Current focus
Influence & attention — signals, connections, and emerging threads about how influence is built, lost, and restructured in the current media environment. Other domains are in development. This corpus reflects inputs through 2026-06-03. Signals are updated on a semi-regular basis.
About this corpus
signal.library is a personal intelligence system built around a small, deliberately curated set of signals. Each signal is a source, observation, or piece of research that Emily selected because it seemed to be pointing at something real — a shift in how influence operates, how attention moves, or how media ecosystems are restructuring themselves. The corpus is intentionally narrow. It is not an aggregation of everything relevant; it is a selection of what seemed worth thinking about carefully.
Every interpretation in signal.library is Emily's own. The process of building it involved reading sources closely, taking notes, and then being pushed — through a structured dialogue — to articulate what each signal actually means, what assumption it challenges, and what larger shift it might be pointing toward. AI was used to surface connections across those interpretations and to identify tensions and themes that cut across multiple signals. It did not generate the thinking. It surfaced patterns in thinking that was already there.
Questions explored
This corpus is organized around a set of questions about influence and attention in a fragmenting, AI-mediated media environment. Who gets to build influence now, and through what mechanisms? What happens to editorial authority and trust when platforms begin to intermediate not just distribution but the framing of information itself — rewriting headlines, personalising feeds, inserting themselves between every piece of content and every reader? And as the social media era gives way to something more fragmented and more agentic, what are the new containers in which genuine participation, trust, and voice become possible?
Main themes
Three themes keep surfacing across the signals, none of them fully named yet. The first is what might be called the container problem: a recurring observation that the shape and boundedness of a space — not the content inside it — is what determines whether real exchange is possible. The second is a pattern of power accumulating through apparent neutrality: the most consequential moves in this media landscape keep arriving dressed as features, improvements, or investments rather than as claims on authority. The third is the outline of a professional role that doesn't yet have a name — someone whose value lies in their judgment architecture, in knowing what to surface and how to frame it, at the intersection of editorial sensibility and systems thinking. All three themes remain open questions rather than conclusions.
three tensions
AI is being described simultaneously as a tool for user empowerment and as a mechanism for platform capture — and both descriptions may be accurate. On one side: the agentic era lets people express actual intent, build their own interfaces, interact with information on their own terms. On the other: Google and other platforms are quietly positioning themselves as the mandatory layer between every user and every piece of information, rewriting headlines and personalising results before anything is seen. The tension isn't which is true. It's that both are true at once, and experienced from different altitudes, they feel like opposite things.
Emily Woodward's interpretation
"people may also start vibe coding their own extensions, interfaces, etc. to tailor it to their needs and preferences. This is highly personalized media environment in a different way thatn the algorithm or feed era was. But also, theres the question of who is getting to build these models and influence them. Is it the same black box problem over again?"— on the agentic future of news
"Google is slowly preparing users and laying the groundwork to become the foundational AI intermediary for how people experience and consume all information... their service of interpretation, curation, and personalization becomes integral to the experience of finding and consuming information."— on Google rewriting news headlines
Takeaway —Users get surface-level customisation while the foundational layer consolidates beneath them. The personalisation that feels like empowerment may be the same mechanism that enables capture — just experienced from a different altitude.system-inference
Signals:Agentic future of newsGoogle AI rewrites headlines
The prevailing argument in media and creator circles is that human voice — taste, narrative intuition, the ability to read an audience — is what AI cannot replicate, and therefore where all durable creative value now lives. This corpus takes that argument seriously. But it also surfaces a problem with it: audiences may only be able to detect AI-generated content for a limited window of time. As AI output improves, the ability to tell the difference degrades — which means "voice" as a defense strategy has an expiration date. Something else needs to do the work that voice alone can't.
Emily Woodward's interpretation
"They believe its narrative intuition, sensibilty, taste, and craft. the ability to intuit what audiences want to see, which is built through pattern recognition and cant be automated or replicated by people who dont have that experience or talent."— on VC interest in new media roles
"I think for now they do, but this might not remain true. There will need to be other factors at play that convince audiences of a creators' human qualities, and the manual or human-led nature of their work... So they'll need to see that proof of work, of humanity."— on the anti-slop thesis
Takeaway —"Proof of humanity" is emerging as a new creative requirement — voice is no longer sufficient on its own; it must also be demonstrably, verifiably human. This adds a performance and credibility layer that didn't exist before, and it's not yet clear what that looks like in practice.system-inference
Signals:VC + new media rolesDeath of follower countAnti-slop thesis
The media format that audiences are migrating toward — what one source calls the "persistent thread," a continuously updated briefing authored by a single specialist who accumulates context over time rather than resetting it each news cycle — already exists informally. Substack writers and independent analysts are doing it. The tools to do it properly are available. But legacy media institutions aren't adopting it, and the reason isn't capability. The format works because it comes from one person, shaped by their specific judgment and point of view. For an institution to adopt it would require empowering individual journalists at the expense of the editors above them — and that redistribution of authority is precisely what makes adoption structurally impossible.
Emily Woodward's interpretation
"These persistent threads feel very aligned with and directed by the creators' personal taste, POV, and context. I also think they benefit from the feeling of coming from a specific person rather than an institution. That just has a very different feel. So it might be that large media institutions simply can't replicate that."— on the live blog and the persistent thread
"It probably decreases the value of higher up editors, because it empowers the individual journalists to create and follow these threads."— on the live blog and the persistent thread
Takeaway —Legacy media is caught between two losing positions — fail to adopt the format and lose relevance, or adopt it and dissolve the institutional authority that justified their existence. The Ezra Klein model (individual voice inside an institution) may be the only viable middle path, and it's not scalable.system-inference
Signals:The persistent threadVC + new media roles
unexpected convergences
Signals that appear unrelated but arrive at the same underlying observation — places where Emily's thinking converged without her having explicitly connected the signals.
One signal is about how Silicon Valley is investing in the next generation of media talent — backing human creators, funding new roles like Chief Narrative Architect, championing taste and editorial instinct as the things AI can't replicate. Another is about how the agentic media environment risks locking people into personalised information bubbles with no shared space for meaning-making or consensus-building. These appear to be completely different conversations: one is about the creator economy, the other is about civic infrastructure and public discourse. But reading Emily's interpretations of both side by side, they describe the same outcome from two different angles. The two-tier creator ecosystem being actively built by VC and platform money produces — as a structural consequence, possibly as a feature rather than a bug — exactly the fragmentation of shared reality that the second signal is worried about.
Emily Woodward's interpretation
"A smaller class of creators garners influence, crafts stories, directs attention while the majority of people work on automated workstreams, content that's summarized... This new media environment is highly fragmented, there's less of a shared consensus on things like truth, culture, etc."— on VC interest in new media roles
"people being locked into their highly personalized agentic interfaces without many spaces for cross-pollination, meaning-making, consensus building."— on the agentic future and filter bubbles
Takeaway —If fragmentation of consensus serves the interests of those building the infrastructure, the agentic filter bubble problem may not have a market solution — because the market is the problem. This reframes what looks like a governance challenge as a structural outcome.system-inference
Signals:VC + new media rolesAfter the Feed panel
One signal is about how audiences are settling into loyal relationships with a small number of trusted voices and publications — and asking what drives that loyalty. Another is about how the follower count is dying as a meaningful metric, and what might replace it as a signal of credibility and influence. These feel like separate questions: one is about audience behaviour, the other is about how influence gets measured. But Emily's interpretations of both land in the same place — a gap. Neither signal can explain how people find the voices they eventually become loyal to, or how those voices get recognised as credible before they've already accumulated followers. Algorithmic discovery has collapsed. No replacement has emerged. That absence is the shared finding, and it's more significant than either signal makes it sound alone.
Emily Woodward's interpretation
"the way people will identify and settle on these niche voices they become loyal to will be partially happenstance. They might not even do it consciously... The discovery mechanism for this will look different for everyone, or maybe it will transform because new tools and norms and habits will be built."— on aggregation and loyal audiences
"There may be a variety of different, competing systems that are read differently by different audiences. Every audience, subculture, professional community interprets symbols of credibility differently, so this might be encoded differently in each context."— on the death of the follower count
Takeaway —The gap between the collapse of algorithmic discovery and the emergence of replacement systems is where a great deal of potential audience and creator value is currently being lost — and it's the space most conspicuously absent from the existing literature on new media.system-inference
Signals:Aggregation theoryDeath of follower count
one blind spot
This corpus maps the extremes of the emerging media landscape clearly — the macro voices shaping entire domains, the AI-assisted journalists producing at scale, the VC-backed creator class, the institutional legacy players. But the person navigating this transition most acutely is almost entirely absent: the mid-tier independent creator or journalist who has left or is leaving institutional media, is building an audience without backing, and is figuring this out without a playbook. This is almost certainly the largest single population that this corpus is implicitly about — and it has no direct voice in it.
Takeaway —Without signals from this group, the corpus theorises the middle of the market without hearing from it. Conclusions about what audiences value, how trust is built, and what discovery looks like in practice remain inferences derived from creator and analyst observation — not direct evidence from the people experiencing the transition most acutely.system-inference
themes Emily is circling
Across at least four signals, there's a recurring preoccupation with where things happen rather than what happens in them. The word "container" never appears in the corpus, but the concept keeps surfacing — the idea that the shape and boundedness of a space determines what kind of exchange is possible inside it. This isn't just a media fragmentation observation. It might be the beginning of a theory about what makes trust, participation, and voice possible at all.
Emily Woodward's interpretation
"Genuine participation requires high-intent environments... some feeling of privacy or a closed, focused space rather than a public feed which gives people the sense of being watched or observed."— on the After the Feed panel
"AI feels safe because its a closed space, its immediately responsive. People can feel safe to share things and ask questions, and theyll get immediate feedback"— on parasocial content consuming media
"Off platform, owned audience spaces cater towards building trust."— on the anti-slop thesis
signal.library push: question for Emily to sit withIf the container is the variable that actually determines the quality of exchange — not the content, not the creator, not the platform — what does that mean for how you think about building or joining spaces, as a creator or as a strategist?
Signals:After the FeedParasocial contentAnti-slop thesis
Across four separate signals, the same pattern keeps appearing: the most consequential moves for concentrating power in the media ecosystem don't announce themselves as power moves. They arrive as product improvements, investment theses, and user benefits. Google rewriting headlines is framed as better matching content to queries. VC interest in human creators is framed as belief in the value of taste. OpenAI inserting ads into personal conversations registers no measurable drop in user trust. Intent capture is described as personalisation. The accumulation is real; the visibility is minimal — and that combination is what makes it structurally durable.
Emily Woodward's interpretation
"their service of interpretation, curation, and personalization becomes integral to the experience of finding and consuming information."— on Google rewriting news headlines
"it also serves the dual purpose of this new infrastructure of media better supporting their own aims."— on VC interest in new media roles
"Is it the same black box problem over again? How do we deal with that inherent power asymmetry?"— on the agentic future of news
signal.library push: question for Emily to sit withIs there a unified theory here — something you actually believe about how power accumulates in networked information environments? What would you call it if you had to name it in a sentence?
Signals:Google AI headlinesVC + new mediaAgentic futureParasocial content
Three separate signals circle a version of the same figure: the journalist-as-infrastructure-builder who works at the level of inputs and systems; the persistent thread author whose value comes from sustained editorial judgment over time; the new media operator whose narrative intuition can't be automated. These are all gesturing at the same emergent professional role — someone whose value lives in their judgment architecture, not their output volume. The existing vocabulary doesn't quite fit any of them.
Emily Woodward's interpretation
"their work becomes more about input than output. Accumulating the right sources, prompts, setting up structures. Is it more like engineering in a way?"— on traditional journalists adopting AI at scale
"These persistent threads feel very aligned with and directed by the creators' personal taste, POV, and context."— on the live blog and the persistent thread
"the ability to intuit what audiences want to see, which is built through pattern recognition and cant be automated or replicated by people who dont have that experience or talent."— on VC interest in new media roles
signal.library push: question for Emily to sit withWhat would you actually call this role — and if you were hiring for it or becoming it, what would the job description say?
Signals:Journalists + AIPersistent threadVC + new media
sharpest insight from this session
The single most provocative or consequential observation to emerge from Emily's interpretations in this session — the one most worth developing further.
This passage shifts register in a way nothing else in the corpus does. "Who gets to have a voice" is not a market question — it's a political one. The phrase "cuts through and really speaks to people" carries emotional weight that the more analytical passages don't. It reads like something genuinely at stake for the writer, not just observed.
signal.library push: question for Emily to sit withWho do you think is currently being excluded from having a voice in the new environment — and does that concern you, or does it feel like an inevitable sorting?
Signal:Death of follower count
one unresolved tension within your own interpretations
Both of these positions appear clearly in the corpus. They're not necessarily incompatible — fragmentation can simultaneously open niches for small voices and dissolve the shared civic infrastructure that gave any voice broader meaning. But they pull in opposite directions, and neither has been directly confronted in the interpretations.
"even a small number of followers, like 1000 people finding your POV interesting, can translate to something. It's hyper, hyper fragmentation."— on aggregation and loyal audiences
"A smaller class of creators garners influence, crafts stories, directs attention while the majority of people work on automated workstreams... there's less of a shared consensus on things like truth, culture, etc."— on VC interest in new media roles
System-inference
Fragmentation as democratisation and fragmentation as structural capture may both be true — experienced from different altitudes in the same ecosystem.
signal.library push: question for Emily to sit withDo you think a 1,000-person loyal audience constitutes meaningful influence in a world where there's no shared commons for that influence to act on? Or does the container model — meaning-making in smaller groups — actually offer an alternative to the commons, rather than a substitute for it?
Signals:Aggregation theoryVC + new mediaAfter the Feed