Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

Advancements in VLMs and VLAs make developing early prototypes extremely fast, enabling small teams to achieve open-loop imitation driving demos in weeks. However, these tools primarily accelerate the steep, early portion of the development curve. Securing the subsequent 'nines' required for superhuman safety, hardware redundancy, and production reliability still demands extensive compounding engineering and roughly ten times the effort for each added nine.

Related Insights

Major AI breakthroughs like Transformers accelerate initial progress but are not silver bullets for the safety-critical long tail. The nature of the problem is that getting a prototype working is relatively easy, but achieving the final "nines" of reliability is incredibly difficult, justifying Google's early, multi-decade investment.

Unlike typical tech development that focuses on capabilities first, Waymo embeds safety as a "non-negotiable foundation" from the start. This means building safety into the model architecture and team mindset, as the approach to achieving 90% performance is fundamentally different from reaching the final "nines" of reliability.

A "vanilla" end-to-end model is insufficient for safety-critical systems. Waymo's foundation model is end-to-end but is augmented with "structured materialized intermediate representation." This allows for crucial runtime validation, richer training, and closed-loop evaluation necessary for superhuman performance at scale.

Drawing from his Tesla experience, Karpathy warns of a massive "demo-to-product gap" in AI. Getting a demo to work 90% of the time is easy. But achieving the reliability needed for a real product is a "march of nines," where each additional 9 of accuracy requires a constant, enormous effort, explaining long development timelines.

Architectures evolve through waves like ConvNets, Transformers, and VLMs. Without strong foundations in infrastructure, data, and rigorous evaluations, teams become trapped in local minima where changing stacks risks regressions. Waymo's primary capability is not current model weights, but the underlying evaluation machinery that enables engineering teams to execute major paradigm shifts without sacrificing safety or stalling deployment progress.

Dmitri Dolgov explains that while AI advancements create hype, they primarily speed up progress on the initial, easier parts of a problem. They don't change the "long tail" of complex, rare edge cases, which remains the core challenge in achieving full, superhuman autonomy.

The main risk to humanoid robotics adoption isn't a lack of impressive capabilities, but the failure to achieve near-perfect reliability. Unlike an LLM where a human can simply re-prompt, a robot's value is in its autonomy. Bridging the gap from 95% to 100% reliability for autonomous tasks is the critical, unsolved challenge.

The current commercial training cadence introduces a dangerous capability overhang. Deploying newly scaled foundation models every six months outpaces the time required to comprehend how existing models function, where their behavioral vulnerabilities lie, or why they fail. As capital requirements jump from tens of millions to billions of dollars, forced pauses between massive training runs could provide a necessary window for researchers to evaluate model bounds rather than blindly stacking parameters.

Achieving near-perfect AV reliability (99.999%) is exponentially harder than getting to 99%. This final push involves solving countless subtle, city-specific issues, from differing traffic light colors and curb heights to unique local sounds like emergency sirens, which vehicles must recognize.

Current robotics demos have low success rates (e.g., 53%). The industry's main challenge, like autonomous vehicles, is bridging the gap from "it works sometimes" to the near-perfect reliability required for commercial deployment, a process that could take a decade.

Modern Foundation Models Flatten Initial Prototyping but Leave High-Nines Reliability Unsolved | RiffOn