From the panic of being "replaced by AI" to the practical productivity of film and television: Ben Affleck discusses the incentive distortions and moral divergences in the film industry.

What is truly worth studying in this material is not as simple as "a Hollywood star understands AI." What Ben Affleck is doing this time actually hits three core contradictions in the future film and television industry: Who exactly should AI replace? Who owns the training data? How should profits from movies be distributed? When you look at InterPositive, Netflix, and Artists Equity together, you will find that Affleck is trying to rewrite not just a post-production software, but the entire production function + ownership structure + incentive mechanism of the content industry. And let me correct a few important facts in your original text. Netflix officially acquired InterPositive in March 2026. Netflix has confirmed that Ben Affleck founded this film technology company, and the entire team joined Netflix, with Affleck becoming a Senior Advisor at Netflix. Subsequent SEC filings from Netflix revealed that the cash consideration for the acquisition completed in March was approximately $587 million; combined with Netflix's announcement and subsequent reports, this transaction corresponds to InterPositive. About Netflix Additionally, "Animals" did not launch on December 9, but rather on October 9, 2026, on Netflix and in select theaters. This has been confirmed on Netflix's official page. Netflix More importantly, Affleck has now confirmed: He used AI in "Animals" in "many places." However, he deliberately does not disclose specific shots because he believes that the essence of film still needs to maintain the "illusion of reality." Reuters Connect This statement is actually very important. ──────────────── 1. What makes InterPositive truly worth $587 million is not the "AI model" Many people's first reaction is: Ben Affleck trained an AI model, and Netflix spent $587 million to buy it. If understood this way, this transaction would seem very absurd. Because today there are many companies training video models. Open-source models are also becoming more and more common. What is truly valuable is actually the combination of three assets: Clean copyright datasets Film industry know-how Workflows embedded in real production processes When Netflix introduced InterPositive, it clearly stated that Affleck and a small team of engineers, researchers, and creators specifically shot their proprietary dataset in a controlled studio, rather than simply scraping internet film and television works. About Netflix This is the core. ──────────────── 2. The biggest moat for future AI may not be the model, but "legal, high-quality, industry-native data" This is a very important investment logic for AI. Today everyone is discussing: GPT, Claude, Gemini, video models. But foundational models are becoming increasingly commoditized. What may become truly valuable in the long term is: Proprietary Data Proprietary data. Especially: High quality + clear rights + highly specialized industry data. This is especially evident in the film industry. One million hours of video on YouTube does not equal: 100 hours of professional film data. Because professional film footage includes: Focal lengths Lighting Camera movements Depth of field Dynamic range Color science Exposure Actor blocking Continuity Production design VFX plates These elements are: Domain Knowledge Domain knowledge. ──────────────── 3. This is actually the same business logic as financial AI and medical AI In the future, general models may be accessible to everyone. What is truly valuable is: Bloomberg's decades of financial data. Hospitals' medical imaging. Insurance companies' claims data. Manufacturers' machine operation data. Hollywood's: Decades of high-quality production data. So AI business models will increasingly shift from: Who has the strongest model? To: Who has data that others cannot legally obtain? This question carries a lot of weight. ──────────────── 4. InterPositive is actually doing a kind of "AI data sovereignty" This may be one of the concepts in the entire story that entrepreneurs should learn from. Affleck publicly explained his workflow by saying: In the film production process, a discrete model belonging to the project can be trained based on the creator's own materials, and the learning outcomes still belong to the production side. LinkedIn This means: Previously: Materials → Post-production company. In the future, it may be: Materials ↓ Film-specific Model ↓ Continuous learning ↓ This model itself becomes an asset of the film. Thus, film assets will no longer only include: Scripts, Original materials, Final films, Copyrights. There will be one more thing: Production Model Production model. ──────────────── 5. This change is very significant: every movie in the future may have its own "small model" Suppose you are filming "Batman." During the training phase, you have shot: Actors' faces Costumes Wayne Manor Batmobile Lighting Gotham scenes. Thus, during the production process, a: Batman Production Model gradually forms, and later you need to supplement: A background. A reflection. A long shot. A weather condition. A missing shot. You may not need to rebuild the set or gather 100 people again. The model already understands: The visual world of this movie. ──────────────── 6. This is much more realistic than "Text-to-Video one-click movie generation" The AI community often makes a mistake: Treating the ultimate future as a near-term business model. Everyone imagines: Input: Generate a two-hour Christopher Nolan movie for me. Then the AI generates it directly. Very appealing. But what actually happens in the industry is usually not: 100% replacement. But rather: 5% → 10% → 20% process automation. Almost all major technologies in history have followed this pattern. ──────────────── 7. Excel did not first replace accountants, but rather replaced calculators Photoshop did not first eliminate photographers. It first changed: Photo editing. CAD did not first eliminate architects. It first changed: Drafting. AI entering the film industry is the same. It first targets: The most expensive, The most repetitive, The most mechanical, The most suitable for computer processing stages. For example: Background restoration Object removal Shot extension Continuity correction Lighting correction Reshoot replacement Previsualization Rotoscoping VFX cleanup. ──────────────── 8. So the AI that is truly worth investing in is not about "replacing directors," but about "eliminating friction costs" This is a very important entrepreneurial principle. Tech entrepreneurs are most likely to say: I want to disrupt the entire industry. The truly high-quality entrepreneurs ask: What are the most expensive, slowest, most annoying, and most repetitive steps in this industry? First solve that. A very typical problem in the film industry is: A shot may need to be reshot due to: Background issues A prop being incorrect Lighting continuity issues Actor sightline problems Reshooting may mean: Finding actors again Renting a studio again Gathering the crew again Re-lighting. The cost may be hundreds of thousands or even millions of dollars. If AI can solve 50% of this: The commercial value is already enormous. ──────────────── 9. This is another layer of logic behind the $587 million valuation: Netflix is buying a "content cost compressor" What is one of Netflix's biggest costs? It is not the app. It is not the servers. But rather: Content Content. Netflix itself clearly stated in its Q1 2026 documents that content is one of its largest expenditure areas and views InterPositive as a tool to generate higher returns on this investment. SEC So from Netflix's perspective: Even if InterPositive ultimately only reduces production costs by: 1% 2% 3%, given Netflix's massive annual content investment, it could still be worth hundreds of millions. ──────────────── 10. This is the correct valuation method for large companies acquiring AI companies Do not ask: How much revenue does this startup have this year? Ask instead: How much can it change the parent company's cost curve? For example: A large insurance company processes: $10 billion in claims each year. If an AI system reduces: 1% of fraud losses, that is: $100 million. So it could be worth: Several billion dollars. Similarly: Netflix invests heavily in content budgets each year. If AI systematically reduces: Reshoots, VFX, Post-production, Production cycles, It becomes a: Margin Expansion Engine Profit margin expansion machine. ──────────────── 11. Netflix's acquisition of InterPositive has a deeper significance: it is creating "AWS for the film industry" Netflix was previously just: A Distribution Platform. Then it became: A Studio. Now it is continuing to move into: Production Infrastructure. If Netflix in the future owns: AI post-production Virtual production Asset management Localization Recommendation Advertising Distribution Then Netflix is not just: "The people who make movies." But could become: The operating system for film production. ──────────────── 12. This is very similar to what makes Amazon truly powerful Amazon initially: Sold books. Later discovered that the infrastructure built internally for operating e-commerce could be sold externally. Thus, AWS was born. Netflix may undergo a similar evolution. Internally, Netflix is building for: Content production. Translation Dubbing Visual Effects Recommendation Editing Marketing The technology that has been trained, may ultimately become the infrastructure of the film industry. InterPositive is likely a piece of the puzzle in this direction. ──────────────── Thirteen, but what makes Affleck truly smart is that he did not start from "AI technology," but from "the pain points of filmmakers." This is a particularly valuable lesson for tech entrepreneurship. The most dangerous way to start a business: I have AI, let's see where it can be used. A better way: I have been in this industry for 30 years, I know where the most painful spots are, then see if AI can solve them. Affleck clearly belongs to the second category. This is called: Founder-Market Fit The fit between the founder and the market. ──────────────── Fourteen, why might a film director be more suited to work on film AI than a top AI researcher? Not because he knows more about Transformers. But because he knows: What problems are worth solving. Engineers might think: Generating a beautiful video, is already great. But what the director sees is: The character's position is wrong. The gaze is incorrect. The focal length is wrong. The blocking is unreasonable. The continuity is broken. The color temperature is inconsistent. The emotion is wrong. This is called: Tacit Knowledge Implicit knowledge. ──────────────── Fifteen, many of the greatest AI companies in the future will come from "people within the industry." This is very important. The next generation of AI startups may not be: AI PhDs looking for industries. But rather: Lawyers + AI engineers Doctors + AI engineers Bankers + AI engineers Directors + AI engineers Manufacturing engineers + AI engineers. Because: Models are becoming easier to obtain. But industry know-how is very hard to acquire. ──────────────── Sixteen, the most important aspect of Affleck's "Late-stage Fine-Tuning" is not the novelty of the technology, but the capital efficiency. His thinking is roughly: Do not retrain a: "What is a person, what is red, what is a car" world model. This foundational knowledge: Open-source models have already spent huge computational resources to learn. So directly: Existing foundational model ↓ Unfreeze weights ↓ Continue training ↓ Incorporate film industry knowledge. Affleck himself also described this process as "unfreeze the weights" at the Bloomberg event. LinkedIn ──────────────── Seventeen, this is actually one of the most important rules for future AI entrepreneurship: do not waste the money that has already been spent on foundational models. In the future, many vertical AI companies will do this: Foundation Model ↓ Domain Fine-tune / Post-training ↓ Proprietary Data ↓ Workflow Integration ↓ Industry Application The real moat is not: "I also trained a large model." But rather: The last mile. ──────────────── Eighteen, the "last mile" of AI is usually the most profitable. The internet is the same way. TCP/IP is not the most profitable. The application layer emerged: Amazon Google Facebook. Smartphones are the same. ARM chips are important. iOS is important. But the huge value is generated in: Uber Instagram TikTok Airbnb. AI is the same. The foundational model may be extremely valuable, but the real trillion-dollar application value will appear in: Vertical Workflow. ──────────────── Nineteen, what Affleck truly opposes is not AI, but "Commodity AI." This distinction must be made clear. He has previously expressed strong skepticism about general AI generating films. But now he himself: Founded an AI company, Sold it to Netflix, And extensively used AI in "Animals." Reuters Connect This is not contradictory. Because what he opposes is: AI replacing judgment. And supports: AI amplifying judgment. These two philosophies are completely different. ──────────────── Twenty, this may be the most important dividing line for future "Creative AI." One model: Prompt: "Make a movie for me." Machine: Generates. Another model: Director: Already possesses visual judgment and narrative judgment. AI is responsible for: Reducing execution costs. The former is: AI as Creator The latter is: AI as Production Leverage Affleck is betting on the second. ──────────────── Twenty-one, this is very similar to the debate when the camera first appeared. After the invention of photography in the 19th century, painters worried: Painting would die. As a result: Painting did not disappear. Instead, it shifted from: Recording reality To: Impressionism Modern art Abstract art. Because mechanical devices took over: "Accurate replication of reality." Humans shifted to: Judgment Expression Style. AI may repeat this process as well. ──────────────── Twenty-two, therefore AI will not devalue Taste, but may actually increase its value. This connects completely with what we discussed earlier about KITH. AI makes: Writing Images Videos Code Cheaper and cheaper. Thus: Production Cost ↓ But at the same time: Content Supply ↑↑↑ As content increases, what becomes truly scarce is: Taste + Judgment Taste and judgment. The greatest value of filmmakers in the future may increasingly not be: "I can operate software." But rather: "I know what is worth doing." ──────────────── Twenty-three, what is most scarce about Christopher Nolan is not camera technology. Any major studio can buy: IMAX cameras. But cannot buy: Nolan's judgment. Similarly: Any brand can buy: Photoshop. But cannot buy: The taste of top designers. So after AI democratizes tools: Judgment Premium The premium for judgment May actually increase. ──────────────── Twenty-four, another truly impressive experiment by Affleck is actually Artists Equity. If we say: InterPositive solves: Production Efficiency. Artists Equity solves: Incentive Efficiency. Putting the two together: The efficiency revolution is complete. ──────────────── Twenty-five, one of the biggest problems in Hollywood is not that movies are too expensive, but that "everyone's economic interests are not aligned." The traditional production structure: Investors: Bear capital risk. Stars: Get a guaranteed fee. Directors: Get a guaranteed fee. Crew: Get paid. Studios: Bear the final risk. Thus, a very typical Principal-Agent Problem arises. In economics, this is called: The Principal-Agent Problem. ──────────────── Twenty-six, what is the Principal-Agent Problem? The boss hopes: The company makes money in the long term. Managers may hope: To maximize this year's bonuses. Investors hope: The movie makes money. Stars may care more about: Getting a $20 million paycheck. Directors may care more about: Artistic ambition. Crew may care more about: Working hours. Everyone is rational. But: The entire system may be irrational. ──────────────── Twenty-seven, "Gigli" is Affleck's own "failed MBA case." He later publicly talked about: How he received a huge upfront fee for "Gigli," But the movie lost money. He expressed that this structure made him uncomfortable; on the other hand, many staff members participated in successful films but often did not see the so-called "backend profits." Yahoo The reported fee for Affleck in "Gigli" was about: $12.5 million. IMDb The success or failure of the film: His downside was limited. This is: Asymmetric Payoff Asymmetrical payoff structure. ──────────────── Twenty-eight, Wall Street has long known how to solve this problem: Skin in the Game. Fund managers invest their own money. Entrepreneurs hold shares. CEOs receive stock. VCs use carry. Why? Because one of the most important management principles is: Let decision-makers bear the consequences of their decisions. Artists Equity is actually just bringing this set: Into Hollywood. ──────────────── Twenty-nine, the breakthrough of "The Rip" is very specific. "The Rip" is a film produced by Artists Equity in collaboration with Netflix, set to launch on January 16, 2026. Home | Artists Equity They negotiated a performance-based bonus with Netflix. The key is: Not just for stars. But also including: grips electricians camera crew and other below-the-line crew. Matt Damon publicly stated: If certain performance targets are met, some staff bonuses could even double their income. GQ TheWrap reported that this arrangement covers about: 1,200 crew members. TheWrap This is very important. ──────────────── Thirty, why giving grips and electricians bonuses may really improve movie quality? Because it changes behavior. Traditional model: "Today I finish shooting, I get paid." New model: "If the movie performs better, I have additional earnings." Then many people may: Be more proactive in solving problems. Care more about the budget. Be more willing to make suggestions. Waste less. This is: Ownership Mentality The mentality of ownership. 31. This is actually Netflix's version of "employee equity" Silicon Valley has long known: Pure salary cannot generate strong ownership. So: Salary + Equity. The backend of the film industry in the past mainly focused on: Top actors Directors Producers. Artists Equity attempts to: Expand Profit Participation. If successful: Hollywood will get closer to: Startup Economics. ──────────────── 32. This is the true meaning of the name Artists Equity Equity is not as simple as "fairness." It also includes: Fairness Ownership Participation in profits. In other words: The real question of Artists Equity is not: "Give the staff a little more money." But rather: Who should own the economic value created by the work? This has become a capital structure issue. ──────────────── 33. However, your original draft's expression of "complete equity for all creators" needs to be cautious You wrote: Creators only take the union minimum wage, and the rest is all equal shares. This model has indeed been discussed by Affleck, and the overall direction of Artists Equity is to expand profit participation. The company's official definition is also to expand profit participation through creator partnerships. Home | Artists Equity However, it is not appropriate to write: All Artists Equity projects are fixed this way. They have clearly stated: Different projects have different scales, participants, and costs, The transaction structure will also vary. Google Groups So it is more rigorous to say: Artists Equity attempts to reduce some fixed upfront compensation based on project circumstances and bind creators to the project's final performance with higher performance bonuses or backend participation. ──────────────── 34. "Saving 30% of the budget" should not be taken as an industry fact If this is an estimate from Affleck's interview, It should be marked as: Affleck's estimate/judgment. Because film costs consist of many factors: Star salaries Locations Unions Insurance VFX Financing Tax credits Marketing Overtime Equipment. Incentive mechanisms can reduce waste, But there is no universal economic law: "Profit Sharing = Saving 30%." ──────────────── 35. However, there is indeed a very large efficiency dividend: Hollywood's "Agency Cost" There is a concept in finance: Agency Cost. When: The person providing the money And The person spending the money Are not the same person, Waste usually increases. This is especially serious in large films. A waste of $1 million, For: An individual on salary, May not matter. For: Their own company, It is completely different. ──────────────── 36. Therefore, Artists Equity essentially reduces Agency Cost Traditionally: Studio Capital ↓ Producer ↓ Director ↓ Talent ↓ Crew. Interests are layered all the way down. Artists Equity attempts to bind: The economic interests of more participants And: Final success. Thus: Agency Cost decreases. In theory: Capital efficiency increases. ──────────────── 37. This is actually very close to the core logic of Private Equity improving companies Good PE is not just about: Layoffs. But rather: Redesigning: Management Incentives. For example: CEO equity. Management has earn-outs. Achieving EBITDA targets: Increased rewards. Why? Because incentive changes: Behavior changes. What Artists Equity is doing in Hollywood is similar. ──────────────── 38. Why would Netflix be willing to accept this model? Netflix used to prefer: Buyouts. One-time payment. In the future: How successful the work is, Costs are relatively certain. This is comfortable for financial management. But the problem is: If everyone has already received their money: Future success has nothing to do with them. Artists Equity proposes: Less Guaranteed Economics, More Performance Economics. Netflix may thus: Reduce front-end risk, While making creators more invested. This is: Risk Sharing. ──────────────── 39. If this model matures, Hollywood may see a kind of "film startup culture" In the future, making movies may be like starting a business. A team: Director Actors Writers Cinematographers Producers Is not simply: Taking a job. But rather: Co-creating an asset. During filming: Income is relatively limited. After success: Share the upside. This will make: Film projects increasingly resemble: Project-based Startups. ──────────────── 40. And AI will make this model more feasible Why? Because AI reduces: Fixed production costs. Assuming previously: Independent film costs: $30M. AI post-production, Virtual production, Previs Brings it down to: $20M. At the same time, creators are willing to convert: Part of the fixed salary Into: Backend. Then capital demand may further drop to: $15M. At this point: The IRR structure of independent films completely changes. ──────────────── 41. Therefore, InterPositive + Artists Equity is actually a complete capital model On one side: Lowering Costs. AI reduces production costs. On the other side: Reconstructing Risk. Profit sharing reduces upfront cash pressure. Ultimately: Improving ROIC. This is the smartest layer of what Affleck is doing. ──────────────── 42. You can even write it as a simple formula Traditional film: Return = Box Office / Streaming Revenue − Huge Fixed Cost Affleck wants to change it to: Return = Better Content Revenue − Lower Production Cost − More Variable Compensation Fixed costs decrease. Variable costs increase. What does this mean? If a film fails: Capital loss is smaller. If a film succeeds: Creators earn more. This is a very nice: Variable Cost Structure. ──────────────── 43. This is very similar to the differences between airlines and software companies The higher the fixed costs: The greater the risk. Movies are originally: High fixed costs Extremely uncertain income. This is a very dangerous combination. Artists Equity hopes to turn: Some fixed costs Into: Performance costs. From a financial engineering perspective: This can improve risk resistance. ──────────────── 44. The biggest impact of AI on film may not be reducing actors, but rather reducing the "capital threshold" In the past, making: An industrial-grade film Might require: $100M. In the future, visual expressions of the same level: Maybe: $40M. In the future: $20M. The result is not: Films disappearing. But rather: More people can make films. This is called: Democratization of Production. ──────────────── 45. Historically, when technology lowers costs, content usually does not decrease, but rather explodes The printing press: Did not reduce authors. Books increased. Digital cameras: Did not eliminate photography. Photos exploded. YouTube: Did not reduce television content. Videos exploded. AI: Is likely not to reduce films. But rather: The supply of film and television will explode. ──────────────── 46. The real crisis is actually "Attention Deflation" Content is increasing. People's time: Is still 24 hours. So: Content Supply ↑↑↑ Attention Supply → Thus: The average attention each unit of content can gain decreases. This is one of the biggest economic problems in the future film and television industry. It is not: Whether content can be produced. But rather: Who can make people willing to spend two hours? ──────────────── 47. Therefore, the biggest moat for films in the future is not Production, but Demand This is especially worth remembering for investors. AI solves: Supply. But AI cannot automatically solve: Demand. You can generate: 100,000 films in a day. But consumers only watch: One. So the most scarce assets in the content industry in the future may become: Brands IP Stars Directors Communities Credible recommendations Distribution capabilities Attention. ──────────────── 48. This is also why Netflix still needs Ben Affleck If in the future AI can really: Automatically make films, Why would Netflix spend nearly $600 million to acquire a: Film director-founded AI company? It precisely illustrates: What Hollywood really needs is: Technology × Taste And not: Technology alone. ──────────────── 49. The most valuable thing Affleck said about AI is essentially: Human Judgment cannot be outsourced When Netflix officially quotes his thoughts, it emphasizes: What needs to be protected most in films is: Judgment. This kind of judgment requires: Decades of experience to form. About Netflix This is completely consistent with our previous analysis of Ronnie Fieg. In the future, AI will make: Execution cheaper. Therefore: Judgment will become more expensive. ──────────────── Fifty, this is a very important wealth rule for the next ten years. In the AI era: Ordinary execution capabilities will decrease in price. Top-level judgment will increase in price. So career growth should shift from: "What can I do?" to: "What can I judge?" Programmers: not just writing code. Need to judge: what products are worth making. Designers: not just making graphics. Need to judge: what is good design. Investors: AI can analyze a million financial reports. But: which company to buy, still requires judgment. Directors are the same. ──────────────── Fifty-one, what will be first compressed by AI is the "middle layer of execution labor." At the bottom: on-site complex labor is temporarily difficult to fully replace. At the top: Judgment Taste Decision-making IP Relationships are still scarce. Middle layer: Repetitive writing Repetitive post-production Basic design Simple analysis Standardized editing are the easiest to be compressed by AI. This is likely to be an important structure in the future labor market: Barbell Economy Dumbbell-shaped economy. ──────────────── Fifty-two, Affleck is worried about young people's "Learned Helplessness," which actually touches on a bigger risk in the AI era. The real danger of AI is not: machines becoming smarter. But possibly: humans stopping to train themselves. Assuming an 18-year-old student: lets AI write all essays. lets AI write all code. lets AI summarize all research. Then after 25: even though AI is strong, the person has not formed: Mental Model. ──────────────── Fifty-three, this is the biggest paradox of education in the AI era. The stronger AI becomes: the easier learning is. But: truly forming abilities may actually become harder. Why? Because: ability comes from: Struggle. Striving. Making mistakes. Modifying. Thinking. If AI takes away all the "difficult parts": the result may be: increased output, decreased ability. ──────────────── Fifty-four, so in the future, truly advanced individuals will actively maintain a "cognitive gym." Just like after the invention of the elevator: people no longer need to climb stairs. As a result: people spend money to run on treadmills at the gym. The same goes for the AI era. Technically, you don't need to: do mental calculations. write. program. research. But to maintain: Mental Fitness, you may still need to train actively. This may become one of the most important concepts in future education: Cognitive Exercise. ──────────────── Fifty-five, finally, stepping to the highest level of capital, connecting the whole matter of Ben Affleck. Many people see: an actor starting an AI startup. What I see is three assets being recombined. First: Intellectual Property film IP. Second: Artificial Intelligence production tools. Third: Equity Incentives capital incentives. Combining these three things: IP creates demand. AI lowers costs. Equity improves incentives. The ultimate result is: higher capital returns. This is the core of business. ──────────────── Fifty-six, if this model succeeds, future Hollywood companies may no longer look like traditional studios. The future may be: 10 top directors 50 producers 200 core employees a large amount of AI project-based freelancers profit-sharing. Instead of: a large permanent organization. This is: Hollywood as a Network rather than: Hollywood as a Factory. ──────────────── Fifty-seven, this is actually a microcosm of the future of the entire knowledge economy. Film is just one of the first cases. In the future: law firms investment banks advertising companies consulting firms architectural design game companies will gradually see: AI productivity small core teams project-based talents profit sharing. Traditional companies: Employee Economy. The future is increasingly likely to be: Ownership Economy. ──────────────── Fifty-eight, so there are six things that entrepreneurs should really learn from Affleck. First, do not reinvent the Foundation Model. Utilize existing infrastructure to do the last 10% that is most valuable in your industry. Second, data must be controlled by yourself. Models will be commoditized. High-quality proprietary data will not. Third, do not use AI to replace expert judgment. Use AI to amplify expert judgment. Fourth, look for the most expensive friction costs in the industry. Not to do flashy demos. But to solve real money problems. Fifth, turn fixed costs into variable costs. Let risks be shared by participants. Sixth, let those who create value share in the value. This is one of the strongest incentive tools. ──────────────── Fifty-nine, if I were an investor, I would not really focus on how beautiful the AI-generated images are from InterPositive. I would focus on five indicators: For each film: how much does it reduce reshoot costs? how much does it shorten post-production time? how many VFX labor hours does it reduce? how many shots actually make it into the final cut? can it be repeatedly applied to hundreds of Netflix contents? Because in the end, whether $587 million is expensive or not: is not about looking at the demo. But looking at: Annual Cost Savings + Incremental Content Value. ──────────────── Sixty, ultimately, the most valuable layer of understanding in this matter Many people think: The AI revolution means: machines replacing humans. Ben Affleck is testing another possibility: allowing the most experienced people to have greater leverage. If this direction holds, the most dangerous people in the future are not: "those who understand AI." Nor are they: "those who only understand the industry." The truly strongest individuals will be: Domain Expert × AI × Capital industry experts × AI tools × capital and ownership. Such individuals may turn: what used to require 1,000 people, into 100 people. Turning: what used to require $100 million projects, into $30 million. But at the same time: quality does not decline, even improves. This is where AI will truly change the capitalist mode of production. So if I were to compress the entire Ben Affleck interview into one memorable sentence, I would write: What AI will first destroy is not creativity, but the expensive, inefficient, repetitive friction surrounding creativity; and when these frictions are eliminated, what will truly be scarce is no longer "those who can produce content," but those who possess judgment, copyright assets, proprietary data, and economic ownership. This is far more important than discussing "when AI can automatically generate a movie."