Seeing is no longer believing.
This breaks into a preamble and a subject-matter on context.
Preamble
In terms of lineage, I'm Ukrainian on my father's side. Candidly, that fact lived in the background of my life for a while, more a name people mispronounced rather than an identity. In February of '22, as I sat at my (~cushy) tech job in San Francisco, I watched in real time an attempt to erase that heritage from the planet. Adjacent to this post, you can read about what we've done since on the misc. page.
Something shifted as I watched a war unfold in the country half my family is from, and I was glued to socials and news. Then, in early March, a video of Zelensky appeared on a hacked Ukrainian news website and spread across Telegram, X, and Instagram. He was telling Ukrainian soldiers to lay down their arms and surrender. Shortly after, additional videos from Putin appeared claiming victory and martial law in the country. Both were crude by today's standards, but the cat was out of the bag — for the first time in the history of our planet, seeing is no longer believing.
This isn't a new concept, but it is a new delivery method. Sun Tzu's "all warfare is based on deception," the Trojan Horse, Operation Bodyguard in WW2 — the lineage is long. What has changed is the medium, and with each new modality (image, voice, video) the lie gets cheaper, faster, and harder to catch. We've seen this crossing from physical warfare into social warfare — the Tom Holland and Zendaya wedding, the list goes on.
Subject-matter
So in a world where seeing is no longer believing, and we actually can't believe our eyes, how does our society shift?
1. Context
At the macro level, the fake news epidemic is a clean example. The concept of fake news is not new, but what is new is the distribution engine and therefore the reach. A "lie" that would once have taken a week to spread through a town now reaches the global population at near zero latency — now, now, and now. Socials, algorithms, and the folding of the old gatekeeping news economy have produced a system in which the velocity of a story is almost entirely decoupled from its underlying source or truth. The medium is no longer the message. The medium is the amplifier, and it amplifies indiscriminately.
At the micro level, the responsibility for context has shifted onto the audience. Nobody signed up to be a part-time forensic analyst. But here we are. A piece of media is no longer self-sufficient. A clip without a source is not evidence; it should be viewed as a rumor with production value. The question we have to start asking is not "what does this show me" but "where did this come from, who is showing it to me, and what would they gain if I believed it."
A caveat: there is no specific "right" way forward. The majority of the information on news channels and social media is subjective. I'm not here to discuss the right and wrong of the world, but I am here to say it is our job as a society to understand the context behind the information we are taking in at an alarming rate.
2. Authenticity Verification
If the first layer of defense is the question we ask, the second is the tooling that answers it. The Coalition for Content Provenance and Authenticity — a standard backed by Adobe, Microsoft, the BBC, Sony, Nikon, and others — calls for a cryptographically signed chain of custody embedded into a piece of media at the moment of capture. Provenance becomes a property of the file itself, not of the platform serving it. This is also a bull case for blockchain. If you know me you know I'm not a huge Bitcoin fan, but I do think the underlying technology has significant opportunity for digital verification through the form of an immutable ledger. Tech firms are also building platform-level detection (Meta, YouTube, and TikTok now run ruleblocks against uploaded video at scale), watermarking standards from model providers like Google's SynthID, and the slow legislative push of the EU AI Act and a growing patchwork of US state laws.
None of these are silver bullets. Watermarks can be stripped. Provenance signatures can be omitted at the source. Detection classifiers are in an open, accelerating arms race against the generative models that produce the content they are trying to flag — and the generative models are improving faster. Every defensive layer buys time, not certainty.
But time, in this fight, is the asset. The goal is not to win. The goal is to tell the consumer that the burden of trust is now directly in their own hands, not in the sources from which they intake information.
Here the argument gets uncomfortable, and I think it is worth sitting with the discomfort rather than resolving it.
In a world where intelligence is cheap and verification is broken, context becomes the scarce input. Context itself is information. And the economy we have built — especially in a capitalist United States — is the "attention economy," the "platform economy," the "generative economy," and it is structurally incapable of pricing information correctly. From my time at Google, I saw first and foremost that we price engagement. We price reach. We price the things that travel, and context, almost by definition, does not travel. We're all shifting to short-form at an alarming rate, and within those short-form TikToks, how can we possibly begin to add context? Context is inherently heavy — a footnote, a source, a backstory, a person willing to slow down and step away from the attention economy. The thing we need most is the thing the system is becoming worse at delivering.
So we arrive at a paradox that I do not think the next decade is going to resolve cleanly: context is simultaneously the most valuable input in a synthetic-media world and the input the surrounding economy is structurally bad at rewarding. The people doing the slow work — verifiers, fact-checkers, archivists, regional reporters, the human infrastructure of "actually checking" — are the people the economy has spent the last twenty years quietly defunding and pushing aside in favor of quick hits.
I do not know how that contradiction breaks. I suspect it breaks in one of three directions: a new class of trust intermediaries emerges and gets paid like the auditors of the analog age; the platforms are forced, by regulation or by liability, to internalize the cost of verification themselves; or we collectively accept a lower epistemic standard and learn to live in a world where most of what we see is provisional.
Footnote: as with all my writing, I can't claim complete originality. We are all taking in influences and biases across our lives in every form, but I will claim authentic throughlines based on my lived experience. At the deepest level, I think the only original thought we have is to cry just after we are born — after that, everything is influenced by our surroundings. Use great tools like Pangram to understand the information you are reading.