Connect with us

News

Elon Musk recaps Tesla FSD neural network at All-In Summit 2023

Credit: All-In Podcast

Published

on

During a conference this week, Tesla CEO Elon Musk reiterated claims that the Full Self-Driving (FSD) beta is nearing higher levels of autonomy. The statements echo recent details learned from previews of a new Musk biography, highlighting the FSD system’s many developments in the last several months alone.

Musk was featured in an interview during the All-In Podcast’s 2023 Summit held on Wednesday, during which he discussed topics like Starlink, X, China, artificial intelligence, and more. Among the topics covered was a brief outro on the FSD beta, which he says is “very close” to becoming safer than a human driver without being monitored.

“Yeah, I think it’s getting very close to being in a situation where, even if there’s no human oversight or intervention, that the probability of a safe journey is higher with FSD and no supervision — like even if you’re asleep in the car — than if a person is driving. We’re very close to that,” Musk said on a video call into the summit.

He also pointed to how much the current version of FSD has improved compared to versions from six months ago, a year ago, and 18 months ago. In that time, Musk explained, Tesla has arrived at what it calls the “final piece of the puzzle,” as the company shifts toward making the whole FSD neural network-based.

“And we’ve got the final piece of the puzzle, which is to have the control part of the car transition from about 300,000 lines of C++ code to a [full-scale] neural network, so the whole system will be neural network, photons into controls out,” Musk added.

Advertisement

During the interview, Musk also referred to miles being driven with FSD as being “much safer” than those without it, emphasizing how close the automaker is to solving that “final piece of the puzzle.”

You can watch Musk’s full interview at the 2023 All-In Summit on X, with the FSD discussion taking place around 44:30.

Tesla Safety Statistics

Tesla releases a regular vehicle safety report detailing quarterly crash results from its Autopilot and FSD systems.

The most recent entry occurred in the fourth quarter of last year, showing fewer crashes when Autopilot technology is engaged than when it isn’t. According to the report, the latest data from the National Highway Traffic Safety Administration (NHTSA) shows that there is an automobile crash in the U.S. every 652,000 miles.

Comparatively, Tesla recorded one crash for every 4.85 million miles driven in Q4 for drivers using its Autopilot systems. Tesla also recorded one crash for every 1.40 million miles driven by drivers in its cars who were not using either of the systems.

The concept of FSD switching to an entirely neural-network-based approach was also detailed earlier this week in previews of a Musk biography written by Walter Isaacson. The biography was published on Wednesday, and it specifically discusses Tesla’s upcoming FSD version 12.

Advertisement

In the past, Musk has said that FSD v12 will no longer be a beta version of the software and will include a significant leap forward.

Tesla FSD gets price cut in the United States

What are your thoughts? Let me know at zach@teslarati.com, find me on X at @zacharyvisconti, or send your tips to us at tips@teslarati.com.

Zach is a renewable energy reporter who has been covering electric vehicles since 2020. He grew up in Fremont, California, and he currently lives in Colorado. His work has appeared in the Chicago Tribune, KRON4 San Francisco, FOX31 Denver, InsideEVs, CleanTechnica, and many other publications. When he isn't covering Tesla or other EV companies, you can find him writing and performing music, drinking a good cup of coffee, or hanging out with his cats, Banks and Freddie. Reach out at zach@teslarati.com, find him on X at @zacharyvisconti, or send us tips at tips@teslarati.com.

Advertisement
Comments

Elon Musk

Elon Musk handed Grok something no other AI company can get their hands on

Elon Musk says SpaceX will feed engineering data into Grok’s next model, avoiding restricted material.

Published

on

By

Artistic concept rendering of SpaceX data being incorporated into a Grok AI model

Elon Musk said Tuesday that SpaceX will feed its internal engineering data into the next major training run for Grok, the AI model now folded into SpaceX following February’s merger. In a post on X, Musk wrote that SpaceX’s “massive corpus of world-class engineering data,” excluding anything restricted under U.S. arms export law, will be added during supplemental training of what he called the “2T run,” a reference to a roughly two trillion parameter model that would nearly double the parameters behind the latest Grok 4.5 that’s rolling out.

The excluded material that Musk is referring to would fall under the International Traffic in Arms Regulations (ITAR), which restricts export of technical data tied to defense and space hardware. That likely rules out propulsion specifics for Merlin and Raptor engines along with guidance and control details for SpaceX’s launch vehicles, but leaves manufacturing knowledge, materials science, and Starlink hardware design on the table.

Advertisement

The announcement extends a pattern that has been building since SpaceX’s Nasdaq debut in June, when the company went public with Grok and xAI’s Colossus supercomputer folded into the pitch to investors.

Days after that listing, SpaceX closed its $60 billion all stock acquisition of coding startup Cursor, giving xAI both enterprise software distribution and a stream of real world developer data to train on. Grok 4.5 launched July 8 running partly on that Cursor training data, with Musk describing it as roughly comparable to Anthropic’s Opus 4.7 but faster and cheaper to run.

Feeding SpaceX’s own engineering data into the next AI model follows the same logic Musk has applied across xAI’s sister companies. Tesla supplies real world driving data and manufacturing expertise, X supplies conversational data, and now SpaceX supplies aerospace engineering data built up since 2002.

Musk did not give a release date for the upcoming AI model, referred to elsewhere as Grok 4.6. He has said the two trillion parameter run is in its final training phase and expected to wrap this week.

Advertisement
Continue Reading

News

Tesla expands ridesharing service in California to new hotspot

Published

on

Credit: Tesla

Tesla has extended its Bay Area ride-hailing service to include pickups and drop-offs at San Francisco International Airport (SFO). The update, shared via the company’s official channels on July 21, allows users in the region to request rides directly to and from one of California’s busiest airports.

The expansion builds on Tesla’s secured limousine permit for SFO operations. Public records show the permit became effective March 20, 2026, and remains active through January 31, 2027. Tesla vehicles operating the service now display authorized limousine permits issued by the City and County of San Francisco.

Tesla’s ride-hailing program in California relies on Model Y vehicles equipped with Full Self-Driving (Supervised) technology. Human safety drivers remain present in compliance with state regulations, distinguishing the service from fully driverless operations.

Advertisement

The Bay Area geofence covers a broad area spanning north of San Francisco to south of San Jose, offering extensive connectivity across the region.

UPDATE: Elon Musk reveals why Tesla didn’t say ‘Robotaxi’ upon California launch

This SFO addition follows earlier progress at other Bay Area airports. Tesla previously expanded service to San Jose Mineta International Airport (SJC) in late 2025. The company had engaged with SFO, SJC, and Oakland International Airport officials as early as September 2025 to secure necessary approvals for passenger transport.

The service provides a new option for travelers seeking electric, app-based transportation integrated with Tesla’s ecosystem. Rides are booked through Tesla’s dedicated ride-hailing application, which handles matching, routing, and payments. Pricing follows standard ride-hailing models, with potential adjustments based on distance, time, and demand.

Tesla’s California ride-hailing program launched in July 2025 with an initial invite-only rollout in the Bay Area. It started alongside operations in Austin, Texas, marking the company’s second major U.S. market.

Advertisement

The Bay Area remains a primary focus in California, with service centered on high-demand corridors connecting residential, commercial, and now major transportation hubs. This latest airport integration represents a practical step in Tesla’s broader mobility ambitions within the state.

Continue Reading

News

Tesla reveals 2026 Summer Update with crazy fixes to Nav and more

Published

on

Credit: Tesla

Tesla has officially revealed its 2026 Summer Update, which comes with a variety of crazy new features, including Navigation fixes that owners have been wanting for months.

Tesla routinely releases a larger update with the Spring, Summer, Fall, and Winter updates, where it ships a variety of new features, bug fixes, and other additions to customer cars.

The 2026 Spring Update featured things like “Hey Grok” voice assistance, a redesigned self-driving app, Unreal Engine visual upgrades, and more.

Tesla’s Summer Release has about ten new features; we’ll show you each and detail them below:

New Grok Voice Commands

“Grok can now make phone calls, search and play music, adjust climate, open the glovebox, and answer questions about your Tesla.”

Self-Driving Stats in Mobile App

“View and share self-driving stats from the mobile app.”

Caraoke With Scoring

“Caraoke now scores your singing while in Park. High scores are saved to your Tesla profile.”

Advertisement

Automatic Navigation

“Automatic Navigation now adapts to your routine.

In addition to Home, Work, and upcoming calendar events, your vehicle can now suggest and route to places you visit regularly – like a school drop-off on the way to work, or the gym on the way home.”

Preferred Routes

“For a more personalized experience, navigation now prioritizes routes that you’ve taken before”

Set Arrival Energy from Mobile App

“Set your desired Arrival Energy from your phone.”

Send Custom Wraps from Mobile App

“Skip the USB drive and upload a custom wrap of your car from the mobile app. Instructions for creating a custom wrap here: https://github.com/teslamotors/custom-wraps.”

Advertisement

Rear Display Lock

“Kids can watch content on the rear screen, but only the front row can control it through the rear screen app.”

Other Improvements

  • Find Superchargers by name when searching for a destination
  • Add Apple Music songs to queue from search and artist page
  • Set your preferred zoom level for the Self-Driving visualization
  • Intro animations for new Model 3 and Y
Continue Reading