Elon Musk claims Tesla is “very close” to developing fully autonomous vehicles despite Tesla's “Full Self Driving” feature not living up to its name since 2016
it hasn't Adrian Potoroaca / TechSpot : Elon Musk: Tesla is close to achieving basic level 5 autonomy for its cars VentureBeat : Elon Musk says Tesla is ‘very close’ to level 5 autonomous driving Chinmay Pandya / CCN.com : Did Greedy Elon Musk Peddle Another Lie to Pocket a Hefty Payday? Usman Khan Lodhi / Neowin : Musk: Tesla is ‘very close’ to achieving level-five autonomous driving technology Chris Davies / SlashGear : Tesla Level 5 autonomy “very close” Musk insists but the true roadmap is murky Darrell Etherington / TechCrunch : Elon Musk sets update on brain-computer interface company Neuralink for August 28 Tweets: Rodney Brooks / @rodneyabrooks : What Elon says here is false. We will not have level 5 autonomous cars on our roads in 2020. It will not happen. https://www.bbc.com/... Costa Samaras / @costasamaras : check out the dates on these posts https://twitter.com/... Grady Booch / @grady_booch : The software is 80% done. https://www.bloomberg.com/... Toby Walsh / @tobywalsh : What's the opposite of a super-forecaster? Someone who always gets predictions wrong. Musk: “I remain confident that we will have the basic functionality for level five autonomy complete this year” https://www.bloomberg.com/... Grady Booch / @grady_booch : “There are many small problems, and then there's the challenge of solving all those small problems and then putting the whole system together, and just keep addressing the long tail of problems.” — Elon Musk This is what we in the field of software engineering call “nonsense”. https://twitter.com/... Oliver Cameron / @olivercameron : To clarify, “the basic functionality for level five autonomy” does not mean you'll see Tesla's with no drivers behind the wheel this year. It means the features needed for L5 may be built (e.g. traffic light detection), but a human will still need to take over frequently. https://twitter.com/... Christopher Mims / @mims : feels like I've heard this before? in nearly every year of the past 5, perhaps? https://twitter.com/... Adam Banks / @adambanksdotcom : He said close to “the basic functionality for level five autonomy”. A very different claim from having a self-driving vehicle that drives itself on actual roads. What he calls the “long tail of small problems” will be a decades-long negotiation between vehicle and road design https://twitter.com/... Matt Rosoff / @mattrosoff : Recall last April he promised 1 million robotaxis on the road this year: https://www.google.com/... https://twitter.com/... @harshmadhusudan : He has been saying “this year” for some time now. https://twitter.com/... Alex Webb / @atbwebb : Hasn't @Waymo been “very close” to developing fully autonomous vehicles for the best part of a decade? That's the whole point. The final 3% that are the hardest. https://www.bloomberg.com/... via @technology
Context & Ripple Effects
The level 5 claim is the latest turn in a decade-long pattern. In 2015 Tesla said self-driving was coming to US roads within months, limited to major highways with regulatory questions unresolved (that three-month timeline never materialized); a year later it promised that every new car already carried full self-driving hardware, with capabilities to arrive via software updates over time.
By 2019 the shortfall was explicit enough that coverage described Tesla shifting the goalposts while customers effectively served as unpaid safety drivers. The new "very close" framing matters because it reuses the same language of imminence that each prior milestone failed to meet — and because the bet against lidar remains the load-bearing technical assumption.
First-order effects
- Tesla owners who bought the "Full Self Driving" option face another cycle where the feature's name outruns its capability, sustaining the criticism that has followed the product since 2016.
- Musk's claim puts his own credibility directly at stake: multiple experts immediately called the level 5 assertion false or misleading, making this a public test of founder claims versus engineering reality.
Second-order effects
- Rivals like Waymo, which took the opposite sensor bet, gain a contrast point: if Tesla's camera-only stack keeps missing deadlines, the lidar-based route looks less like over-engineering and more like prudence.
- The subsequent "extremely slow and cautious" FSD beta rollout functions as an implicit walk-back of the "very close" claim — Tesla shipping narrowly rather than broadly signals the gap between marketing and deployability.
Third-order effects
- Four years later, experts were still arguing that Tesla's imitation learning approach needs AI breakthroughs that may not arrive soon (the WSJ critique) — suggesting the structural issue is methodological, not a matter of iteration time.
- If the pattern of imminent claims followed by slipped timelines holds, regulators and buyers are pushed toward discounting vendor autonomy declarations entirely, raising the evidentiary bar for any level 5 certification claim.
The trend: Consumer autonomy is splitting between founder-declared timelines and demonstrated capability, with Tesla's camera-only bet increasingly defined by what it still cannot ship.