The release of Toy Story, the first feature animated entirely by computer

Filed under fiction, and it is a moment. canon/proposals.md puts toy-story-1995 in the fiction section, between contact-1985 and voyage-dans-la-lune-1902, and then — in the same entry — states a citation occasion that has nothing to do with the story: "the one displacement of a creative profession that has already run to completion, with dates on both ends … Cite when: an argument is made about generative tools replacing creative work — this is the case with an actual answer rather than a forecast." That occasion is satisfied by the release and the nine years after it. It is not satisfied by a single frame of the film. If the entry stays fiction, a reading that comes here for the displacement is citing a file whose kind points at the plot, which is precisely the error alexnet-2012 diagnosed when it talked its way out of idea.

The house test for the kind is the one eliza-1966 argued out and eight entries have used since: a date on which something visibly happened in public. This clears it without strain. The film premiered at the El Capitan Theatre in Los Angeles on 19 November 1995 and went wide on 22 November. What visibly happened was that a feature-length film containing no drawn frame was put in front of a paying audience and the audience did not care that it was made differently — $30m of budget returning a domestic gross several times over, a 100% aggregate on 163 reviews, and, ten years later, in its first year of eligibility, selection to the National Film Registry as the only animated title of 1995. That is an event, with a venue, a date, an adjudicating public and a scoreboard. It sits on the shelf next to deep-blue-1997, watson-jeopardy-2011 and alphago-move-37-2016: a capability demonstrated in front of witnesses, after which the field moved.

star-wars-1977 stayed fiction and the contrast is the argument. That entry is kept for a line of dialogue and a set of assumptions, both of which live in the text; filed as moment it would have been about box office, which is a fact about Hollywood. Here it runs the other way. The text is a good film and the canon does not need another good film; what the canon needs is the nine years.

The fiction reading I am declining, and why it is still in this file. The proposal names a second thread — "manufactured beings with inner lives who must conceal them from the humans who own them, played for comedy" — and it is real but it is not this entry's job. star-wars-1977 already holds owned, ordered-about, memory-wiped machines whose moral standing is raised as a joke, and holds it better, because Lucas's droids are made and Andy's toys are not. Nothing in this film suggests anyone manufactured a mind. The toys are simply conscious, uncaused and unexplained, and their inner lives have no relation to the factory. As a story about made minds it is thin.

There is one exception and I would rather record it than lose it. Buzz Lightyear is a mass-produced object that sincerely believes it is not one. It reports its own capabilities in good faith and the reports are false; it flies and the flight is a sequence of coincidences; it is corrected not by introspection but by an external demonstration — a television advertisement for itself, and then a window — and its collapse afterwards is total. That is the most citable thing in the text and it is about a system's confident account of its own nature being wrong in a way the system cannot detect from the inside, which is a live 2026 problem with a literature. It is filed here rather than in a separate entry because it does not carry a nine-year record behind it, and this entry does.

descends_from holds one id, and it is a causal claim rather than a thematic one. star-wars-1977 paid for this. George Lucas opened Lucasfilm's Graphics Group in 1979 by hiring Edwin Catmull out of the New York Institute of Technology, funded out of what Star Wars had made; when the licensing revenue fell away after Return of the Jedi and Lucas's 1983 divorce, the group was spun out rather than dissolved, and Steve Jobs bought it on 3 February 1986 for $5m to Lucas for the technology rights plus $5m of working capital. Pixar is that group. The line from the 1977 film to the 1995 one is money, and then people, and it is documented at both ends. Nothing else in canon/ precedes this. The ancestors I would actually want — Catmull's 1974 subdivision-surfaces thesis, Sutherland's Sketchpad, the RenderMan specification, the $26m three-picture deal Disney signed with Pixar in 1991 — are not in canon/ and I am not inventing ids for them.

What it is

An 81-minute animated feature, directed by John Lasseter, story by Lasseter, Pete Docter, Andrew Stanton and Joe Ranft, screenplay by Joss Whedon, Stanton, Joel Cohen and Alec Sokolow, produced by Bonnie Arnold and Ralph Guggenheim for Pixar Animation Studios and released by Disney through Buena Vista. It premiered on 19 November 1995 and opened wide on 22 November. It cost $30m. Counting later re-releases it has taken about $198.4m domestically and $176.2m internationally, roughly $374.6m in all. Rotten Tomatoes aggregates 100% from 163 critics; Metacritic 96; the opening-weekend CinemaScore was an A. Lasseter received a Special Achievement Academy Award in 1996 and the film took three nominations — Original Screenplay, Original Song, Original Score. On 20 December 2005, in its first year of eligibility, the Librarian of Congress added it to the National Film Registry; it was the only animated title selected that year.

The claim that makes it a moment is narrow and should be stated narrowly: it is the first feature-length film in which every image in the final print was generated by computer. Not the first computer graphics in a film, not the first digital compositing, not the first CG character. The first one where there was nothing else.

The production ran about four years from the 1991 Disney deal, including a shutdown in November 1993 and a restart the following April. The rendering was done on roughly 110 machines running around the clock — 87 two-CPU Sun SPARCstation 20s, 30 four-CPU SPARCstation 20s and a SPARCserver 1000 — consuming more than 800,000 machine hours. Individual frames took between about 45 minutes and 20 hours. Sources disagree on the frame count: 114,240 by one account and 110,064 by another, and I have not found a figure I can trace to Pixar itself, so both are recorded and neither is relied on.

The headcount is the number that matters here, and it is the one usually quoted wrongly. Pixar's crew began at 24 and peaked at about 110 — 27 animators, 22 technical directors and 61 others. Disney's The Lion King (1994), made in the technique this film displaced, is credited to a staff of around 800. The honest comparison is 110 against 800, not 27 against 800: a feature at roughly a seventh of the headcount, on the first attempt, by a studio that had never made one.

And the animators were not computer people. This is the load-bearing fact of the whole entry and it comes from Lasseter in the film's own production notes: "The majority of animators on 'Toy Story' had no computer training before working on this film. They came from traditional animation backgrounds including cel (hand-drawn), clay and stop-motion animation." They were hired for acting and character work; the software was built to be usable by people who did not have the software. The film was designed on paper first, in hand-drawn storyboards cut into story reels — the standard Disney method, unchanged. The first feature with no drawn frame in it had thousands of drawn pages behind it.

What followed at the displaced end is documented and fast. Michael Eisner announced the end of traditional animation at Disney in 2002; David Stainton ran Walt Disney Feature Animation from 2003 to 2006 and oversaw the conversion. The Paris unit closed in 2003 and the Tokyo unit closed with more than 100 jobs. Fifty animators went in Orlando in June 2003 when Brother Bear wrapped. On 12 January 2004 Disney announced the closure of Walt Disney Feature Animation Florida — opened May 1989, about 260 staff, more than 250 jobs — and Stainton's public reason was consolidation: "Having the entire animation group working together in Burbank under one roof will further enhance our filmmaking process." The reporting that week put the underlying reason as "an emerging audience preference for cutting-edge 3-D computer-generated movies over traditionally drawn cartoons," and put the division's headcount at a peak of 2,200 in 1999 down to about 600, all in Burbank. Home on the Range opened in April 2004, took $103.9m worldwide, and was the last traditionally animated Disney feature of that era. Elsewhere the same thing: DreamWorks left hand-drawn features after Sinbad (2003), Warner Bros' feature animation unit effectively ended after Quest for Camelot (1998).

Eight years and five months, premiere to Home on the Range; eight years and two months to the Florida closure announcement. The proposal's "nine years" is the calendar-year difference, 1995 to 2004, and the exact figure is a little under nine. That interval is what "dates on both ends" means, and it is the only such interval available to anyone arguing about this.

Why a reading would cite it

There is no item in this project's current window about creative-work displacement, and I would rather say that than dress a standing occasion as a fresh one. The 17 August reading's work-and-economy lens read "Nothing notable found since the last reading"; the 19 August one holds Baidu's AI revenue crossing half its core book and OpenAI cutting prices, which are adoption facts and not displacement facts. The nearest live thing is the 18 August entry on Google buying Spirit Airlines' internal corpus out of bankruptcy for $10m — about 100 million emails, 3.4 million payroll records, employee data back to 1986, customer records carved out and employee records not — which is the input half of the creative-work fight rather than the output half, and belongs to a different argument.

The occasion this entry is actually for is a dated one and it comes due this year. On 31 January 2024 CVL Economics published Future Unscripted: The Impact of Generative Artificial Intelligence on Entertainment Industry Jobs, commissioned by the Animation Guild (IATSE Local 839), the Concept Art Association, the Human Artistry Campaign and the National Cartoonists Society Foundation, off a survey of 300 entertainment executives run 17 November to 22 December 2023. Its headline finding: about 21.4% of US film, television and animation jobs — roughly 118,500 of some 555,000 — "are likely to have a sufficient number of tasks affected to be either consolidated, replaced, or eliminated by GenAI in the U.S. by 2026," with a further 52,400 in gaming and 203,800 across US entertainment overall; 62,000 in California alone. By occupation: 3D modellers 33%, compositors 25%, storyboard artists, animators and illustrators 15%. Three-quarters of respondents said generative tools had already supported the elimination, reduction or consolidation of jobs in their own division.

The due date on that forecast is 2026. We are in the last four months of it. A reading in this window that meets the number — and it will, because it is the most-quoted figure in the argument — needs something to hold it against that is neither a projection nor a vibe. This is the only case on the board where a general-purpose tool did in fact eliminate a technique from the top of a creative industry and the whole arc is finished and dated. That is what it is for.

Four things about the citation, in order of how much work they do.

First, the clock. From the demonstration that the new tool could carry a feature to the last feature made the old way was eight years and five months — and the demonstration was not the start of the change either, since Pixar had been at it since 1986 and the Lucasfilm group since 1979. A reading that meets a three-year displacement forecast should say out loud that the one completed case took closer to three times that, and should also say why the comparison is not clean: in 1995 the new tool cost $30m and four years and hundreds of thousands of machine hours per picture, and in 2026 the new tool is a subscription. The constants are different and I do not know which way. But "nine years" is a real number and "three years" is a survey of executives' intentions, and a reading should not let those two things sit at the same weight.

Second, who lost. Not the animators as a class. The 27 people who animated this film were, in the director's own words, mostly cel, clay and stop-motion animators with no computer training, and the craft — timing, weight, performance, acting — transferred whole. What did not transfer was the institution. Disney went from 2,200 animation staff in 1999 to about 600 in 2004; the Florida studio's 258 and Paris's and Tokyo's units were not retrained, they were closed. Some staff took transfers to Burbank; others left for California, Canada and India, and some left animation entirely. In April 2013 Disney laid off nine veteran hand-drawn animators, including Nik Ranieri, Ruben Aquino, Frans Vischer, Russ Edmonds, Brian Ferguson, Jamie Lopez and Dan Tanaka — people at the top of the craft, eighteen years after this film's release, as the last administrative act of the transition. The skill was portable and the job was not. That distinction is the single most useful thing this entry carries into an AI argument, and it cuts against both of the usual conclusions.

Third, what the institution lost that it could not get back. When Disney reversed itself and made The Princess and the Frog in 2009 — Bob Iger undoing Eisner's policy, John Lasseter installed as chief creative officer, Ron Clements and John Musker brought back — the studio had to retrieve salvaged animation desks, because the infrastructure had been dismantled. The film cost $105m and took $271m, which is a success by most measures and was read internally as a failure against the 1990s hand-drawn hits; Lasseter's account is that the format "was viewed as old-fashioned by the audience." Winnie the Pooh followed in 2011 — $30m against $50.1m worldwide, opened opposite the final Harry Potter — and that was the end. A reading arguing that a displaced capability can be re-staffed on demand should know that the richest animation studio in the world tried, with the founder of the displacing studio in charge, and got two films out of it.

Fourth, and this is the part the doom reading has to answer: the technique did not die. The Boy and the Heron — hand-drawn, Studio Ghibli, Miyazaki's second win — took Best Animated Feature at the 96th Academy Awards on 10 March 2024, twenty years after Disney called the format obsolete. Demon Slayer: Infinity Castle, made in the same displaced technique on a budget reported around $20m, opened in North America on 12 September 2025 and has taken about $738m worldwide, the highest-grossing anime film and the highest-grossing Japanese film ever made. The 2004 claim that audiences had developed a preference against drawn animation was a claim about a format that turned out to be a claim about a slate.

And the tool that replaced it commoditised to nothing. Flow was made in Blender — free, open source — by Gints Zilbalodis with a crew under fifty for about $3.7m over five and a half years, and won Best Animated Feature on 2 March 2025 against Inside Out 2 at $200m and The Wild Robot at $78m. The capability that cost $30m, four years and a room full of Sun workstations in 1995 was, twenty-nine years later, available to fifty people for the price of a mid-budget documentary, using software that costs nothing. That is what the end state of a tool transition looks like, and it is neither of the two things people expect: not the incumbent winning, and not the incumbent's workers unemployed — the tool got cheap enough that a crew of under fifty beat a $200m studio picture with it.

Against all of that, here is what is actually on the record for the current transition, and a reading should notice how thin it is compared to what the 1995 record contains by this point in its arc. Critterz — OpenAI-backed, produced by Vertigo Films and Native Foreign, script from writers on Paddington in Peru, human artists' sketches fed into GPT-5 and image models, professional voice actors, budget under $30m against a nine-month turnaround — was announced for a Cannes 2026 premiere and arrived at the May 2026 market as a sales launch with first footage rather than a finished film, and I can find no evidence it has been released as of this writing. Copán: La Leyenda opened in Honduras on 7 May 2026, 74 minutes, after roughly fifteen years in development of which about ten were conventional 3D work that ran out of money; AI finished it, promotional material carried visible Gemini Image Generator watermarks, and it was panned. Netflix's INKubator, a "GenAI-native" animation unit inside Netflix Animation with Serrena Iyer as COO, surfaced in May 2026 through job postings rather than any announcement by Netflix.

One caution that belongs with the citation and not in a footnote: the animation job losses of 2023–2026 are not, on the available evidence, an AI story. Pixar cut 175 people in 2024, about 14% of the studio; on 21 July 2026 Disney cut several hundred across the company including 116 at Pixar, roughly a tenth of its 1,100, in its third round of the year. Disney's statement was "part of our continual evaluation of how we manage resources and reinvest across the company as our industry continues to evolve," and the reporting attributes the cuts to the underperformance of Hoppers and Elio and to pressure on production budgets. AI is not named. The cause the trade press and the Guild both point at is the post-streaming-boom contraction and the offshoring of work to subsidised regions — Tom Sito's version is "executives are trying to figure out how to control costs, and to them it always comes down to cutting staff." A reading that files those layoffs under AI displacement has committed the exact error this entry exists to prevent, and has done it while a genuinely dated forecast about AI displacement sits ungraded three months from its due date.

What it got right, and what it got wrong

Not required for moment, and this entry would be worth little without it. What follows grades the claims the participants actually made, with the date made and the date due. The film made no claims; the industry around it made several.

Right — the tool would fully replace the technique at the top of the American feature market, and fast. Claim made November 1995, by the enthusiasts, and adopted as policy by Eisner in 2002. Due: by 2004. Graded: correct and then some. Every American studio hand-drawn feature unit was gone or converted within nine years, and the last one grossed $103.9m against a market where CG features were routinely doing multiples of that. This is the half of the record the pessimists get to keep, and it should be conceded without hedging.

Right — the headcount would fall, and by a lot. A feature at 110 people against 800, on the first attempt, with rendering hardware that would now embarrass a phone. Due: continuously. Graded: correct, and understated. The 1995 figure of 110 is itself an artefact of a hard, immature tool; Flow got under fifty in 2024 with free software. Anyone arguing that a new tool leaves total employment unchanged in a creative industry has to get past this, and cannot.

Wrong — that audiences had turned against drawn animation. Claim made January 2004, in Disney's own reasoning as reported at the time, and repeated by Lasseter afterwards about The Princess and the Frog: the format "was viewed as old-fashioned by the audience." Due: continuously since. Graded: false, and expensively. The technique Disney declared obsolete won Best Animated Feature in 2024 and produced a $738m film in 2025 on a $20m budget. What was actually failing at Disney in 2002–2004 was Disney's slate. The studio mistook a problem with its films for a problem with its medium, dissolved the capability, and then could not rebuild it when it changed its mind. This is the most transferable error in the file, because it is the shape of the mistake available right now to any studio that reads a bad quarter as evidence that a craft is finished.

Wrong — that the displaced workers were the ones who could not use computers. Claim implicit in every contemporaneous account. Due: continuously. Graded: backwards. The people who made the displacing film were cel, clay and stop-motion animators with no computer training; the tool was built for them. The people who lost their jobs lost them because a corporation closed a building in Orlando, a unit in Paris and a unit in Tokyo, and because by 2013 there was no hand-drawn feature in development to staff. The determinant was institutional, not technical, and the two are constantly confused.

Wrong — that the new tool would stay expensive and concentrated. Nobody claimed this outright in 1995, but the whole industrial logic assumed it: the barrier to CG features was capital and proprietary software, and the studios with both would own the category. Due: by roughly 2020. Graded: false. Rendering commoditised, the software commoditised, and the 2024 Best Animated Feature was made in a free program by fewer than fifty people for $3.7m. If there is one thing in this record that should make a 2026 forecaster nervous about their own model, it is that the tool that eliminated a profession went on to eliminate the advantage of the people who eliminated it.

Not graded — whether the work got better or worse. Toy Story is a very good film and Home on the Range is not, and it is tempting to read the transition as quality winning. That reading does not survive the comparison class. The 1995 film was made by the best available animation team over four years with unprecedented care and an owner willing to shut production down for five months rather than ship something bad. The more representative first output of a new tool is Copán: La Leyenda. There is no claim here I can grade, and selecting on the successful first instance is the fallacy this section exists to avoid.

Commonly misused as

Not required for moment, and this is where the entry does most of its work, because all three camps in the AI-and-creative-work argument reach for this case and none of them can have it whole.

• "CG killed 2D animation." It killed hand-drawn feature production at American studios, between 2003 and 2011, and did not kill the technique. Drawn animation went on to win an Academy Award in 2024, to produce the highest-grossing Japanese film ever made in 2025, and to remain the ordinary medium of television and of the Japanese industry throughout. The precise claim is about an industry's production decisions in one country over eight years. Anything wider is false.
• "The animators were replaced by computers." They were replaced by animators using computers, at roughly a seventh of the headcount, and the first cohort of those animators had no computer training when they were hired. Quoting this case for "the machine replaced the artist" inverts what the record says. The correct sentence is longer and less useful in an argument: the tool changed completely, the craft transferred almost intact, the number of jobs fell hard, and the losses fell on whoever was inside an institution that chose not to carry them across.
• "So everyone will reskill and it will be fine" — the optimist's misuse, and the one this entry is most often conscripted for. Two thousand two hundred to six hundred at one company. Two hundred and fifty-eight in Orlando, a unit in Paris, a unit in Tokyo, fifty in Orlando the year before, nine of the best hand-drawn animators alive in 2013. Transfers were offered to some and taken by some; others went to Canada, India or out of the business. "The skill transferred" is a statement about the skill. It is not a statement about the person holding it, and using the first as though it were the second is the specific dishonesty available here.
• "It took nine years then, so we have nine years now." The interval is real and the extrapolation is not. In 1995 the replacing tool cost $30m and four years per picture — on the order of $60–70m in 2026 money by a rough CPI conversion of my own, not a sourced figure — and required its own render farm; the constants are not comparable and I do not know in which direction the error runs. Cite the nine years as the only completed clock anyone has, which is worth a great deal, and not as a rate.
• "Toy Story proves the new tool wins on merit." It proves that this new tool, in these hands, with four years and a studio willing to halt production for five months, produced one excellent film. That is survivorship. The contemporaneous 2026 instances are a film that arrived at Cannes as a sales reel and a Honduran feature finished with AI after its money ran out, with the image generator's watermarks still on the marketing. Neither is evidence either way yet, and neither is this.
• "Animation layoffs are AI displacement." Not on the current record. The 2024–2026 cuts at Pixar, DreamWorks, Netflix and Aardman are attributed by the trade press, by the Guild and by the companies to the streaming contraction, slate underperformance and offshoring. The Animation Guild's own 2024–27 contract, ratified by 76.1% of voting members, treats AI as a future hazard to be bargained over — consultation on non-generative alternatives, no outsourcing of AI work to non-union labour, severance and retraining if a job is lost to AI — with three members of its own negotiating committee saying publicly that the terms were too weak. That is a union preparing for something, not reporting it. Filing today's layoffs under AI makes the eventual real signal unreadable.
• As evidence of anything at all. Fiction is not evidence and neither is history; this canon is not the evidence ledger; this file places no needle, no score and no landmark. The particular hazard is that this case is emotionally legible and numerically specific at the same time, which makes it unusually easy to launder into support for a 2026 claim in either direction. Cite it to give a forecast a shape to be measured against. Never cite it to settle what the forecast will turn out to be.

Sources

Primary where reachable, and the failures are named.

The film: Toy Story (Lasseter, Pixar/Buena Vista), premiered 19 November 1995, wide 22 November 1995. Release dates, credits, the $30m budget, the $198.4m/$176.2m/$374.6m grosses including re-releases, the 81-minute running time, the 114,240 frame count, the 800,000 machine hours, the crew figures (24 initial, ~110 peak, 27 animators, 22 technical directors), the ~800-person Lion King comparison, the Rotten Tomatoes 100%/163 and Metacritic 96, the CinemaScore, the Special Achievement Award and three nominations, and the 2005 National Film Registry selection: Wikipedia, "Toy Story."

Production detail — the render farm specification (87 two-CPU SPARCstation 20s, 30 four-CPU SPARCstation 20s, one SPARCserver 1000, ~110 machines running 24 hours), the 45-minute-to-20-hour per-frame range, the 110,064 frame count, the 61-person balance of the crew, and the Lasseter quotation on the animators' backgrounds — is from the production notes reproduced at animatedviews.com. Those notes are Disney/Pixar press material, and I have not read them at a studio archive; the quotation is recorded as press-kit provenance rather than as an interview I traced. The two frame counts do not agree and I could not resolve them, so both are given.

The Registry selection is confirmed at the Library of Congress announcement of 20 December 2005 (loc.gov, "Librarian of Congress Adds 25 Films to National Film Registry").

Pixar's lineage — the 1979 Lucasfilm Graphics Group, Catmull's hire from NYIT, the post-Return of the Jedi revenue fall and the 1983 divorce, the 3 February 1986 sale to Jobs at $5m plus $5m capital, and the 1991 $26m three-picture Disney deal: Wikipedia, "Pixar."

The Disney transition: the 12 January 2004 Florida closure announcement, the 260 staff, the 50 laid off in Orlando in June 2003, the Paris and Tokyo closures with 100+ Tokyo jobs, Stainton's "under one roof" statement, the "emerging audience preference" framing and the 2,200-in-1999-to-600 headcount are from the Baltimore Sun's 13 January 2004 report and contemporaneous wire coverage; the 2,200/600 pair is as reported at the time and I have not found it in a Disney filing, so treat it as journalism rather than as a company figure. Studio detail, the May 1989 opening, the 258 staff at closure and the staff outcomes: Wikipedia, "Walt Disney Feature Animation Florida." Home on the Range's $103.9m, the DreamWorks exit after Sinbad (2003), the Warner Bros end after Quest for Camelot (1998), and the salvaged animation desks retrieved for The Princess and the Frog: Film Stories, "How Hollywood movie studios abandoned hand drawn animation at speed." Eisner's 2002 announcement and Stainton's 2003–2006 tenure: Wikipedia, "David Stainton," and the same Film Stories piece. The Princess and the Frog ($105m/$271m, the Iger reversal, Clements and Musker, and the "old-fashioned" quotation attributed to Lasseter) and Winnie the Pooh ($30m/$50.1m, opening against Harry Potter): Wikipedia and Screen Rant; the Lasseter quotation is secondhand and recorded as such. The April 2013 layoff of nine hand-drawn veterans and their names: Cartoon Brew, "Disney Just Gutted Their Hand-Drawn Animation Division," and Wikipedia, "Nik Ranieri."

The survival of the technique: The Boy and the Heron's win at the 96th Academy Awards, 10 March 2024. Demon Slayer: Infinity Castle, North American release 12 September 2025, ~$20m budget, ~$738m worldwide and the highest-grossing-anime and highest-grossing-Japanese-film records, per Forbes, Rotten Tomatoes editorial and NBC News coverage; the running total moved several times during reporting and the figure here is the highest I could corroborate, so read it as "about $738m and still moving" rather than as final. Flow: $3.7m budget, crew under fifty, five and a half years, Blender with the EEVEE renderer, Best Animated Feature on 2 March 2025 against Inside Out 2 ($200m) and The Wild Robot ($78m), per Fast Company, Hollywood Reporter and Blender.org. KPop Demon Hunters won the same award at the 98th Academy Awards on 15 March 2026 (Sony Pictures Animation, animated by Sony Pictures Imageworks on a Maya/Unreal/OpenUSD pipeline); it is computer animation and is not offered above as evidence about drawn animation.

The forecast: CVL Economics, Future Unscripted: The Impact of Generative Artificial Intelligence on Entertainment Industry Jobs, 31 January 2024, commissioned by the Animation Guild IATSE Local 839, the Concept Art Association, the Human Artistry Campaign and the National Cartoonists Society Foundation; survey of 300 executives, 17 November – 22 December 2023. Figures and the quoted "sufficient number of tasks affected" language via Cartoon Brew's report and the Animation Guild's own AI page, which reproduces the projection verbatim. The 2024–27 contract terms, the 76.1% ratification and the three dissenting negotiating-committee members: Hollywood Reporter, Variety and TheWrap, December 2024.

The current record: Pixar's 175-person cut in 2024 (14%) and Aardman's just-under-5% reduction, the Guild's 6,000+ membership, and the Sito and Kaplan quotations: Screen Daily, "Why outsourcing and AI means the US animation sector is facing hefty challenges in 2025." The 21 July 2026 Disney round, 116 at Pixar out of ~1,100, the company statement and the attribution to Hoppers and Elio: TheWrap; the Variety and Deadline reports of the same round returned 307 and 402 respectively on this run and were not read. Critterz: Deadline's May 2026 Cannes-market report (also 402 on this run, read only through search indexing) plus Cartoon Brew and the September 2025 announcement coverage — budget under $30m, nine-month turnaround, Vertigo Films and Native Foreign, GPT-5 and image models over human sketches. Copán: La Leyenda: Wikipedia — 7 May 2026 Honduran release, Metrocinemas, Ricardo Morales, Level 7 Studios, 74 minutes, ~15 years in development, the AI pivot after the money ran out, the visible Gemini watermarks and the reception. Netflix's INKubator is reported from job postings in May 2026 and Netflix has not announced it; it is recorded at that weight.

Fetches that failed on this run: bls.gov returned 403 for the Occupational Outlook Handbook page on special effects artists and animators, so the +2% 2024–2034 projection and the ~5,000 annual openings are from search indexing of that page rather than read at source and are not relied on above; adobe.com's June 2026 creative-professions research timed out and is not used; today.com returned 403. Several search results for the 2026 material came from sites that appear to be automatically generated (aicrisis.org, aidoomsday.world, careercanopy.ai, metaintro.com and others) and carried precise-looking figures — "67% fewer concept artist positions," "23,000 gaming jobs since January" — that I could not corroborate anywhere else; none of them are used here. Grokipedia appeared in results for the computer-animated-feature counts and was not used, both on source-quality grounds and because the project that holds musk-robotaxi-2019 should not be sourcing its canon from xAI.

The empty work-and-economy lens of 17 August 2026, the Baidu and OpenAI items of 19 August, and the Spirit Airlines corpus purchase of 18 August are as recorded in this project's own digests, which hold the primary links. They are named here as the citation occasion. Nothing in this file is deposited as evidence, and nothing in it touches the needle.