Latest / Soft Ratios Research / Abundance moves the tollbooth (on creative bottlenecks)

Transcript

Abundance moves the tollbooth (on creative bottlenecks) · Soft Ratios Research

• 0:09
  Figure: Soft Ratios Research.
• 0:12
  Ground: The voices are synthetic [pause] no act, just the thinking.
• 0:18
  Figure: Hoyd Breton, a designer at Soft Ratios Studio, teaching himself in the open [pause] which is half a conversation. [pause] A review is the other half.
• 0:28
  Ground: One subject at a time, twice a week [pause] Tuesday and Thursday. [pause] Today, creative bottlenecks.
• 0:36
  Ground: Picture six versions of working software on the screen before lunch. Yesterday, there was only a blank folder and a vague sentence about what the thing should do. [pause] That should feel like release. The expensive part has collapsed. More can be tried, and less has to be defended before it exists. [pause] But the relief has an unpleasant edge. If six plausible versions are cheap, the finished object isn't the whole job anymore. It may not even be the difficult part.
• 1:07
  Figure: The familiar claim is that cheaper production opens a field. More people gain the tools. More work gets made. Old specialists lose their monopoly on execution.
• 1:19
  Figure: All of that can be true while power gathers somewhere else.
• 1:24
  Ground: Then access is being described too loosely. Having the tool isn't the same as having a durable place in the work that follows.
• 1:32
  Figure: People keep bundling four gains together: using the tool, releasing something, having a fair chance of being found, and having enough bargaining power to earn from it. New technology reliably improves the first two. The others move unevenly.
• 1:47
  Ground: That distinction can become a way to dismiss the opening. A person who couldn't afford a team can now test an idea alone. The possibility is real before the market does anything with it.
• 2:02
  Figure: Cheap production does create work that otherwise wouldn't exist. It lets in outsiders, small experiments, unlikely successes. The correction is narrower: a lower entrance fee doesn't tell us who'll own the room. [pause] Three earlier shifts make the pattern visible. Copying became cheap, then distribution became cheap, and now competent first drafts are becoming cheap. Each time, the difficulty moved to the next scarce thing.
• 2:28
  Ground: And each move changed the work itself. Not merely who collected the money.
• 2:34
  Figure: Start with a printed book. Before mechanical reproduction, much of the labor returned with every copy. A scribe couldn't finish one manuscript and pull another hundred from it. [pause] Printing spread across Europe, and the arithmetic changed. Between 1450 and 1500, book prices fell by roughly two-thirds. By the end of that period, presses had produced about twenty-seven thousand recorded editions and an estimated seventeen million copies. [pause] The scarce copy was becoming an edition.
• 3:04
  Ground: A page still had weight. Paper had to be bought before anyone bought the book. Type had to be set, corrected, and held in place. A room could fill with finished sheets that had already consumed the money meant to save the business. [pause] So the printer escaped repetition and inherited prediction. How many copies. Which title. Which language. Which town might pay. The hand stopped copying every sentence and began gambling on demand.
• 3:38
  Figure: Paper alone could take about half the money needed for an early printing project. The press reduced the labor in each copy, but the first sale still sat behind type, wages, paper, and a whole run of books. Cheap copies rewarded whoever could finance the batch and survive what didn't sell. [pause] Distribution hardened into another scarcity. Books were heavy, easily damaged, and often shipped unbound. One late-sixteenth-century route of roughly three hundred and sixty kilometers could add a fifth to the sale price. [pause] The press made the pile. Merchants, fairs, and booksellers made the market.
• 4:15
  Ground: That seems less like the end of craft than a split inside it. The visible gesture shrank. Around it grew selection, finance, correction, storage, transport, and sales. More hands were involved, but fewer of them could decide what entered the press.
• 4:33
  Figure: England made that control explicit. In 1557, the Crown gave the Stationers’ Company authority over commercial printing. A “copy” entered in its records belonged to a printer, bookseller, or publisher-financier inside the guild, not ordinarily to the person who wrote it. [pause] The arrangement tied commercial protection to political control. The Crown gained a trade that could police forbidden material. The licensed trade gained protection from rivals.
• 5:03
  Figure: Reproduction had become easier. Permission to reproduce at scale had become property.
• 5:09
  Ground: Still, the reader gained something enormous. Distant words became obtainable. More titles survived. More people could become known through writing. Calling the publisher a toll collector shouldn't erase the expansion beneath the toll.
• 5:25
  Figure: The expansion stands. The market enlarged, professional authorship grew, and printed knowledge traveled farther. The freedom wasn't false. It was lopsided: many more people could read and write for print, while a smaller group controlled capital, rights, and routes to buyers.
• 5:43
  Ground: Then the first lesson isn't that cheap tools are a trap. It's that output and bargaining power are different things.
• 5:51
  Figure: Bargaining power gathers around what stays scarce. Once copying was abundant, deciding, funding, licensing, and distributing became more valuable.
• 6:00
  Ground: A person choosing what to learn might hear that and rush toward management. Stop learning the difficult technique. Learn to own the press.
• 6:09
  Figure: That would misread the shift. The technique didn't become worthless. It stopped being sufficient. Accurate composition, editing, and presswork still mattered because errors could spread across every copy. The valuable ability widened from execution into responsible coordination.
• 6:27
  Ground: There is a bodily difference between making one excellent thing and arranging for a thousand excellent things to remain excellent. Scale doesn't remove care. It gives care more places to fail.
• 6:43
  Figure: Then a different burden collapsed. Digital publication pushed the cost of copying and delivering one more file toward almost nothing. A song, book, image, or piece of software no longer needed a shelf in every town. [pause] Online inventories could contain millions of works where a physical record shop might sustain ten or fifteen thousand albums. Self-publication removed the need to persuade a traditional gatekeeper before release.
• 7:11
  Ground: The shelf disappeared. The length of a person’s day did not. [pause] That's where the new cost entered. Every added work asked for a small portion of the same finite attention. Nothing prevented publication, yet almost everything could remain unseen. [pause] Selection moved from the room behind the publisher’s door into daily life. Search results, recommendations, playlists, feeds, reviews, messages from friends. The burden of choosing spread across the whole audience. [pause] Whoever made that choosing feel easy acquired power without having to finance most of what was being chosen.
• 7:56
  Figure: Platforms replaced the hard gate with a continuous soft one. Uploading was open. Visibility depended on ranking, recommendation, moderation, advertising, payments, and terms that could change after release. [pause] Music makes the separation unusually plain. In the United Kingdom, the number of artists releasing music roughly doubled from two hundred thousand in 2014 to four hundred thousand in 2020. Listening prices fell by more than a fifth in real terms.
• 8:27
  Ground: So access improved on both sides. More people released music, and more music became affordable to hear. That isn't a disguised loss.
• 8:38
  Figure: Both gains stand. But more than sixty percent of streams went to the top four artists in every thousand. Fewer than one in a hundred artists were heard enough in a year to bring in twelve thousand pounds.
• 8:52
  Ground: Permission to enter became cheap. Repeated attention remained dear.
• 8:57
  Figure: A playlist position could change a career because the platform now directed the scarce attention. Placement on one major global playlist was estimated to add hundreds of thousands of streams each day. The upload button widened participation. The ranking decided who got discovered.
• 9:15
  Ground: This is the part that feels cruel. Promotion used to stand beside the work, slightly embarrassing and separable. Under abundance, it moves inside the workday. Titles, thumbnails, release cadence, short extracts, audience replies, platform experiments. The made thing arrives carrying its own small sales department.
• 9:37
  Figure: Authors spent an average of seven and a half hours each week marketing their books. Only a small share saw a substantial rise in book income. Lower delivery costs didn't remove labor. They reassigned some of it to the constant pursuit of visibility.
• 9:52
  Ground: Then “build an audience” isn't a resilient answer by itself. It can mean renting a second job from the same platform that made distribution cheap.
• 10:02
  Figure: A follower count on a service whose recommendations can change isn't much to own. A direct, consenting relationship is different because it remains useful when ranking rules shift. The distinction isn't between promotion and purity. It's between tenancy and having somewhere else to go.
• 10:20
  Ground: That other route can still be tiring. A mailing list must be served. Subscribers expect continuity. Directness removes an intermediary and creates obligations that were hidden inside it.
• 10:33
  Figure: The cost doesn't vanish there either. But the relationship can strengthen someone's bargaining position. [pause] Radiohead’s direct release of In Rainbows earned more online before retail than its preceding album had earned across all formats. The advantage came from reputation and an audience already willing to arrive without a retailer directing them.
• 10:52
  Ground: That example contains its own warning. A direct route works best after people already care.
• 11:00
  Figure: Reputation compounds. It can't simply be switched on when the intermediary becomes inconvenient. A record of useful work, repeat customers, and remembered promises has to exist before it can strengthen anyone's position.
• 11:14
  Ground: The second lesson is harsher than the first. When distribution becomes abundant, publication stops being the finish line. The work extends into being found, being remembered, and earning the next voluntary encounter.
• 11:30
  Figure: The platform arranging those encounters can charge both sides: access to the audience, knowledge about the audience, promotion inside the distribution channel, and a share of payment.
• 11:42
  Ground: Now the question shifts again. If software can produce the first credible version, what becomes scarce immediately after that version exists?
• 11:52
  Figure: Generated drafts make well-defined execution cheap. In one writing exercise with working professionals, assistance cut completion time by forty percent while the finished work was judged eighteen percent better. [pause] The gain is tangible. A blank page becomes editable material sooner. A small team can compare more directions. Someone blocked by technique can produce a proposal that's good enough to test.
• 12:17
  Ground: That sounds broader than printing or digital delivery. Those shifts cheapened copies and routes. This one reaches into the act that used to prove competence.
• 12:29
  Figure: It reaches some acts unevenly. With product ideas, market groupings, and persuasive writing, consultants using generated assistance worked faster and produced stronger results. When the task exceeded the tool’s reach, its polished answers made them much less likely to find the correct one.
• 12:47
  Ground: The dangerous draft isn't the broken one. It's the smooth one that removes the feeling that anything still needs checking.
• 12:54
  Figure: Code shows the same boundary. In a set programming exercise, developers with assistance finished in less than half the time. But experienced maintainers working in mature repositories they knew well took longer with later tools. Prompting, waiting, reviewing, and correcting cost more time than the saved typing.
• 13:15
  Ground: So “software is cheap” names only the visible birth of the artifact. It says nothing about fitting that artifact into existing promises, protecting data, handling an odd case, or keeping the thing legible after its first author leaves.
• 13:31
  Figure: What changed was the price of variation. Once competent proposals arrive quickly, value moves upstream to choosing the problem and downstream to selecting, fitting, checking, maintaining, and taking responsibility.
• 13:47
  Ground: “Judgment” is a comfortable word for experienced people. It can turn a threatened hierarchy into a flattering story: the machine does the junior work, while the senior person’s taste becomes more precious. [pause] But taste wasn't issued with the senior title. It formed through awkward first attempts, corrections, cheap assignments, and watching someone notice what the beginner couldn't yet see. [pause] Remove those attempts, and the old expert may become more productive for a while. The path toward the next expert grows thin underneath. [pause] The missing bottleneck is time spent becoming capable of judgment before anyone is willing to pay for judgment alone.
• 14:33
  Figure: That changes the claim. Judgment isn't merely the scarce product after generation. Forming it is scarce too, and cheap execution can consume the ground where it used to grow. [pause] Junior writing once included summaries, outlines, routine copy, and first drafts. Junior design included variations, cropping, retouching, and production files. Assistant editors learned a body of footage while logging, sorting, and assembling it. Those tasks were modest deliverables, but they were also places to watch closely. [pause] Freelance work shows fewer short, replaceable writing and translation jobs, while the remaining work demands more complexity. A contract may become more valuable on average while the paid repetitions that once built a career disappear.
• 15:22
  Ground: A private gain becomes a collective shortage with a delay. Each firm has a reason to delete the elementary task. No single firm has a strong reason to preserve the profession’s future supply of experienced people.
• 15:37
  Figure: The useful distinction is assistance that reveals thinking and assistance that replaces it. In a large classroom experiment, unrestricted answers improved work done with help but left people worse when the help was removed. A guarded tutoring mode improved the practice without leaving the same loss behind.
• 15:56
  Ground: The resilient learning pattern is slower at the important moment. Attempt first. Expose the reasoning. Compare it with the generated alternative. Name the miss. Try again without support. [pause] That feels inefficient because the artifact appears later. But the artifact and the capability are two different outputs.
• 16:19
  Figure: A workshop that counts only completed drafts will choose replacement. One that also checks what people can do alone, how well they correct mistakes, and whether they're growing into independent responsibility can see whether the assistance is teaching or concealing.
• 16:36
  Ground: Then the decision about what to learn can't come from a list of tasks the tool performs today. Tasks are unstable. The deeper question is which kinds of understanding remain necessary when the first pass is free.
• 16:52
  Figure: Several survive all three collapses. Problem choice survives because abundant execution makes a bad choice more expensive. Knowledge of the field survives because plausible errors look ordinary from outside it. Integration survives because new work must enter old obligations. Verification survives because fluency isn't evidence.
• 17:16
  Ground: And consequence survives. Someone still has to release the work under a name, answer when it fails, repair the relationship, and decide whether the thing should remain in the world.
• 17:28
  Figure: Judgment becomes observable rather than ceremonial through a clear brief. Conditions for acceptance. Rejected alternatives. Source checks. Tests. Rights records. Edit history. Named approval. Evidence of what happened after release.
• 17:47
  Ground: A beautiful explanation of taste is weaker than a trail showing where taste prevented harm or produced a result. The value isn't that a person touched the work. It's that a person could tell the difference and accepted the cost of being wrong.
• 18:04
  Figure: Trust follows the same pattern. Generated polish makes surface quality easier to imitate, so identity, track record, and warranties become more valuable. In one online auction test, an established seller’s history raised what buyers would pay by just over eight percent for otherwise matched goods.
• 18:22
  Ground: Reputation is useful because it can be lost. A name without exposure to consequence is decoration.
• 18:29
  Figure: A signed history can support that exposure, but only narrowly. It can identify the credential that signed a claim, record which tools or ingredients were declared, and show whether the covered file changed. It can't by itself establish the person behind the signature, prove the claim is true, show the rights were cleared, or show the work deserves attention.
• 18:53
  Ground: A receipt isn't a conscience. Still, it's better than a cloud of confidence with no origin anyone can inspect.
• 19:01
  Figure: The same abundance may create a premium on presence. A recording can be copied without limit. A live performance still requires bodies sharing time. A generated image can resemble a handmade object. It can't duplicate the negotiated encounter of a tattoo, a site-bound installation, or a commission changed through dialogue.
• 19:21
  Ground: Those aren't refuges from technology. They're cases where removing the relation changes the product. The scarce thing isn't difficulty for its own sake. It's contact, consent, place, response, or bodily consequence.
• 19:38
  Figure: Infrastructure remains another tollbooth. A cheap prompt rests on expensive chips, data centers, electricity, training data, and hosted services. A low price at the screen doesn't mean the ownership beneath it is widely shared.
• 19:52
  Ground: That leaves a person building software in an odd position. Competing in a flood of cheap artifacts while paying rent to the providers that made the flood possible.
• 20:02
  Figure: The defensive move isn't complete independence. That's rarely available. It's noticing dependencies before mistaking them for assets: a hosted model isn't owned capability, a platform following isn't a portable audience, generated code isn't maintained knowledge, and a polished draft isn't verified truth.
• 20:21
  Ground: There is also a temptation to move entirely upstream — to become the chooser who never makes. That loses something. Judgment without repeated contact with material can become theatrical. The hand needs resistance to keep the eye honest.
• 20:40
  Figure: So execution should remain, but its purpose changes. Some of it produces the artifact. Some of it maintains the ability to inspect, alter, and refuse what the tool produces.
• 20:51
  Ground: Keeping a few skills deliberately expensive preserves that ability. An unaided sketch. A small piece of code understood line by line. A source checked at its origin. A conversation with the person who bears the consequence. Not nostalgia. A check against smoothness.
• 21:12
  Figure: Across all three collapses, abundance didn't abolish scarcity. It handed the advantage to whoever controlled what the cheap part still needed.
• 21:20
  Ground: None of those later scarcities is automatically owned by the person doing the work. That's the political part hidden inside the productivity story.
• 21:30
  Figure: Ownership can be altered. Rights can be retained or bargained for collectively. Audiences can be made more portable. Recommendation power can be examined. Training can be treated as paid work. Open tools and shared infrastructure can keep the ability to produce from settling behind one gate.
• 21:49
  Ground: But no policy turns abundance back into innocence. Every easier act creates another place where choosing, caring, or coordinating has to occur.
• 21:59
  Figure: So the practical question is larger than which tool to learn. It's which scarce part the work strengthens: knowledge no generic draft can fully contain, a relationship that can survive a platform change, a right that can be retained, a record that makes judgment credible, or a practice that still forms judgment in the first place.
• 22:21
  Ground: Cheap making isn't the disappearance of difficulty. It's difficulty changing address.
• 22:27
  Figure: The durable advantage belongs less to the person who can summon another version than to the person who can explain why this version should exist, prove that it holds, and remain present after release.
• 22:39
  Ground: When the first usable version costs almost nothing, which part of the work is worth choosing to remain responsible for?