No one goes to Hanks for his spreadsheets. They go for a darn good pizza. Lately though, the shop's been quiet. So Hanks decides to bring back the $1 slice. He asks Co-Pilot in Microsoft Excel to look at his sales and costs and help him see if he can afford it. Co-Pilot shows Hanks where the money's going and which little extras make the dollar slice work. Now Hanks has a line out the door. Hanks makes the pizza. Co-Pilot handles the spreadsheets. Learn more at M365 Co-Pilot.com This episode is brought to you by Indeed. Stop waiting around for the perfect candidate. Instead, use Indeed sponsor jobs to find the right people with the right skills fast. It's a simple way to make sure your listing is the first candidate see. According to Indeed data, sponsor jobs have four times more applicants than non-sponsored jobs. So go build your dream team today with Indeed. Get a $75 sponsored job credit at Indeed.com slash podcast. Terms and conditions apply. This episode is brought to you by State Farm. You know those friends who support your preference for podcasts over music on road trips? That's the energy State Farm brings to insurance. With over 19,000 local agents, they help you find the coverage that fits your needs. So you can spend less time worrying about insurance and more time enjoying the ride. Download the State Farm app or go online at StateFarm.com. Like a good neighbor, State Farm is there. Tesla entirely wiped the word autopilot from their navigation software and officially rebranded the car's processing unit from the FSD computer to the AI computer. Yeah and that semantic shift is actually backed by a massive underlying code replacement. I mean the trajectory of this conversation moves right through an exploration of a complete neural network compiler rewrite, a growing physical hardware divide creating two distinct classes of vehicles on the road and the literal gamification of autonomous driving software. So does rewriting a vehicle's neural network compiler actually make a two-ton machine react faster to a squirrel darting into the road? Well to really understand that connection we have to look closely at the MLIR compiler rewrite. That is the multi-level intermediate representation compiler that was just pushed inside the latest full self-driving update. It completely changes how the software interacts with the physical silicon. Hold on, what exactly does an AI compiler do inside a car? Think about it like you were trying to translate a massive complicated book into a new language using a physical dictionary for every single word. Right, exactly. Like looking up each word one by one you eventually get the meaning of the sentence but the process is incredibly slow and you know completely disjointed. Exactly. The old software was essentially doing that. It was constantly looking up instructions to figure out how to communicate with the car's hardware. Yeah, the code had to be translated step by step before the silicon could actually execute a steering command or a breaking decision. Right. But the new MLIR compiler eliminates that dictionary entirely. It allows the neural network to fluently speak the native language of the car's hardware. Wow. And the direct result of that fluency is a 20% reduction in reaction time. 20%. Yeah, it's huge. When you reduce latency in the system like that you allow the car to process far more frames from the external cameras every single second. The physical consequences on the road are immediate and honestly highly noticeable for anyone sitting in the driver's seat. You feel it right away. Yeah. Like you get significantly smoother stops because the vehicle is digesting visual data faster and you know planning its breaking curve earlier. Definitely. The vehicle also improves its handling of complex compound intersections. Those are the worst. Right. We are talking about those really frustrating poorly designed city intersections with curved roads, faded lane lines and multiple confusing traffic light configurations. Yeah. The faster the compiler processes the geometry of the road, while the more confident the steering inputs become. The system even gains the ability to maintain control during temporary system degradations. Like if you are driving directly into harsh sunlight, one of the front cameras is temporarily blinded by the glare, the vehicle can automatically recover. Right. It relies on the remaining sensors. Exactly. Without immediately screaming at you to intervene and take the wheel. Which brings us directly to the training process. The engineers actually dedicated a reinforcement learning phase specifically to small animals. Interesting. Yeah. They fed the system incredibly difficult examples of animals darting into the street from behind parked cars or bushes. The specific focus on those scenarios is fascinating because it relies on granting the system digital rewards for proactive safety. I mean, the neural network operates in a simulation and every time it successfully anticipates a sudden unpredictable movement like from a simulated squirrel or dog without causing an erratic swerve, it earns a mathematical reward. It is literally trained to crave the reward of anticipating the unpredictable. Right. But a faster digital brain requires sufficient physical muscle to process that data flow. And this exposes a permanent physical limitation for older vehicles currently on the road. Yeah, let's lay out the physical hardware differences stratifying the fleet. So the newly named AI computer known internally as hardware for boasts a 20 core CPU. It has 16 gigabytes of RAM and five megapixel cameras. We back up a little over one megapixel on the older ones. Yeah, the older hardware three vehicles rely on a much older 12 volt architecture and cameras with just over one megapixel. That is early smartphone camera quality guiding a vehicle at highway speeds. It sounds wild when you say it out loud, but yes. And the issue goes far deeper than just the camera resolution, right? The older 12 volt wiring harnesses simply like the physical bandwidth to process. It's Toyota truck month. Time to get a truck that works as hard as you do. Tacoma, Tundra, built for the work site, ready for the trail and packed with tech that makes every drive smarter available with 360 degree panoramic cameras for perfect visibility in tight spots, powered tailgate for easy loading and a high tech connected screen to keep you on the grid no matter where you are all backed by the brand known for its legendary reliability. The rugged Toyota Tacoma and the full size Tundra are built to handle it all. And right now your local Toyota dealer has great financing and lease options available to qualified customers, meaning there's no better time to test drive the Toyota truck you need. Find a great Toyota truck month deal today. When you visit buyatoyota.com, that's buyatoyota.com. Toyota, let's go places. And we're back in the miller's yard. Despite the heat, their true green lawn is thriving. They got a lawn like a golf course here in Maryland without wasting a weekend. And PJ to a golf has started chewing up like this, bro. Amazed this grass looks this good and this heat has to dodge the bikes and the recycling bins. Oh, perfectly struck. True green, the easiest way to get a golf course quality lawn. Click the screen now and get started today at truegreen.com. Exclusion supply, see truegreen.com for details. This NBA playoffs with fan duel, you're not just watching the playoffs, you're a part of them. Steve's got the phone opens fan duel and he's going for threes. Be a part of the action on fan duel. All customers get a profit boost every NBA playoff game day. 21 plus and present in Maryland. Opt in required. Bonus issued as not with garbled profit boost tokens. Restriction supply, including any token expiration and max wage or amount. See full terms at fan duel.com slash sportsbook gambling problem. Getting help is your best bet. Please play responsibly for help. Visit MD gambling help.org or call 1 800 gambler. The uncompressed high resolution video streams that the new AI computer demands on its 16 volt system. Yeah, this is just the harsh reality of computer science colliding with physical limits. I mean, you physically cannot push the massive data required by these advanced neural networks through older lower capacity wiring. Yeah, it is a matter of electrical gauge and data pipelines. Eight cameras sending uncompressed video simultaneously create an enormous stream of information. The older 12 volt pipeline is simply too narrow to carry it all without dropping packets of data. I understand the physical limits. I get the computer science and the bandwidth constraints. Yeah. But if you are an owner who purchased one of those older vehicles, assuming you possessed all the necessary sensors for total autonomy, you were likely incredibly frustrated. Oh, absolutely. You were sold a promise of a self driving future. Finding out that a literal physical wiring harness is holding your car back from the latest updates is a really tough pill to swallow. I mean, people paid thousands of dollars for those software packages. The frustration is completely understandable. The unavoidable consequence is that this bandwidth limit forces the older fleet to run late versions of the software. It permanently caps their capability ceiling. Right. The older cars get optimized, compressed versions of the neural network just to function while the newer models receive the full uncompressed suite of advanced processing. And the most highly visible feature strictly gated behind the newer processors isn't even steering related at all. It is the conversational assistant residing in the dashboard. Yeah. This involves the full integration of the Grock AI assistant, which is now accessible entirely hands free via the Hey, Grock wake word. The vehicle utilizes its upgraded infotainment processor to run local natural language processing that completely changes the cabin interface because you no longer need to tap a microphone icon or use rigid specific voice commands to get the car to understand you. It also has location based contextual awareness. Like imagine you are driving and you realize you need milk. You can sit in the driver's seat and simply ask the car to remind you to pick up groceries when it detects you are near your home. Yeah, that is so cool. The vehicle uses its GPS coordinates to trigger the assistant's memory the moment you cross an invisible geographic boundary around your house. Along with that sophisticated voice assistant, we have a completely redesigned self driving app situated right on the main display. Okay. It features a simplified one tap subscription model, removing any friction from purchasing the software. More importantly, it features a dashboard specifically dedicated to tracking self driving streaks alongside your total miles driven. It is exactly like the UI design you find on fitness trackers. Closing your daily activity rings or keeping a language learning streak alive is replaced by maintaining your autonomous driving streak. Right. They take that fundamental human psychological desire to keep a numerical streak alive and apply it directly to vehicle operation. Treating the driving experience like a daily habit app opens up an entirely new revenue approach. When you think about the business model, the entire interface is designed to boost recurring software subscriptions. Yeah, for sure. By keeping drivers constantly engaged with their stats, checking their streak and viewing their autonomous mileage, you make the software feel indispensable. They also pushed lighter UI updates that directly change the physical cabin environment. The ambient cabin lights now glow a bright flashing red to warn you of blind spot dangers. Oh, nice. Yeah, integrating safety directly into the interior lighting scheme. And pet mode includes new customizable avatars for the main screen. If you leave your air conditioning on for your pet, you can select a digital dog, a cat or a cyber hog, wearing pixel sunglasses to display on the screen for anyone looking through the window. That is funny. But while tracking driving streaks and customizing digital hedgehogs entirely alters the digital cabin experience for you, the software updates are fundamentally rewriting how specific models utilize their physical breaks and defend against hackers. Right. We see this clearly with a new feature called comfort braking, which is an exclusive update for newer model Y vehicles. It leverages a brand new dual master cylinder to let the AI smoothly modulate brake pressure during a complete stop. The dual master cylinder allows for an incredibly fine level of physical control. When the vehicle is coming to a halt, the AI can micro adjust the hydraulic pressure on the brake calipers in the final fractions of a second. So smooth. It releases just enough pressure right as the tires stop rolling, which completely eliminates those abrupt jerky halts that toss your head back against the seat. Then you have the Cybertruck exclusives. There is an active road noise reduction system that uses microphones planted inside the seats themselves. Wait, inside the seats. Yeah, inside the seats. Those microphones pick up the low frequency drone of the tires on the highway and instantly generate anti noise through the cabin speakers. It creates isolated quiet zones for the occupants by literally canceling out the sound waves. Wow. And there's also a parked blind spot warning system specific to that vehicle. If you are parked on a busy street and the external cameras detect an approaching cyclist coming up along the side, the software physically disables the interior door unlatch button on your first press. It forces you to pause, preventing you from accidentally throwing your door open directly into the path of the cyclist. Exactly. Wait, hold on. Let's reset the pace here for a second. So software now dictates physical mechanics, but that exact software is under constant attack. Yeah. A recent PIN2 owned security competition highlighted this vulnerability perfectly. A research team managed to achieve route permissions on the vehicle system via a sophisticated USB attack. Right. By plugging a device into the physical port, they bypassed the user level permissions and gained direct access to the core operating system. And once they had route access, they exposed private data, access the vehicle's stored Wi-Fi credentials, and potentially even gain the ability to execute physical vehicle controls. Scary stuff. Yeah. Tesla countered the attack by immediately deploying an over the air security patch to the entire fleet. The ability to patch a car automatically while it is parked overnight in your garage drastically limits the lifespan of a cyber threat. The vulnerability existed, but it was closed almost immediately without anyone needing to visit a dealership or plug in. This episode is brought to you by MGM plus from and its horse returns for season four centered around a town that traps all those who enter terrifying creatures that come out at night horrors. Some was to stay buried where desperate hope leads to darker truths. Dangerous new arrivals. A sinister figure. A shocking revelation from season four premieres April 19th on MGM plus. Some doors should remain closed. Pepsi prebiotic cola in original and cherry vanilla. That Pepsi taste you low with no artificial sweeteners and three grams of prebiotic fiber, Pepsi prebiotic cola. Unbelievably Pepsi. Your favorite local grocery stores like Kroger, Ralph's, Fred Meyer and more are now delivering on Uber Eats. Get 40% off your order of fresh quality ingredients. Whether you just got home to an empty fridge or suddenly got a craving to whip up something new, you can get everything you need delivered in as little as 25 minutes. Get 40% off your order with code Kroger 2026. Plus members get zero dollar delivery fees. Order now on Uber Eats. Orders of $30 or more save up to $25 and four 3026. Yeah, for details. The diagnostic tool. However, that reality permanently links passenger safety to the absolute integrity of a digital firewall. You are relying on the manufacturer to find and patch those holes faster than malicious actors can exploit them. Yeah, the physical environment inside the vehicle, like the doors, the brakes, the steering is completely dependent on how secure that software layer remains against external manipulation. Tesla software updates are transforming vehicles into hardware-dependent AI nodes. Reaction times, stopping smoothness and voice interfaces evolve automatically while parked, assuming the car possesses the physical processing power to handle the code. As the software demands outgrow the physical wiring of millions of cars already on the road, will the unified driving experience initially promised by the manufacturer permanently fracture into different classes of capability? If you're not subscribed yet, take a second and hit follow on whatever app you're using. It helps us keep making this. We appreciate you being here.
