I watched Peter Leyden’s Big Think video and it gave me a useful name for a feeling I have not been able to shake.
I distrust the claim that we can identify a single decisive year while living through it.
What interests me is how many systems are changing at the same time.
AI, media, software, economics, work, trust, energy, biology, creative tools, personal infrastructure. Changes in one affect the others. It is difficult to adjust to so many of them at once.
The exact year matters less than the compression of change.
The Useful Part of Leyden’s Frame
Leyden’s big historical pattern is that America periodically goes through periods when institutions fail, people contest their authority, and new institutions are built. In the video, he points to the founding era around 1787, the post-Civil-War era after 1865, and the post-World-War-II era after 1945.
I don’t know if the 80-year-cycle thing is “true” in any clean historical sense. A recurring interval does not establish a reliable historical rule. These frameworks can make contingent human choices sound predictable.
But as a mental model, it helps.
The useful part is not “this exact year is definitely the turning point.” The useful part is that existing institutions are struggling at the same time new tools are becoming usable.
That feels right to me.
You can see it everywhere. Media businesses are struggling to adapt. Search is changing. Trust in institutions is shot. Hollywood is trying to figure out whether AI is a tool, a legal exposure, a labor threat, or all three. Software teams are trying to decide what junior work even means when an agent can scaffold a feature, write tests, and explain the code back to you.
Meanwhile, individuals are rebuilding their own personal infrastructure because public information is difficult to evaluate and keep track of.
These changes affect people at the same time.
The Convergence Is the Story
Leyden talks about three big technology tipping points: AI, clean energy, and bioengineering. OpenAI introduced ChatGPT on November 30, 2022. The 2020 Nobel Prize in Chemistry went to Emmanuelle Charpentier and Jennifer A. Doudna for developing a method for genome editing. The International Energy Agency describes solar PV and lithium-ion batteries as learning-curve technologies, with costs falling as cumulative capacity doubles.
The effects extend into several areas of everyday work.
AI is not just a chatbot. It changes software, research, design, writing, customer support, search, scheduling, legal review, coding, editing, and the basic expectation of how fast a thought can become a draft.
Media used to be easier to talk about as “content,” which was always a dumb word but at least pointed at something. Now it is distribution, trust, identity, recommendation systems, fandom, propaganda, newsletters, podcasts, YouTube clips, TikToks, Discord servers, and whatever the hell X is today.
Work gets flattened into remote versus office because that is the argument people know how to have. The real shift is deeper: meetings, email, project management, hiring, training, documentation, review cycles, approvals. The routine processes that organize work are changing too.
Economics is the same. Inflation and interest rates matter, obviously, but the deeper question is whether people believe the deal still works. Housing, healthcare, education, subscriptions, software rent, creator income, job security, retirement, the cost of being a functioning adult. Many people find those costs difficult to manage.
Trust is not just misinformation. It is the loss of shared reality. Screenshots can be fake. Audio can be fake. Expertise is politicized. Institutions overplayed their credibility. Random people on the internet are sometimes right before the official channels are. That makes it harder to decide whom to trust.
Creative tools are crossing old boundaries. Unreal Engine is not just for games. It is previs, post-vis, virtual production, digital puppetry, and real-time filmmaking. AI image tools are not just making pretty mood boards. They are becoming part of pitch decks, look development, location planning, costume visualization, and concept exploration.
Then there is personal infrastructure, which has become more important to how I keep track of information. Notes, archives, backups, automations, password managers, local files, RSS feeds, read-it-later queues, personal websites, media libraries, AI assistants, private search across your own stuff. When the outside world gets noisier, your own system matters more.
People have to adopt or assess the tools while dealing with existing workload, financial, and institutional problems.
So the feeling is not “AI exists.” The feeling is “AI exists while everything else is already unstable.”
That is different.
Production Is a Good Example
Production makes this easier to see because production requires technology to meet practical constraints.
On a real show, nobody cares that your AI tool can make a beautiful impossible room. Can construction build it? Can the camera move through it? Can VFX use it? Does it match the budget? Does legal approve the training data? Does the union care? Does it save the art department time, or does it create a thousand new little cleanup tasks?
The answers determine whether the tool is useful on that job.
The practical stuff is much less glamorous and much more important: searching a video archive with natural language, roughing out a shot list, tone-checking an email, generating a first-pass lookbook, turning a location photo into a rough 3D reference, testing furniture in a virtual set, cleaning plates, doing roto prep, comparing schedule options, summarizing production notes.
These uses reduce the time needed for particular tasks.
The distance between idea, reference, test, revision, and decision gets shorter. That changes the work even when the crew size does not change. It changes expectations. It changes what “fast” means. It changes who can make a credible pitch. It changes how much polish people expect before a thing is even real.
Same with Unreal. The interesting part is not that a game engine can make pretty images. The interesting part is that film people, game people, editors, directors, VFX supervisors, and production designers can all start touching the same virtual object earlier in the process.
That changes how departments coordinate, as well as which software they use.
Software Is the Same Pattern
Software is going through its own version of this.
The first-order take is “AI writes code.” Fine. True sometimes. Overstated often. Also not the whole story.
The better take is that software work is being reorganized around faster loops.
You can ask an assistant to inspect a repo, find the relevant files, make a scoped change, write the tests, run the checks, explain the diff, and leave documentation. Not perfectly. Not magically. But well enough that the shape of the job changes.
The bottleneck moves.
It used to be “can I produce the code?” More and more, it becomes “do I understand the system well enough to ask for the right change, review the result, and know what failure would look like?”
That requires understanding the system and evaluating changes to it.
The same thing is happening in media. The bottleneck used to be access to publishing. Then it was attention. Now it is trust, taste, verification, and persistence. Anyone can publish. Anyone can generate. Anyone can clip. That makes actual judgment more valuable, not less.
Existing Institutions Still Matter
This is where Leyden’s historical frame is helpful without treating it as a prediction.
When old systems stop working, the people who benefited from them do not usually say, “Fair enough, let’s redesign this thoughtfully.” They defend the old system. Sometimes because they are cynical. Sometimes because their whole identity is built on it. Sometimes because the replacement really is undercooked and dangerous.
That part is not mysterious.
You can see it in media companies trying to defend distribution models that no longer make sense. You can see it in studios trying to use new tools without triggering labor blowback. You can see it in universities trying to respond to AI with plagiarism panic instead of redesigning assignments. You can see it in software orgs pretending nothing has changed while individual engineers quietly rebuild their workflows. You can see it in politics, where institutions built for a slower information environment are getting battered by rapid cycles of online outrage.
Existing institutions retain power, while the proposed replacements remain unproven.
What I’m Doing Differently Because of This
I am taking my own infrastructure more seriously.
I want to choose how I store and evaluate information rather than simply accept each service’s defaults.
I want my notes searchable. I want my files backed up. I want my important work in formats I can move. I want my website to be mine. I want my media library to survive platform churn. I want AI in the workflow, but I also want receipts, sources, diffs, tests, and human judgment around it.
I am paying more attention to production workflows than AI demos.
The demo is usually the least interesting part. The real question is where the tool lands in the pipeline. Who uses it? What does it replace? What does it make faster? What new mess does it create? What approval process does it need? What happens when it is wrong?
I am treating trust like a design problem.
Screenshots, citations, provenance, version history, reproducible checks, clear authorship, local archives. These records help me check claims and recover earlier work.
I am trying to stay flexible without becoming gullible.
Some of this is hype. Some of it is real. The trick is not to pick a team and defend it forever. The trick is to keep updating as I see how the tools perform in practice.
What I Take From It
Leyden’s proposed historical cycle may be too neat. I do not need to accept it to find his account of simultaneous change useful.
AI affects software; software changes how work is organized; changes in work and media affect whom people trust. Those relationships make it difficult to assess one technology in isolation.
For my own work, the response is practical: keep recoverable records, check sources and outputs, and judge tools by what they help a team complete. Those habits remain useful even if the historical prediction is wrong.