349: The Agent Wave Is Coming

L1 Automotive Training is referenced as a provider of specialized training content for automotive diagnostics and programming topics. The hosts highlight it as a resource for learning module programming and related security/electrical diagnostics.

Keith Perkins is mentioned as the expert behind the training resource, with a focus on module programming, J2534, and related diagnostic topics. The host credits his videos with helping them understand the tooling and obstacles involved in programming modules.

J2534 refers to the SAE J2534 standard for “Pass-Thru” programming devices used to communicate with a vehicle’s modules. It’s commonly used for tasks like reprogramming modules without needing the OEM’s proprietary tool.

Key and immobilizer work involves programming or pairing keys to the vehicle’s security system so the car will start. Immobilizers prevent unauthorized starting by requiring the correct key authentication before the engine can run.

Module programming is the process of updating or re-flashing a vehicle’s electronic control units (ECUs) with new software. It’s often required after repairs, part replacements, or when addressing drivability and diagnostic trouble codes.

“eProm work” refers to programming or replacing data stored in EEPROM/EPROM-type memory chips used by some vehicle modules. In diagnostics and repair, it can come up when a module’s calibration or security data needs to be read, modified, or restored.

Electrical diagnostics is troubleshooting faults in a vehicle’s wiring, sensors, power/ground circuits, and module communication. It often involves interpreting scan data alongside measurements like voltage drop, resistance, and current draw.

Drivability diagnostics focuses on drivability-related symptoms like misfires, hesitation, poor acceleration, stalling, and rough idle. It typically combines scan tool data (DTCs, live data) with mechanical checks and test drives to pinpoint the root cause.

Open Claw is described as an open-source “harness” for AI agents. The idea is to give agents access to tools (like email or coding), define skills and workflows, and even schedule tasks so the agent can perform them automatically.

An AI agent is more than a chatbot: it can take actions to complete tasks. In this segment, the host describes agents as systems that can execute steps like sending emails or creating files, not just generating text.

LLM stands for “large language model,” which is an AI system trained to understand and generate language. In the segment, the host contrasts an LLM (text generation) with an AI agent that can execute tasks.

Scheduling refers to setting recurring tasks for the AI agent to run automatically at specific times. In the example, the agent is set to produce a daily report and email it to technicians.

An ABS module is the electronic control unit that manages the anti-lock braking system. When it’s replaced or reprogrammed, the vehicle often needs calibration and security/authentication steps so the module can communicate correctly with the rest of the car’s systems.

VCI typically refers to a vehicle communication interface—an adapter that lets a diagnostic laptop talk to the car’s onboard systems through the diagnostic port. It’s commonly required for programming tasks like module re-flashes or configuration.

Some module programming procedures require the vehicle’s key/immobilizer setup to be present in a specific way—often meaning the shop must have two keys available for authentication or configuration. This is common when security systems are involved in the programming flow.

“VSP” here appears to refer to a specific set of vehicle/service paperwork that must be completed for the job. The speaker emphasizes that these forms are captured automatically by their system, reducing missed steps and ensuring compliance/documentation is handled consistently.

A knowledge base in a shop context is a structured library of repair information—like steps, pitfalls, and vehicle-specific notes—so technicians can repeat successful procedures. The speaker describes using voice input to automatically format and store entries, which helps standardize documentation across jobs.

OTIS is Audi’s online technical information and service platform used by technicians for repair procedures and service data. In this segment, it’s referenced as part of the workflow for programming/repair steps related to an ABS module.

The Audi Q5 is a compact luxury SUV, and the 2020 model is part of Audi’s second-generation Q5 lineup. In this segment, it’s mentioned as the vehicle the shop worked on, specifically in the context of an ABS-related repair.

The hosts discuss how fast-food drive-throughs are starting to use AI at the ordering window. This is a “where you encounter it” moment—AI is becoming part of everyday services, which parallels how vehicles may increasingly integrate AI features.

This is about a navigation/communication failure when a phone stops working. For drivers, it underscores the value of redundancy—like built-in GPS, offline maps, and non-cellular guidance—so you’re not stranded when your phone fails.

GPS (Global Positioning System) is a satellite-based navigation system that helps determine your location and provide turn-by-turn directions. The hosts are discussing how common GPS became and how it changes the way people learn routes and avoid getting lost.

TomTom is a well-known brand of in-car and portable navigation devices. In the segment, it’s referenced as part of the early wave of GPS becoming mainstream for drivers.

Garmin is another major GPS navigation brand, commonly used in cars and for personal navigation. The hosts mention it to describe the period when GPS devices became widely used by everyday drivers.

Turn-by-turn navigation is the feature where a GPS system gives step-by-step directions as you drive. The hosts contrast relying on GPS with learning geography and street layouts, especially in a city where “true north” doesn’t match local orientation.

AI chatbots are software agents that generate responses based on patterns learned from data. In the transcript, they’re used as an example of AI systems communicating and evolving their own “language,” which is a common concern in discussions about advanced AI behavior. For cars, the takeaway is that future vehicle software may communicate with other systems, increasing the importance of monitoring and safety controls.

A “sandbox” is a controlled testing environment where an AI or software system is isolated to limit what it can access. The transcript describes putting a model in a sandbox to see if it could escape, which is a way researchers test containment and safety boundaries. In automotive terms, this maps to the idea of isolating vehicle subsystems so one compromised component can’t take over the whole car.

Cybersecurity risks refer to the possibility that connected systems can be hacked or manipulated. In modern vehicles, that matters because cars increasingly rely on software, networks, and cloud services, so an attack could affect functions or data. The key idea is that as cars become more connected, the “attack surface” grows beyond just mechanical systems.

The hosts describe a shift toward custom, niche software built for very specific automotive or diagnostic needs. Instead of large companies building everything at scale, individuals can now create targeted tools that solve a narrow problem for a small group. In diagnostics, that can mean utilities that validate files, streamline workflows, or reduce human error.

A “Nissan file” here refers to a specific software/data file used with Nissan diagnostic or programming workflows. In practice, technicians verify they have the correct file for the vehicle so the right calibration or configuration is applied. Using the wrong file can lead to failed programming or incorrect vehicle behavior.

A “mobile programming business” is a service model where a technician travels (or works remotely) to perform vehicle software programming and diagnostics. It often relies on having the right tools and data files on hand, plus the ability to quickly verify compatibility. This segment frames how niche software tools can directly support that kind of work.

“Rate of acceleration” refers to how quickly acceleration changes over time, which is a concept from vehicle dynamics and physics. In driving terms, it’s related to how abruptly a car’s speed is increasing as you press the accelerator and how that change evolves.

The segment discusses how automation and AI-driven customer support can reduce demand for certain roles, such as middle management tasks and customer service representatives. For automotive businesses, this can shift work toward exception handling, technical escalation, and human-in-the-loop review.

The hosts are describing AI systems that automatically review customer interactions (like phone calls) and score or categorize them. In automotive service and diagnostics, this kind of tooling can help route issues, flag poor experiences, and improve scripts or workflows over time.

CRM stands for Customer Relationship Management. In automotive contexts, CRMs often track leads, service requests, call outcomes, and customer history, and they may integrate AI to summarize calls or recommend next steps.

“AI coding” refers to using AI tools to generate or assist with software code. In automotive-adjacent businesses, this can change how technicians and service workflows interpret codes, search for fixes, and communicate with customers—sometimes leading to mismatched assumptions if the underlying diagnosis isn’t verified.

The story describes a common diagnostic failure mode: installing parts based on a customer’s assumption (often tied to a code or AI-suggested fix) without confirming the root cause. When the part doesn’t resolve the issue, it creates repeated “no dice” repairs and confusion about what the customer is actually trying to accomplish.

Spark plugs ignite the air-fuel mixture in each cylinder. During misfire diagnostics, technicians often inspect and replace spark plugs, since worn plugs can contribute to ignition problems or misfire codes.

A misfire means one or more cylinders aren’t combusting properly. It can be caused by ignition issues (like coils or spark plugs) or fuel/air problems, and it often triggers diagnostic trouble codes.

An ignition coil converts battery voltage into the high voltage needed to fire the spark plug. If a coil is weak or failing, it can cause a cylinder misfire, and swapping coils can help confirm the diagnosis.

The intake manifold is the part that routes air to the engine’s cylinders. In this case, the coil was “buried on the intake,” meaning access required removing intake components, which increases labor time and cost.

A “scope” refers to using a diagnostic oscilloscope to view electrical waveforms from sensors and ignition components. Scoping can help confirm which cylinder is misfiring and whether ignition signals are present and correct.

The speaker criticizes relying on AI (GPT) to prescribe parts without verifying the vehicle’s actual condition. In diagnostics, “wide net” recommendations can waste money if the failure is specific or if multiple unrelated codes exist.

An oxygen sensor (O2 sensor) code indicates the engine control system detected an issue with exhaust oxygen readings. The speaker specifically mentions a heater-circuit-related code, which can be separate from misfire causes.

A heater circuit code for an oxygen sensor points to a problem with the sensor’s built-in heating element. The heater helps the sensor reach operating temperature quickly; if it fails, the sensor may not provide accurate readings.

Isaac Rodel is mentioned alongside Auto Rescue Tools as part of the show’s sponsor/read. The segment implies involvement with programming laptops or setup for multi-brand vehicle programming.

Auto Rescue Tools is mentioned as a sponsor in the segment. The context suggests they provide programming or diagnostic equipment used by technicians.

Diagnostic equipment is a broader category of tools used to test and troubleshoot vehicle systems beyond basic code reading. Depending on the job, it may include oscilloscopes, battery testers, wiring/connector testers, and programming interfaces.

Scan tools are handheld or laptop-based devices that read diagnostic trouble codes (DTCs) from a vehicle’s onboard computer. They’re used to pinpoint which system is malfunctioning and to view live sensor data for troubleshooting.

Key cutting equipment is used to physically cut a replacement key blade to match the vehicle’s lock or ignition. Many modern keys also require electronic programming (transponder/immobilizer), so cutting is only part of the process.

The hosts discuss using ChatGPT-like AI for automotive diagnostics and coding tasks. The key point is that AI can be helpful for general research, but it may be unreliable for very specific, high-stakes procedures where accuracy and context matter.

Cloning a module means copying the data from one electronic control unit (ECU/module) to another so the replacement behaves like the original. This is often discussed in the context of immobilizers, instrument clusters, or other security-related modules.

A “used module” refers to installing a previously used electronic control unit (ECU/module) instead of a brand-new one. Many modern cars require programming/adaptation and sometimes specific procedures to ensure the used module communicates correctly with the rest of the vehicle.

Parking sensors are the proximity-detection system that warns you when you’re close to objects while parking. Calibration is the process of teaching the system the correct reference points so it can interpret sensor readings correctly.

In this context, the “parking light” is the indicator/warning behavior that stays on when the parking sensor system isn’t fully cleared. The hosts describe a specific procedure (cycling the mirrors) that resolves the light after sensor calibration.

Folding the mirrors in and out is described as a required service/initialization step to clear a parking-light condition on newer Lexus vehicles. It’s a reminder that some vehicle systems need a specific “state change” or initialization sequence after calibration or troubleshooting.

The BMW 5 Series is a mid-size luxury sedan that’s known for advanced electronics and systems that can require specialized diagnostic procedures. In the podcast context, it’s mentioned alongside an “E39 module” reference, which points to older 5 Series-era components being identified, programmed, or discussed for diagnostic work. That kind of mention typically comes up because module replacement or coding can be a key part of troubleshooting.

A “pinout” is the mapping of connector pins to specific circuits or signals on a module. In diagnostics and module cloning, pinouts are critical so you can wire or interface correctly without guessing.

“Programming and diagnostics” refers to using scan tools to read fault codes and also to update or configure electronic modules. Many modern cars require module coding/programming after repairs, replacements, or certain service actions.

“ADOS” is mentioned as part of the speaker’s diagnostic/programming knowledge base. Without the full context or expansion, it’s unclear whether this is a specific tool name, software suite, or internal acronym used in their shop.

Blind spot indicator lights are part of a driver-assistance system that warns when a vehicle is detected in the blind spot. The speaker notes a behavior where the lights stay on until a speed threshold is reached, which can be normal system logic rather than a failure.

Tape measures are mentioned as part of a hands-on diagnostic or fabrication workflow—measuring repeatedly to confirm fitment, routing, or dimensions. In automotive work, accurate measurements can be critical when diagnosing issues related to alignment, clearance, or harness routing.

Heat shrink is an insulating sleeve used on wiring connections. When heated, it contracts to provide mechanical strain relief and electrical insulation—common in automotive repair and harness work.

“Buck connectors” likely refers to quick-connect or butt-style electrical connectors used to join wires in automotive harness repairs. The key idea is that proper connectors and crimping/insulation are essential for reliable electrical connections.

“ECU” stands for engine control unit, and Bosch is a major supplier of automotive electronics. An ECU part number is the specific identifier used to make sure you’re ordering or referencing the exact control module that matches the vehicle’s configuration.

Stellantis is the parent company behind multiple automakers (including Chrysler). A “module” in this context is a vehicle control unit or electronic subsystem that may have specific data access and integration requirements for diagnostics and automation.

“E90” refers to the BMW 3 Series generation (the 3 Series built in the E90 chassis). In this segment it’s mentioned alongside “modules,” implying the discussion is about accessing or modifying vehicle electronic control units (ECUs) or diagnostic/telematics components on that platform.

The Chevrolet Silverado is GM’s full-size pickup line, and “2020 Silverado” narrows it to that model year. Here it’s grouped with an “E90 module” and talk of cloning/unlocking, suggesting the conversation is about vehicle electronics security and access to modules on different vehicles.

“Unlocking” typically means gaining authorized or exploit-based access to a vehicle module’s functions (or enabling a feature) without necessarily duplicating the module. In modern vehicles, this is often constrained by cryptography, secure boot, and manufacturer security policies.

This segment is primarily about whether progress in the “cloning world” is slowing down, and why new cracked tools aren’t appearing as quickly. It frames the episode’s discussion around the pace of automotive electronic security research and tooling.

The speaker is describing a “stagnation” in the ecosystem of tools used to access or modify vehicle modules—suggesting fewer new breakthroughs or cracked security methods. In automotive diagnostics/security, this often happens when manufacturers patch vulnerabilities faster than the community can develop reliable tools.

“Proprietary” here refers to manufacturer-controlled diagnostic/security logic that’s not publicly documented. The point is that AI or automation may still run into restrictions when systems are designed to prevent unauthorized access or copying.

“Local models” refers to running AI models on your own hardware instead of relying on a hosted service. In the context of this discussion, the speaker suggests local AI could help develop programs/protocols for reading or working with protected data, while avoiding some platform restrictions.

Job displacement is the idea that automation and AI can replace human workers for certain tasks. In discussions like this, it usually refers to roles that involve routine decision-making, communication, or service work that software/robots can perform.

“The Matrix” is a reference to a fictional scenario where humans are used as power sources by machines. The speaker uses it as a metaphor for fears about humans being exploited or controlled by advanced AI/robots.

“Human battery” is a metaphor for treating people as energy storage or a resource for machines. It’s not an automotive term, but it’s a notable sci-fi concept being used to express concern about how AI could exploit humans.

The hosts reference a psychological study to explain why people feel different based on upbringing and life experiences. In car terms, this is analogous to how driver behavior and habits are shaped by early exposure and consequences.

The speaker emphasizes that choices have consequences, which is a useful mindset for diagnostics: small issues can become bigger problems if ignored. This frames why proper maintenance and timely repairs matter in vehicle ownership.

A “bike chain” issue is used as an example of a mechanical problem you might need to troubleshoot on your own. While it’s not a car part, the analogy maps to how drivers should be able to recognize and respond to mechanical failures.

The hosts mention that school kids started learning on iPads around COVID. In a car-enthusiast context, this is more about how technology changes learning habits than about vehicles, but it’s still a notable tech reference that frames the discussion.

Durability and reliability tests are how automakers prove components can survive real-world use over time. The hosts argue that modern testing has become overly “math-driven,” which can miss practical issues that show up once cars are in the hands of customers.

A vehicle recall is when a manufacturer (or regulator) identifies a safety-related defect and requires repairs at no cost. The hosts claim that every automaker is dealing with large-scale recalls, suggesting systemic issues rather than isolated problems.

A transmission is the drivetrain component that manages engine power and gear ratios to match driving conditions. The segment suggests a high volume of transmission removals and related repairs, implying widespread drivability or durability problems.

The hosts specifically call out the Toyota Highlander model years 2016 through 2024 as having transmission issues. They claim the transmission problems are “dropping worse” than another vehicle comparison, implying a notable reliability concern for that range.

This appears to refer to the U.S. National Highway Traffic Safety Administration (NHTSA), which tracks vehicle safety issues and publishes recall data. The hosts use its “most recalled” list to support their claim about which brands are seeing the most recalls.

Ford is named as the #1 most-recalled brand in the hosts’ referenced list. In a recall context, this typically means Ford vehicles have a higher count of safety-related issues requiring manufacturer fixes.