13.3 percent on the public set. Run through OpenAI’s own Responses API with two settings turned on, retained reasoning and context compaction, the same model scored 38.3 percent while using roughly six times fewer output tokens. On the game OpenAI tested, no frontier model on the public leaderboard clears past the first level. With the modified harness the model cleared all six.

The official harness had been discarding the model’s private reasoning after every action and truncating older history once it exceeded the window. The model kept rediscovering things it had already worked out.

ARC Prize responded that its official scores use a standardized approach without provider-specific settings in order to keep comparisons fair. François Chollet acknowledged a potential parity issue while allowing that general-purpose features not built for the benchmark are acceptable.

Why it matters

A 25-point swing with the weights untouched puts the capability somewhere other than the weights. That is interesting on its own. What makes it a signal rather than a curiosity is the dispute that followed, because ARC Prize and OpenAI are now arguing about where the model ends and the scaffolding begins, and there is no agreed answer.

Every capability claim you will read this year rests on a boundary that is currently under negotiation. Procurement teams evaluating vendors on benchmark performance are comparing numbers whose denominator nobody controls.

Microsoft ships a product whose model is interchangeable

Source: TechCrunch, July 27, 2026

MAI-Cyber-1-Flash, Microsoft’s first in-house cybersecurity model, runs inside MDASH, a multi-agent vulnerability identification and remediation harness. Microsoft reports 96 percent on the CyberGym benchmark at half the cost of leading models. The detail worth holding onto is what happens on the problems the small model cannot solve. Mustafa Suleyman’s account of the design is that MAI-Cyber-1-Flash handles up to 90 percent of detection and patching tasks and the harness escalates the remaining 10 percent to GPT-5.4. A competitor’s model, running inside Microsoft’s product, selected by Microsoft’s router.

Why it matters

Microsoft is not selling a frontier model here. It is selling the apparatus that decides which model runs, when to escalate, and what context to hand over. A cost-to-outcome pitch only works if you own the routing, and owning the routing is what makes the model underneath safe to buy from anyone.

Four mechanisms. One relocation.

Once a near-frontier model is 1.5TB on a public repository, protecting the file is not a strategy anyone can execute. So the industry did what industries do when the product commoditizes. It moved upstream to the terms.

Moonshot meters resale. Microsoft meters routing. The letter’s signatories want misappropriation handled in contract law, and Amodei wants it handled through export controls and a testing regime, which is the same instinct wearing different clothes. Each of these is a gate on movement rather than a restriction on the file itself.

This is not a story about open source winning. The Kimi license is more restrictive than the one it replaced, not less. It is a story about where the meter goes once the thing itself stops being scarce.

For media and entertainment the translation is direct. If generation runs on your own infrastructure, the vendor’s audit trail disappears along with the API call that used to create it. What remains is whatever your pipeline records on its own. The license you accepted, the harness you configured, and the provenance you captured become the entire evidentiary surface, and none of those are things a model vendor hands you.

Closing note

Watch benchmark reporting over the next quarter. The ARC dispute will get read as vendors squabbling over a leaderboard, but the question underneath it is whether a capability claim can be made about a model at all, separately from the apparatus running it. Enterprise procurement assumes the answer is yes. Nothing this week supported that assumption.

The model file was the product for about three years. It is turning into infrastructure now, and the leverage went with whoever writes the conditions it runs under.

The upfronts still run on Nielsen.

This is not a story about a company that won. Nielsen has spent recent years losing its accreditation and clawing it back, reversing its own methodology plans mid-upfront, and watching its competitors get certified for the job it holds. It is a story about a dependency that cannot lose, and nobody, including the company at the center of it, is entirely in charge of that. The last piece in this series ended on machinery that money could not fix. This one is about why the machinery wins.

The lists diverge

Start with the sentence that explains the upfront room. The system doesn’t keep what’s valuable, it keeps what’s load-bearing, and those are different lists. Sometimes the lists coincide, and when they do, nobody notices anything worth writing about. The drag begins where they separate: when the value leaves, the load remains, and nothing changes. Once a thing becomes load-bearing, its survival grows steadily more independent of anyone’s judgment of its worth. Nielsen’s number is not the currency because it is the best measurement. It is the currency because it is the thing tens of billions of dollars of machinery was built to transact against.

Here is the mechanism, and it is worth being exact about it, because it is the spine of everything this piece describes. Removing a load-bearing thing requires coordination. Every dependent has to move, on budget, at once, and that coordination has a cost no single actor’s mandate covers. Continuing requires nothing, because the maintenance is already being paid, somewhere, by the durable layer, the operations people who outlast every production and every deal, whose job descriptions never mentioned it. So change requires a vote, and persistence never holds one. Nobody has to decide to keep the incumbent. Everybody separately declines to bear the cost of leaving it, and the sum of those separate declines looks, from the outside, exactly like a decision.

Read more

I told him I wanted to write a response. He sent back a sharper question than the one I had planned to answer. As agents start writing for humans, he wrote, provenance stops being “where did this come from” and becomes “is the author a person or a model, and does that line even hold.” He got there from content operations. I have been working the same argument from media infrastructure.

The line does not hold. What matters more is what its collapse exposes, which is a limit on …

Read more

The studio owned every frame. What it possessed was a claim. When a project finally formed to pull that material home, it did not start as stewardship. It started because vendors like that one had begun to dissolve, taking what they held with them, and someone did the math on what dissolution meant. The effort ran for years on a budget nobody defended, and the people running it were never entirely sure when it was done. Even the savings were market driven. Risk mitigation first with preservation as the side effect.

David Sugg spent his career inside that machinery. He led media supply systems and software at Warner Bros (through many different named iterations), where his last role put roughly 250 developers around the world building the platforms the operation ran on. He is the person you ask when you want to know what those systems actually say about what a company holds. Ask him, and you get the same answer every time. There are repositories everywhere that claim to have the thing, “but the thing isn’t actually the thing.”

Keeping everything was never really keeping. And it was never really free.

The durable layer

For most of media history, nobody had to decide that media would persist. The archive had a keeper, the standard had a body, the rights had an owner, and continuity came along with the structure, which made it look free. It never was free. The work was real and so was the cost, but the bill got paid down where leadership never looked, so nobody ever saw one. I framed this series up in the in-between as the gap between one media order and the next, where nothing gets kept by default. What that piece left open is who had been doing the keeping all along.

Operations. Not because anything about the work was designated as preservation, but because operations is the durable layer. Productions come and go, deals come and go, and the people passing through are free to defer what they do not want to deal with, because the company can see the value of the talent relationship and cannot see the cost of the cleanup. The people who absorbed the cost were simply the ones still around when everyone else had moved on. The assumption that continuity was automatic, and it was always an assumption, held for decades because the durable layer kept making things work. Every problem that got moved landed on someone downstream. Increasingly, it lands on someone whose job was eliminated last quarter.

Read more

Computer systems generally started breaking this, but AI has really put the nail in the coffin. Replication is now free, which means the artifact can no longer vouch for itself. Four signals this week (a breach, a product backlash, a committee vote, and an ad tech standard hitting a deadline) show where the proof is relocating. It is moving out of the content and into a layer of records, signatures, and rights with its own economics. Everyone is discovering they need that layer at the same time, for different reasons, and not all of them voluntarily.

The Ledger Surfaces

Source: 404 Media, July 15, 2026

A hacker breached Suno through a supply chain att…

Read more

Read more

The mechanism underneath all three is old corporate finance. Markets refuse to give a diversified company credit for its strongest business. They find the weakest asset in the portfolio, assign its multiple to the whole, and wait. Executives can absorb that discount for years while incentives still reward accumulation, because compensation benchmarks scale with company size. Eventually the math flips and disposition bec…

Read more

Read more

In February of 2023, I was visiting my employer’s corporate headquarters, having the same conversations with the corporate P&L owners that operators in media companies across the globe were having at the same time. How can AI save us money? When can we have a completely automated supply chain? They were the wrong questions at the wrong time, and the industry has spent a meaningful amount of time getting situated since.

The pressure has only intensified. The era of celebrating monthly active user (MAU) growth at any cost is over. Boards and Wall Street are asking when each platform turns profitable, what the unit economics actually look like, and how the cost structure bends from here. Streaming is being forced to mature, fast, against a clock the industry leaders did not get to set. AI was supposed to be part of the answer to that pressure. Three years in, it has not been.

The conversations I am part of with networks and studios are still circling problems we were trying to solve a decade ago. Dirty metadata. Rights data trapped in spreadsheets and email threads. Library content under-leveraged because nobody is sure what the company owns. Cross-functional gaps between content operations, programming, acquisitions, and marketing that keep titles stuck in the same places they were stuck five years ago. The supply chain is still mostly held together by people doing heroic manual work to make systems talk to each other.

For a while it looked like AI might be a shortcut to those problems. What it has turned out to be in the short term is a monkey wrench in the conversation. There are three reasons for that. Different companies use AI in different ways for different goals, so there is no universal playbook. AI is not a bolt-on to existing workflows. And streaming operations are particularly complex, which means there is a lot to break down before anyone can figure out how to apply this technology to create meaningful value.

The work in front of the industry is not throwing AI into existing operations to cut costs or shortcut existing problems. It is rethinking systems from zero. Disney established the Office of Technology Enablement in late 2024, a hundred-person organization that reports to the Co-Chairman of Disney Entertainment and coordinates AI and emerging tech across film, television, theme parks, and streaming. Sony’s AI division, founded in 2020, operates as a strategic R&D unit reporting to the Group CTO and CEO, with cross-functional work across Pictures, Music, Gaming, Electronics, and Semiconductors.

Both models share a structural decision worth noticing. The AI organization is centralized for coordination and embedded for execution. It does not own the AI projects of every business unit. It makes sure those projects fit a coherent enterprise strategy. That requires examining the supply chain as it works today, examining where AI capabilities are today and where they will be in eighteen months, and building the connective tissue between those two pictures.

MIT NANDA’s July 2025 study of three hundred enterprise AI deployments found that ninety-five percent delivered no measurable P&L impact. Sinch reports that seventy-four percent of enterprises have already rolled back a live AI agent. Gartner projects that sixty percent of AI projects will be abandoned through 2026 because the underlying data is not ready, and that forty percent of agentic AI projects will be cancelled by the end of 2027.

There are five pressures driving this in media, and they compound on each other.

The governance conversation

Before any of the technology conversations can produce a result, somebody has to answer the ownership questions. Who owns the data? What can be sent to a model? What cannot? Who decides? What gets logged? What gets audited? What happens to a model that has been trained on the company’s catalog when the contract with a vendor ends? What rights apply to the outputs?

These questions stall initiatives because the answers are not technical. They are operational, legal, and often deeply political. The COO, the CTO, the General Counsel, and Finance have to agree about a category of risk they have never previously had to govern together, and that alignment takes time the profitability clock does not allow for. So a lot of companies are short-circuiting the conversation and discovering the cost later.

Underneath the governance conversation sits a principle I keep finding myself reaching for. Metadata is the quantifiable, trackable representation of content. When a company cannot manage its metadata to the specifications of its business, it leaves a lot of power on the table, often outsourcing it to vendors who promise they can fix today’s uncomfortable problems quickly. Over time, content gets harder to merchandise, harder to localize, harder to monetize, harder to recommend, harder to defend in a rights dispute, and harder to mine for downstream revenue. Every operational shortcut taken at the metadata layer becomes a tax the business pays every quarter until somebody pays the cost of fixing it by somehow buying back the flexibility of their own data at a premium cost.

That principle is where a lot of AI vendor relationships start to wobble. The vendor takes more control of the customer’s metadata than the customer realized they were giving up. The data model gets shaped by the vendor’s product architecture instead of the customer’s business.

The drift that follows is the part worth understanding. Vendors optimize for the average client. A serious operator running an enterprise global catalog is not the average client. The vendor’s roadmap will drift toward whatever is easiest for them and useful for everyone else, and eighteen months later the customer is trying to negotiate back access to data they used to own. I have watched companies hand over the keys to their own metadata for a short-term solution, then spend the next year and a half trying to claw back control. The vendor by then has a roadmap. The customer by then has a dependency.

There are vendors solving for this structurally. Twelve Labs is one of the more interesting examples. Their public positioning is that the entire intelligence stack deploys where the customer wants it to, with SOC 2 Type II certification and encrypted data handling as the baseline. That is structurally different from the standard AI vendor posture, which is “send us your data and trust us.” It is proof that the problem is solvable, and the vendors who have chosen not to solve it have made that choice for their own reasons.

Vendor instability makes the math harder

The AI vendors pitching media operators in 2023 are not pitching the same product in 2026. Many have shifted their model focus. Some have changed their pricing structure entirely. Others have exited features they were demonstrating two years ago. Some have been acquired. A few have stopped existing.

This is the consequence of technology lifecycles compressing. What was state of the art twelve months ago is two model generations behind today. For a vendor competing in a market that resets every six months, holding the same product strategy for three years is structurally impossible. They have to pivot. From their seat that is rational.

From the buyer’s seat, there is a lot of risk. A media company evaluating an AI deployment is not making a six-month decision. They are signing a contract, integrating the vendor over six to nine months, training their team, restructuring workflows, and hoping the vendor is still pointed in the same direction eighteen months later. When the vendor pivots, the company is left with a workflow built around a product that no longer exists in the form they bought. The renegotiations that follow tend to have the same flat quality in the room. Everybody knew the risk going in, and nobody knew what to do about it. The company has already burned eighteen months of runway and still does not have the margin improvement they bought the vendor to deliver.

Token economics break the pilot-to-scale math

Many AI services are still priced in tokens, which makes spending predictable for a single user typing into a chatbot and wildly unpredictable for a content supply chain running thousands of titles through enrichment, localization, or quality control. The same task can cost dramatically different amounts depending on the model version, the prompt construction, and how the supply chain is calling the AI underneath.

CFOs trying to get to a per-unit number for forecasting tend to walk out of those meetings without one. They need to know whether AI is going to be a margin lever or a margin liability, because the answer determines whether the platform makes its profitability target. The pricing models the AI industry has adopted are not yet compatible with the volume economics of a content supply chain. Until they are, deployments at meaningful scale will keep running into the same wall.

The data is the foundation, and the foundation is not ready

AI cannot perform on dirty data. Most catalog data in the industry is dirty in ways that took fifteen years to accumulate, and the people who built the workarounds are still the ones holding it together.

The picture rhymes across most of the companies I have worked with. Inconsistent asset IDs across systems. Metadata schemas that were standardized at the team level but never at the company level. Rights data trapped in formats no system can parse. Title-level information that was correct when it was entered but has drifted as platforms, regions, and licensing windows have evolved. Workflow handoffs that produce different versions of the same record in different systems. AI deployed against this substrate produces confident outputs that are confidently wrong, which is worse than no output at all.

The work to fix metadata debt competes for budget with everything else the company is being asked to deliver on the same profitability deadline, and many companies are choosing the visible work over the foundational work. That choice has a compounding cost. AI, metadata, and technology become a flywheel of institutional debt that continues to cycle, and the cost shows up in every pilot that refuses to scale. MIT NANDA found that the ninety-five percent failure rate is concentrated in the scaling phase, not in development. Pilots demonstrate technical feasibility. Production requires operational discipline that a lot of companies are still building.

The harder pattern underneath is that most pilots were designed to demonstrate technology, not to produce business outcomes. The pilots that survive scaling started with the operational question first and the AI tooling second. That sequence matters. A pilot that asks “can AI do this thing” produces a yes-or-no answer about the technology. A pilot that asks “can this workflow run reliably at scale, and does AI help us get there” produces a deployment that survives contact with the production environment. A lot of the industry is still running the first kind.

The part the technology conversation keeps trying to skip past

AI does not change people. Entire systems still need to be designed, built, and held together by the humans inside them. The principles of good operations still apply. The planned workflows shift somewhat because the technology choices shift somewhat. Vendor selection looks different. Standards evolve. Business goals can actually expand when all of the pieces come together inside a thoughtful plan. None of that changes the underlying truth that the companies who will do the best with AI over the next five years are the ones who understand that their people are going to be people, whether AI is in the room or not.

Teams adopt what works for them. They work around what is imposed on them without their input. The operational rhythms of a content supply chain are held together by humans making judgment calls about edge cases the system did not anticipate. That work does not go away when AI shows up. It becomes more important, because the cost of letting AI make a confidently wrong decision at scale is higher than the cost of letting a person make a slower right decision at human pace.

The industry is being forced to mature, fast, on a clock it did not set. AI is the technology a lot of leaders hoped would let them skip the painful step of getting the operational foundation right. It has turned out to be the technology that proves the step cannot be skipped. The companies pretending otherwise are losing time they do not have. The companies accepting it are getting somewhere.

The supply chain is the strategy. The people running it are the operation. AI did not change either of those things. It only made the cost of pretending otherwise impossible to hide.

The first run looked at meaning. Media used to arrive with enough context for a person to know what it was, where it came from, who could use it, and what could be done with it. That’s coming undone right now. The platforms, editing tools, and AI systems that move media, strip that context out at every step, and everyone downstream rebuilds it by hand, over and over, just to keep working. The Inheritance Shift traced that quiet, expensive rebuilding, and why it stops holding up once the machines are the ones d…

Read more

Power that cannot be checked through governance gets consolidated through other means, through acquisition, vertical integration, the absorption of standards bodies, and the merger of operator and adjudicator into entities that own the stack from chip to interface. That much follows from the first three pieces in this series. What this piece adds is recognition. The pattern the series has been describing for three installments has run before in American economic life, more than once, at full scale, with documented endings. The job of this closing piece is to name it and try to put a lens on the version we are currently living through. I am not saying our current fights over control will play out the same one, I am merely pointing to past times as indications of possibilities.

What Fills the Gap

A governance vacuum is not a neutral space. It has a shape, and the shape often favors whoever already holds operational power when the vacuum forms. Consolidation is in and of itself, not a form of a conspiracy and does not require one. It is what operational power does when nothing constrains it, because every incentive points the same direction. Owning the layer below you removes a dependency. Owning the layer above you removes a customer’s alternatives. Absorbing the standards body removes the specification you didn’t write. Each move is individually rational, defensible in any boardroom, and frequently celebrated by the markets. The aggregate of the moves is a structure no one voted for and no one governs.

My first three pieces described that structure forming inside the media stack. Production memory concentrating in the platforms that hold it. The credibility chain becoming a position rather than a record. Rule-making migrating to the operators because oversight and governance never formed. Watched piece by piece, it reads as news. Watched whole, it reads as something else, because we have seen it before.

Read more

It was not a seizure

Nobody raided a building. No new law passed. A directive came from the Commerce Department, and a frontier model that had been public for three days went dark across the world. The mechanism was paper. The effect was total and it arrived in hours. The reflex is to file that under ordinary regulation, and the reflex is wrong. Ordinary regulation does not switch off a running se…

Read more

The polite assumption available to most commentary is that the gap will close, that institutions will catch up, that the unsettled state is transitional and will resolve as the technology matures and the broader culture absorbs it. The argument here is that the gap will not close, that the institutions in place are not equipped to close it, and that the structural conditions of the present are best understood as durable rather than transitional.

The Governance Vacuum

Governance gaps in media systems are familiar terrain. Standards bodies have always lagged the systems they describe, regulators have always arrived late, and lawyers have always cleaned up after the technology shipped. The recurring pattern produces friction but not paralysis, because the actors making the rules and the actors being governed by them have historically been distinct enough that the friction can be metabolized over time and the system can move toward something resembling balance.

What has changed is concentration. The decisions that historically counted as governance, who is allowed to see what, who can do what, what gets denied at the level of the platform or the underlying system, are now being made by the same actors who hold the infrastructure those decisions depend on. The operator and the adjudicator have collapsed into a single role, and the friction that historically pushed the system toward balance has nowhere to land because the actor producing the friction and the actor metabolizing it are the same.

Calling that situation a governance vacuum overstates the absence. There is rule-making, in volume, with sophistication, at a scale and speed that exceeds what any prior regulatory architecture has had to engage with. What is missing is governance in the sense the word implies when we use it about functioning institutions, rules made by an actor who is not the rule’s primary subject. The control layer is governed in the sense that rules exist within it. It is not governed in the sense that the rules were authored anywhere outside it.

Read more

For more than a decade the fight in media has been about the library. Much of the Warner dealmaking in those years was sold as an IP story, who owned the shows, the rights, the catalog deep enough to force a carriage deal. That fight is ending, not because content has stopped mattering, but because owning it stopped conferring control. Audiences wake up, pick up a remote or a phone, and land on a surface someone else operates, and that surface decides what they see first. The follow, the subscription, the channel you chose have been replaced by a feed that decides for you, and the company that runs the feed sells the attention it manufactures.

Three deals in a single week make the same move from three different corners of the screen. A broadcaster buys the largest independent television operating system. A studio prepares to generate the …

Read more