ILM is embracing AI like any other tool, and you should too
The year, 1993. A rudimentary computer-generated T. rex—a reptilian skin stretched over a wire frame—played on a loop in a computer at Industrial Light & Magic in California. Three film legends—VFX supervisor Dennis Muren, animator Phil Tippett, and director Steven Spielberg—watched silently as the implications sank in. 'Cinema history changed,' Rob Bredow recounts in his April 2023 TED Talk, which has just been published on YouTube. Tippett, a stop-motion pioneer, dryly told Spielberg, 'I feel like I'm going extinct.' As most movie buffs know, that line landed in Jurassic Park. Tippett's fear, however, turned out to be unfounded.
The legendary effects company fused Tippett's stop-motion puppetry with nascent CGI, using a 'dinosaur input device'—a rigged armature with motion encoders—to digitize frame-by-frame animation. The result? A seismic shift that expanded artists' tool kits worldwide and opened a new era in filmmaking.
Bredow was only 19 when that happened. Now, as SVP of creative innovation for Lucasfilm and chief creative officer of ILM, he sees a direct parallel to today's artificial intelligence debates. 'Headlines say, 'AI is coming for our jobs,'' he says in the TED Talk. From the Dykstraflex—the computer-controlled motion camera that enabled Star Wars 's iconic dogfights—to the StageCraft—a 270-degree LED curved wall that projected hyperrealistic 3D environments for The Mandalorian and now many other shows and movies—Bredow argues that ILM's 50-year history is a demonstration of how technological leaps redefine, rather than replace, artistry.
Stop-motion transformed and merged with 3D effects. So did physical models, full-size sets, and matte paintings. Just like it's happening now with AI. 'Innovation thrives when old and new technologies are blended,' Bredow argues.
ILM was late to the AI game. This became painfully obvious when the effects company created a rejuvenated version of Mark Hamill for the Season 2 finale of The Mandalorian (despite fans cheering on Luke Skywalker's return to the screen). Done with traditional computer face tracking and 3D models— the same technique used to create Peter Cushing as Grand Moff Tarkin and Princess Leia— Return of the Jedi Luke was slammed for being unrealistic. Then a Star Wars fan and AI aficionado called Shamook re-created the scene using AI. The former took hours. The latter took weeks. There was no doubt about which one looked more realistic.
The difference was so obvious that the company realized it had to act: ILM hired Shamook days after the deepfake remake was released. He worked on Indiana Jones and the Dial of Destiny, where ILM merged generative AI, trained on Harrison Ford's past performances, with a meticulously handcrafted CG model to de-age the actor. The AI captured Ford's micro-expressions; artists fine-tuned subtleties like eye moisture and skin texture. Ford himself said it was pretty good and really felt like him. Because it did.
AI is just another tool in the toolbox
The ethos that now guides its AI integration has been in ILM's DNA since its origin. It was what drove George Lucas to pair engineers with artists to solve visual storytelling challenges. 'We're designed to be creative beings,' Bredow says. 'We love seeing tech and creativity work together.'
Bredow hinted at ILM's embrace of AI tools in a Fast Company interview back in August 2024: 'I do see a path forward with some combination of the algorithmic tools that we've had and some machine learning-based tools that we either already have or can imagine developing, that are really going to help accelerate artist workflows.' Now he has made clear that AI has reached a point in which it is just another toolbox in ILM's toolbox. His stance about the technology is one that I have been seeing more and more since independent filmmaker Paul Trillo, one of the pioneers in using generative AI for his shorts, told me the same years ago. Trillo thinks that AI will enable indie projects to achieve blockbuster-grade VFX: 'It is just a powerful tool in a creative's arsenal.'
It's just too bad that the example that Bredow presented in the TEDTalk was so underwhelming: A video that shows some uninspired sci-fi animals that looked like Photoshop-made images quickly turned to video using Kling, a commercial AI video generation tool developed by Chinese tech company Kuaishou. He described it as ILM's 'moving mood board,' but it falls short of what you would expect from the mother of all VFX houses.
But his points and the lesson from ILM's half-century of visual innovation stand. The Dykstraflex didn't kill cinematography—it birthed a new process and visual language. CGI dinosaurs didn't erase animators—they just demanded hybrid skills. Now, as AI reshapes VFX, Bredow says we are witnessing another T. rex moment: one where artists, armed with generative tools, push storytelling beyond current limits.
Adaptation is nonnegotiable. New tools should be embraced as long as they are not unethically taking advantage of other people's artwork. 'The next game changer,' he said, 'will light up screens worldwide.' The credits won't fade on human creativity. Hordes of people's names will keep rolling. At least for a few more years to come.
Hashtags

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


The Verge
38 minutes ago
- The Verge
Razer's new mousepad wirelessly charges mice with a Logitech-style puck
Razer has announced a new mousepad that can wirelessly charge your wireless mouse while you're using it. It's not Razer's first attempt at the idea, but the new HyperFlux V2 doesn't require you to use a single specific mouse — it works with several models in Razer's current lineup, charging their batteries continuously using an included adapter puck. It looks and sounds very similar to Logitech's patented Powerplay Wireless Charging System. The HyperFlux V2 Wireless Charging System comes in two versions for $119.99. One with a hard surface that's available now, and one with a softer cloth surface, which is expected to launch later this year. The original Razer HyperFlux was sold as part of a $249.99 bundle that included the Mamba HyperFlux mouse. The V2 model doesn't come with a mouse, but is compatible with the Razer Basilisk V3 Pro 35K, Basilisk V3 Pro, Cobra Pro, and Naga V2 Pro mice, which range in price from $129.99 to $179.99. The Mamba HyperFlux mouse that worked with the original version of the HyperFlux pad featured a supercapacitor inside that retained power for about 5 to 10 seconds when it was lifted off the pad. That was enough time to allow you to reposition the mouse without it dying, but to use it anywhere else required a USB cable. The new HyperFlux V2 instead uses a puck inserted into the bottom of the supported mouse models that charges their batteries wirelessly while they're on the pad, like its Logitech competitor. An LED indicator at the top of the pad indicates the battery's charge level by changing colors. Like the original version, the new HyperFlux V2 mouse pad also serves as a receiver, allowing Razer's mice and keyboards to wirelessly connect to a PC (through the pad, which connects to a computer over USB) using its proprietary HyperSpeed technology that's faster and more reliable than Bluetooth. The mouse pad is also compatible with Razer's wireless keyboards, including the BlackWidow V3 Pro and DeathStalker V2 Pro, but it can't charge them wirelessly.


Fox News
40 minutes ago
- Fox News
Expert speaks out on 'perfect solution' to federal hiring issues creating national security concerns
EXCLUSIVE: Fox News Digital sat down with SkillStorm CEO Justin Vianello, who addressed issues the federal government faces hiring workers, sometimes raising national security concerns, and explained what his company is doing to streamline that process. The federal government has struggled for decades with staffing issues in key roles like cybersecurity, tech and other high-skill areas, an issue flagged as far back as 2001, according to the Government Accountability Office. Vianello discussed how SkillStorm is attempting to solve those issues. "If we look at the procurement process and the way it's been structured, there's significant delays," Vianello told Fox News Digital. "So, it can take years to actually get to a point where a solicitation is actually awarded. And then, ironically or paradoxically, post that award, the agency will expect … the particular company to be able to deliver a team in 10 days. So, this process is inefficient and somewhat outdated." Vianello explained that the current hiring process is "lengthy" and "laborious," sometimes taking years rather than months and creating delays that teams need to properly mobilize and deploy. "One of the solutions to that issue is to actually allow for an on-ramp time where people can spend between two to four months to custom build teams that have the right skills, that have (the) right certifications that are based in the right locations to rapidly deploy teams and to accelerate IT transformation and automation. And that's really where the SkillStorm model comes in," Vianello said. Vianello says the company has spent millions of dollars in recent years building a Performance Acceleration Center for Excellence that is essentially a learning management training system with a customized curriculum and content along with a "stable of trainers" in a position to "rapidly upskill and deploy people." "How do we leverage that infrastructure to build out a solution for the federal government?" Vianello said. "Well, what we do is we leverage that infrastructure to accelerate and train teams. And the way the model works is we both bring people into our program. We train them for anywhere between 10 and 16 weeks. We pay them while we're training them. We help them achieve their certification, and then we deploy them. And we recover the investment that we make by billing them hourly." That system, Vianello explained, means SkillStorm takes "all the risk up front" and recovers it by billing hourly to the client. "Now this is the perfect solution to being able to custom-build tech teams, create net new talent for the ecosystem and being able deploy these people over time. But the government is gonna have to change the procurement system to not require people to be deployed within 10 days but allow companies to build these teams over two, three, four months." Another issue, Vianello told Fox News Digital, is the current hiring process can get tied up with security clearances and become a national security risk. "That's absolutely part of it, but I think there's a bigger issue here if you look more generally at our model and some of the issues that are facing the market," Vianello said. "Well, if you look at SkillStorm's model, SkillStorm has an innovative cost-effective solution to custom-build U.S.-based tech teams for rapid deployment. "Now, we have a student debt crisis in this country, and, at the same time, what are we doing? We're offshoring our children's roles to other countries, and we're using visa holders to take up the place of entry-level tech roles. Now, if we don't invest in programs like SkillStorm, if we do invest in these outcome-driven, apprenticeship-type programs, where's the next generation of cybersecurity experts going to come from? "Where's the new generation of AI innovators going to come from? This is a national security issue that is essential in driving innovation. Right now, there are 500,000 open cybersecurity roles as of January 2025. We are the domestic models, like these apprenticeship models, that can support that gap to make sure that we're protecting national security." Former General Services Administration (GSA) head Emily Murphy, who previously spoke to Fox News Digital about the GSA's work to streamline government in the era of DOGE, said she has "seen firsthand how outdated federal systems have become one of the most serious yet least discussed threats to national security. "Agencies charged with safeguarding cybersecurity and digital infrastructure are losing the talent battle to the private sector, and the slow, outdated process for onboarding cleared workers doesn't match the urgency of today's threats." Murphy explained that the federal government needs a "new pipeline" that "delivers clearance-eligible, project-ready professionals trained on mission-specific tools." "SkillStorm is doing exactly that, deploying "Stormers," technologists trained on specific tech platforms, at a significant discount. It's a smarter, faster way to secure the talent our government urgently needs. Vianello told Fox News Digital SkillStorm and the Department of Government Efficiency (DOGE) have similar goals in making government more efficient. "I think DOGE is really focused on IT automation and IT transformation and doing it on an efficient and cost-effective basis," Vianello said. "We believe, going forward, there's probably going to be more of a push to less full-time employees and more of a push towards efficient contractors coming in and accelerating project delivery. So, again, this really does come back in our belief. "To the solicitation process, how do we tighten it up? How do we make sure that once an award is made and that technology is implemented, it's not outdated? Because, if that continues to happen, how are you going to continue to attract technologists, young technologists who want to be part of the change?"


TechCrunch
40 minutes ago
- TechCrunch
Windsurf says Anthropic is limiting its direct access to Claude AI models
Windsurf, the popular vibe coding startup that's reportedly being acquired by OpenAI, said Anthropic significantly reduced its first-party access to the highly popular AI models Claude 3.7 Sonnet and Claude 3.5 Sonnet. Windsurf CEO Varun Mohan said in a post on X that Anthropic gave Windsurf little notice for this change, and the startup now has to find other third-party compute providers to run Claude AI models on its platform. 'We have been very clear to Anthropic that this is not our desire – we wanted to pay them for the full capacity,' said Mohan on X. 'We are disappointed by this decision and short notice.' In a blog post, Windsurf said this it has some capacity with third-party inference providers, but not enough, so this change may create short-term availability issues for Windsurf users trying to access Claude. With less than five days of notice, Anthropic decided to cut off nearly all of our first-party capacity to all Claude 3.x models. Given the short notice, we may see some short-term Claude 3.x model availability issues as we have very quickly ramped up capacity on other inference… — Varun Mohan (@_mohansolo) June 3, 2025 The decision comes just a few weeks after Anthropic seemed to pass over Windsurf during the launch of Claude 4, the company's new family of models, which offer industry leading performance on software engineering tasks. Anthropic gave several popular vibe coding apps — including Anysphere's Cursor, Cognition's Devin, and Microsoft's GitHub Copilot — immediate access to run Claude Sonnet 4 and Claude Opus 4. Those apps started supporting the new Claude 4 models on launch day. Windsurf said at the time it did not receive direct access from Anthropic to run Claude 4 on its platform, and still hasn't — forcing the company to rely on a workaround that's more expensive and complicated for developer to access Claude. Anthropic's AI models have become a favorite among developers, and in the past, Anthropic has worked with Windsurf to power its vibe coding tools. The AI-assisted coding sector, also know as vibe coding, has heated up in recent months. OpenAI reportedly closed on a deal to acquire Windsurf in April. At the same time, Anthropic has invested more in its own AI-coding applications. In February, Anthropic launched its own AI coding application, Claude Code, and in May, the startup held its first Code with Claude developer conference. Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you've built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | REGISTER NOW 'We're prioritizing capacity for sustainable partnerships that allow us to effectively serve the broader developer community,' said Anthropic spokesperson Steve Mnich in an email to TechCrunch on Tuesday, noting that it's still possible to access Claude on Windsurf via an API key. 'Developers can also access Claude through our direct API integration, our partner ecosystem, and other development tools.' Windsurf has grown quickly this year, reaching $100 million ARR in April, in an attempt to catch up with more popular AI coding tools such as Cursor and GitHub Copilot. However, Windsurf's limited access to Anthropic's models may be stunting its growth. Several Windsurf users that spoke with TechCrunch were frustrated by the lack of direct access to Anthropic's best AI coding models. Ronald Mannak, a startup founder that specializes in Apple's programming language, Swift, told TechCrunch that Claude 4 represented a significant jump in capabilities for his workloads. While Mannak has been a Windsurf customer since late 2024, he's switched to using Cursor in recent weeks so that he can vibe code more easily with Claude 4. As a short-term solution to support Claude 4, Windsurf allows users to connect their Anthropic API keys to their Windsurf accounts. However, developers have noted that this 'bring your own key' solution is more expensive and complicated than if Windsurf provided the models itself. When it comes to vibe coders, optionality is the name of the game. Every few months, OpenAI, Google, and Anthropic release new AI models that seem to outperform the industry on coding tasks. Because of that, it benefits vibe coding startups to support AI models from all the leading developers. Windsurf spokesperson Payal Patel tells TechCrunch via email that the company has always believed in providing optionality for users. In this case, it seems Anthropic has made that a bit more challenging.