Runway Turns AI Bug Into Feature
Runway's AI video model bug turned into a feature, a lesson for companies building foundation models
Runway's AI Video Model Bug Turned Into a Feature
Runway spent weeks trying to engineer its way out of a stubborn bug: AI-generated avatars would drift off-center during real-time video generation, as reported by VentureBeat. The fix wasn’t a back-end patch — it was a new front-end feature that just worked around the problem. Ryan Phillips, head of enterprise product at Runway ML, walked through this lesson at VB Transform 2026, arguing that even companies not building foundation models themselves can learn from how Runway builds, evaluates, and ships them.
This approach highlights the importance of flexibility and adaptability in AI development. By turning a bug into a feature, Runway demonstrated that sometimes, the most innovative solutions come from unexpected places. As the AI landscape continues to evolve, companies must be willing to think outside the box and explore new approaches to problem-solving.
The implications of this approach are significant. If companies can learn to turn bugs into features, they may be able to reduce the time and resources spent on debugging and focus on more strategic initiatives. This could lead to faster development cycles and more innovative products, which could in turn drive business growth and competitiveness.
As AI adoption becomes more widespread, the ability to think creatively and turn potential liabilities into assets will become increasingly important. Companies that can master this skill will be better positioned to succeed in a rapidly changing technological landscape.
Visa Uses Mythos to Hunt for Bugs in Its Payment Network
Visa used Anthropic’s Claude Mythos to hunt for bugs in its own payment network, a vast infrastructure that spans more than 200 countries and territories, moves money in roughly 160 currencies, and connects nearly 5 billion payment credentials to over 175 million merchant locations, according to VentureBeat. The model stitched minor weaknesses deep in the stack into working exploit chains that would traditionally have surfaced only late in penetration testing.
Rajat Taneja, Visa’s president of technology, walked the VB Transform 2026 audience through what came next, including why Visa released the harness that governed the entire hunt. This move highlights the importance of collaboration and knowledge-sharing in the pursuit of cybersecurity.
The use of AI-powered tools like Mythos to hunt for bugs is a significant development in the field of cybersecurity. By leveraging the power of AI, companies like Visa can identify and address potential vulnerabilities more quickly and effectively, reducing the risk of cyber attacks and protecting sensitive data.
As the payment landscape continues to evolve, the use of AI-powered tools to enhance security will become increasingly important. Companies that can stay ahead of the curve and leverage the latest technologies to protect their networks and customers will be better positioned to succeed in a rapidly changing environment.
Instacart's CTO on AI and Tech Debt
Instacart is posing the provocative question: What if most of the work your engineers do today should, in fact, be done by machines? At VB Transform 2026, CTO Anirban Kundu argued that dev teams continue to waste their time on draining, repetitive, high-volume work; this should be absorbed by AI agents so that humans can focus on problems that require judgment, intent, and exception handling, as reported by VentureBeat.
In fact, in 97% of cases, Instacart’s builders don’t even read code anymore. This shift highlights the potential for AI to automate routine tasks and free up human resources for more strategic and creative work.
The implications of this approach are significant. If companies can leverage AI to automate routine tasks, they may be able to reduce the burden of tech debt and focus on more innovative and high-value initiatives. This could lead to faster development cycles, improved product quality, and increased competitiveness.
As AI adoption becomes more widespread, the ability to think creatively about how to leverage AI to automate routine tasks will become increasingly important. Companies that can master this skill will be better positioned to succeed in a rapidly changing technological landscape.
Pangram Raises $9M to Detect AI-Generated Content
Pangram has raised $9 million to scale its AI detection software, as reported by Startups | TechCrunch. The startup has also released a new AI text detection model, Pangram 4, and an AI image detection model in research preview.
This development highlights the growing need for tools that can detect and mitigate the impact of AI-generated content on the internet. As AI-generated content becomes more prevalent, companies and individuals will need to be able to identify and distinguish between human-generated and AI-generated content.
The implications of this development are significant. If companies can develop effective tools for detecting AI-generated content, they may be able to reduce the risk of disinformation and misinformation and promote a more transparent and trustworthy online environment.
As the AI landscape continues to evolve, the need for effective tools to detect and mitigate the impact of AI-generated content will become increasingly important. Companies that can develop and deploy these tools will be better positioned to succeed in a rapidly changing environment.
Bot-Detection Startup Spur Nabs $200M from Insight
Spur Intelligence has raised a $200 million round from Insight Partners for its tech that can identify legit human traffic from bots, as reported by Startups | TechCrunch.
This development highlights the growing need for tools that can detect and mitigate the impact of bot traffic on the internet. As bot traffic becomes more prevalent, companies and individuals will need to be able to identify and distinguish between human and bot traffic.
The implications of this development are significant. If companies can develop effective tools for detecting bot traffic, they may be able to reduce the risk of cyber attacks and fraud and promote a more secure and trustworthy online environment.
As the cybersecurity landscape continues to evolve, the need for effective tools to detect and mitigate the impact of bot traffic will become increasingly important. Companies that can develop and deploy these tools will be better positioned to succeed in a rapidly changing environment.
The Bottom Line
The stories highlighted above demonstrate the significant impact that AI is having on the tech industry. From turning bugs into features to detecting AI-generated content, companies are leveraging AI to drive innovation, improve efficiency, and promote security.
As the AI landscape continues to evolve, companies will need to stay ahead of the curve and leverage the latest technologies to succeed. This will require a combination of technical expertise, creative thinking, and strategic vision.
- Companies can turn AI bugs into features by thinking creatively and leveraging the power of AI.
- AI-powered tools can be used to detect and mitigate the impact of cyber attacks and fraud.
- Companies can leverage AI to automate routine tasks and free up human resources for more strategic and creative work.
- AI detection software can be used to detect and mitigate the impact of AI-generated content on the internet.
- Companies that can develop and deploy effective bot detection tools will be better positioned to succeed in a rapidly changing environment.
🚀 Built by Mapt
Like this site? Mapt builds websites, brands & growth engines — over text.
📄 Full episode transcript
Runway just turned a catastrophic bug in its AI video model into a feature, because after weeks of trying to engineer a fix, the company realized it was easier to just work around the problem. This might sound like a crazy approach, but according to Ryan Phillips, head of enterprise product at Runway ML, it's actually a key lesson in how to build and ship AI models effectively. Phillips explained at VB Transform 2026 that even companies that aren't building foundation models themselves can learn from Runway's approach, which emphasizes flexibility and creative problem-solving. The bug in question caused AI-generated avatars to drift off-center during real-time video generation, which could have been a major problem for the company's users. But instead of trying to fix the bug, Runway decided to turn it into a feature, effectively finding a way to make the problem work for them.
This approach might seem unorthodox, but it's actually a testament to the company's ability to think outside the box and find innovative solutions to complex problems. And it's not the only company that's been pushing the boundaries of what's possible with AI. Visa, for example, recently used Anthropic's Claude Mythos to hunt for bugs in its own payment network, and the results were impressive. The model was able to identify minor weaknesses deep in the stack and stitch them together into working exploit chains, which could have potentially been used by hackers to breach the system. But instead, Visa was able to use the model to identify and fix the vulnerabilities, making its payment network more secure as a result.
The fact that Visa was able to use AI to identify and fix vulnerabilities in its payment network is a big deal, and it's a testament to the power of AI to transform industries and revolutionize the way we do business. And it's not just payment networks that are being transformed by AI - companies like Instacart are also using AI to change the way they approach tech debt. According to Instacart's CTO, Anirban Kundu, the company has stopped worrying about tech debt because AI has made it possible to automate so many tasks that used to require human intervention. In fact, Kundu says that in 97% of cases, Instacart's builders don't even read code anymore, because AI agents are able to handle most of the work.
As AI continues to transform industries and revolutionize the way we do business, it's also raising new challenges and concerns. One of the biggest challenges is the flood of AI-generated content that's hitting the internet, which can be difficult to distinguish from human-generated content. That's why companies like Pangram are working to develop AI detection software that can identify AI-generated content and help us to separate the real from the fake. Pangram has just raised $9 million to scale its AI detection software, and the company has also released a new AI text detection model and an AI image detection model in research preview. This is a big deal, because it could help us to build a more trustworthy and transparent internet.
But Pangram isn't the only company that's working to detect and prevent AI-generated content - bot-detection startup Spur has just raised $200 million from Insight Partners to develop its tech that can identify legit human traffic from bots. This is a crucial area of research, because it could help us to build a more secure and trustworthy online environment. And as we continue to navigate the rapidly evolving landscape of AI and machine learning, it's clear that companies like Pangram and Spur are going to play a key role in shaping the future of the internet. So what's next for AI and its impact on our daily lives - tune in tomorrow to find out.