FSD – vår tids største bløff?

[Motorwatt]

Det er nok ikke bare avvikene som har tatt tid. Dårlig planlegging, og for høye ambisjoner helt fra starten må nok også ta en stor del av skylden. Husker ikke om det er 3 eller 4 "rewrites" som er gjennomført siden 2016. I tillegg kommer til tider stor gjennomtrekk i utviklingsavdelingen, og en ustyrlig daglig leder med et vell av løfter med urealistiske tidshorisonter som sannsynligvis er tatt med minimal kommunikasjon, har nok også gjort sitt.

[kodax]

Det har kanskje noe med at flere her har hatt Tesla med FSD siden de lanserte det, og har sett hvor sakte det går.
...

Er det ikke det som er selve definisjonen på gamlisinnstilling, da

[JorEl]

Tesla har jobbet rask og godt. Mobileye startet i 1999 og Tesla ca 15år senere.
Les hva Mobileye sier om seg selv. 1700 ansatte, kun kamera, ingen tro på radar ol. Og en lovnad/mål om "ready Fully Autonomous Vehicles with BMW and Intel, with production launch planned for 2021"

https://www.mobileye.com/about/
Mobileye, an Intel company, was launched in 1999 with the belief that vision-safety technology will make our roads safer, reduce traffic congestion and save lives. With a cutting edge team of more than 1,700 employees, Mobileye has developed a range of software products that is deployed on a proprietary family of computer chips named EyeQ®.

Mobileye is a leading supplier of software that enables Advanced Driver Assist Systems (ADAS), with more than 25 automaker partners including some of the world's largest. Beyond ADAS, our technology has rapidly evolved to also support the three pillars of Autonomous Driving – Sensing, Mapping, and Driving Policy. As a result of this broad and well-advanced product offering, we have achieved a partnership to develop production-ready Fully Autonomous Vehicles with BMW and Intel, with production launch planned for 2021, and another partnership with the Tier-1 supplier Delphi for a "turnkey" system to be productized starting from 2019 with customer OEMs.

We are dedicated to continuing our two-decade tradition of leveraging the newest technology, including AI, deep learning and crowdsourcing, to create the hardware and software needed to help our over 25 OEMs partners enable ADAS systems and eventually fully autonomous vehicles. These partners choose Mobileye due to our track-record of innovation, but also because our systems have been validated to the highest level of safety and accuracy over millions of miles.

Inspired by Human Vision
Mobileye's strength comes in part from our early realization that a single-lensed camera (mono-camera) would become the primary sensor to support Advanced Driver Assistance Systems (ADAS) and eventually Autonomous Vehicles. The mono-camera is the most versatile of sensing technologies, with the ability to identify shapes, like vehicles and pedestrians, as well as textures, like lane markings and traffic-sign text. This realization marked a revolutionary leap in the market, as sensors like dual-lensed cameras (stereo-vision) and radar were previously regarded as superior. Today, most global automakers have chosen mono-vision as the primary information source for Advanced Driver Assistance Systems, and Mobileye continues to lead innovation in the field.

"The Mobileye mono-camera was inspired by human vision, which only uses both eyes to obtain depth perception for very short distances," says Amnon Shashua, co-founder, chairman and CEO of Mobileye. "Therefore, the added benefit of a second camera lens is only relevant for short distances. Driving-scene interpretation is based on much longer distances. All depth-perception cues for farther distances – such as perspective, shading, texture, and motion cues, that the human visual system uses in order to understand the visual world – are interpreted by a single eye. Therefore, Mobileye understood that a single-lens camera could be the primary sensor to enable autonomous driving."

Mobileye's vision safety technology for ADAS is deployed on over 40 million vehicles and counting, making today's roadways safer for all. We work closely with global automakers to continuously deploy advancements in safety and convenience. Over the next several years, as more stringent safety regulations are implemented in major markets, it is expected that the percentage of newly-produced vehicles that include ADAS will increase.

Building on ADAS products like Automatic Emergency Braking and Lane Keeping Assistance, that form the building blocks for higher level autonomous vehicles, Mobileye is now using the latest techniques in Artificial Intelligence and deep learning to deploy 360-degree surround-view mono-vision sensing; to build high-definition, crowdsourced maps; and to provide vehicles with the computer intelligence required to negotiate complex driving situations. We believe that these technologies, along with important technological inputs developed by our automaker and supplier partners, can eventually result in a world where vehicle injuries and fatalities can be reduced by a very significant amount.

[daktari]

Det kan være verdt å se denne videoen fra New York City fra MobilEye. Det er oppunder 40 minutter uten intervensjoner downtown. Ifølge dem bare kamera.

Håper en av beta-folkene til Tesla prøver samme ruta, men i bykjøring går det foreløpig noe rundt  10 min mellom hver gang FSD må manuelt håndteres. På amerikanske "ring 2" er fsd bedre.


https://youtu.be/50NPqEla0CQ

[PKEllefsen]

Har gått nok så sakte opp til nå ja, men er villig til å vedde en god del på at det vil gå raskere herfra og utover. Tar litt tid når man starter på noe helt nytt å forstå det man jobber med. Så langt har det vært fokus på å finne riktig fremgangsmetode noe som har ført til nytt hardware, rewrites, dropping av radar og nå snart Dojo. Mulig enda en rewrite er nødvendig om de finner ut at det er for mye rusk og bugs i systemet etter å ha fjernet radar + å kutte ned fra forskjellige stacks på forskjellige lokasjoner til å nå kun bruke en stack på alle steder. Forhåpentligvis vil dette gå fint slik at det ikke forekommer ytterligere forsinkelser på den fronten. Tipper personlig at HW4 vil bli nødvendig for selvkjørende biler da Greentheonly på twitter sa at de allerede hadde cappet kapasiteten til HW3 i januar og at de måtte kombinere prosessorkraft fra begge chippene istedenfor å kunne bruke dem separat som var først tenkt. Hadde ikke vært overraskende om HW3 måtte byttes ut med tanke på at det er et 1.gen produkt fra Tesla. Finner alltid ut mye gode ideer i etterkant som gjør at andre generasjon av produktet er det som blir gjeldende. Ikke rart at Elon trodde HW3 hadde nok kraft heller da med tanke på hvor mye bedre det var enn HW2.5, men han gleder seg nok til HW4 han også nå tenker jeg.

[Vaduz]

Jeg tror det er helt omvendt jeg, utviklingen gikk raskt i starten fordi systemet bare var en filholder for motorvei og cruise control, men kompleksiteten øker eksponensielt desto mer bilene skal klare selv. Tror det er det samme med AI generelt, masse hype, lite som skjedde i starten, men litt etter litt vil det krype på og om 50 år er nok verden ganske annerledes. Ville ikke basert bilkjøp på dette her. Måtte le litt da noen skriver om «snart HW4 og ny AI supercomputer så da skjer det snart!» déjà vu all over again som man sier;)

[mthoring]

Jeg er også systemutvikler og har null tro på at utviklingen går raskere og raskere. Det er heller omvendt, som andre har sagt, 80-20.
Ekte L5 kommer, men ikke i år eller neste år. Å tippe innen 2030 er uhyre optimistisk, men en smule mer realistisk. Utbredt selvkjøring innenfor avgrensede, forutsigbare, områder vil komme først.

[Counterpointer]

Veldig fristene å si fra oss voksne at barna får bare holde på med sitt, de skjønner nok hvor vanskelig det er etterhvert [emoji6]

[turfsurf]

Det har kanskje noe med at flere her har hatt Tesla med FSD siden de lanserte det, og har sett hvor sakte det går.
...

Er det ikke det som er selve definisjonen på gamlisinnstilling, da

Vet ikke jeg, om det så er står jeg for det 

[turfsurf]

Det er nok ikke bare avvikene som har tatt tid. Dårlig planlegging, og for høye ambisjoner helt fra starten må nok også ta en stor del av skylden. Husker ikke om det er 3 eller 4 "rewrites" som er gjennomført siden 2016.

I tillegg kommer til tider stor gjennomtrekk i utviklingsavdelingen, og en ustyrlig daglig leder med et vell av løfter med urealistiske tidshorisonter som sannsynligvis er tatt med minimal kommunikasjon, har nok også gjort sitt.

Akkurat rewrites har jeg ikke så mye i mot. Finner man systematiske feil så er det noe som må til.

Gjennomtrekk i utviklingsavdelingen er ett stort problem, men dessverre noe alle de store sliter med, og mye av grunnen til at store software løsninger gjør mange rare hopp i livsløpet.

[Motorwatt]

Veldig fristene å si fra oss voksne at barna får bare holde på med sitt, de skjønner nok hvor vanskelig det er etterhvert [emoji6]

EM om FSD: I didn't expect it to be so hard 

[turfsurf]

Ekte L5 kommer, men ikke i år eller neste år. Å tippe innen 2030 er uhyre optimistisk, men en smule mer realistisk. Utbredt selvkjøring innenfor avgrensede, forutsigbare, områder vil komme først.

Akkurat det siste kan komme rimelig fort, kanskje en ti års periode. Gjerne som en regulering fra EU. Blir større krav til veimerking, kanskje vi kan få bedre merkede veier i Norge og. Landevei med midstripe vil så følge.

[Motorwatt]

Ser ut som FSD Beta V9 klarer seg utmerket uten veimerking.

[Counterpointer]

Veldig fristene å si fra oss voksne at barna får bare holde på med sitt, de skjønner nok hvor vanskelig det er etterhvert [emoji6]

EM om FSD: I didn't expect it to be so hard 

Elon er på mange måter en stor guttunge [emoji23]

[daktari]

https://motherfrunker.ca/fsd/

Full copy below for archival
September 2014
They will be a factor of 10 safer than a person [at the wheel] in a six-year time frame
Source
December 2015
We're going to end up with complete autonomy, and I think we will have complete autonomy in approximately two years.
Source
January 2016
In ~2 years, summon should work anywhere connected by land & not blocked by borders, eg you're in LA and the car is in NY
Source
June 2016
I really consider autonomous driving a solved problem, I think we are less than two years away from complete autonomy, safer than humans, but regulations should take at least another year
Source
October 2016
By the end of next year, said Musk, Tesla would demonstrate a fully autonomous drive from, say, a home in L.A., to Times Square ... without the need for a single touch, including the charging.
Source
January 2017
At what point will Full Self-Driving Capability features noticeably depart from? - Elon: 3 months maybe, 6 months definitely
Source
March 2017
I think that [you will be able to fall asleep in a Tesla] is about two years
Source
May 2017
Update on the coast to coast autopilot demo? - Still on for end of year. Just software limited. Any Tesla car with HW2 (all cars built since Oct last year) will be able to do this.
Source
March 2018
I think probably by end of next year [end of 2019] self-driving will encompass essentially all modes of driving and be at least 100% to 200% safer than a person.
Source
November 2018
Probably technically be able to [self deliver Teslas to customers doors] in about a year then its up to the regulators
Source
January 2019
We need to be at 99.9999..% We need to be extremely reliable. When do we think it is safe for FSD, probably towards the end of this year then its up to the regulators when they will decide to approve that.
Source
February 2019
We will be feature complete full self driving this year. The car will be able to find you in a parking lot, pick you up, take you all the way to your destination without an intervention this year. I'm certain of that. That is not a question mark. It will be essentially safe to fall asleep and wake up at their destination towards the end of next year
Source
April 2019
I think it will require detecting hands on wheel for at least six months.... I think this was all really going to be swept, I mean, the system is improving so much, so fast, that this is going to be a moot point very soon. No, in fact, I think it will become very, very quickly, maybe and towards the end this year, but I say, I'd be shocked if not next year, at the latest that having the person, having human intervene will decrease safety. DECREASE! (in response to human supervision and adding driver monitoring system)
Source
April 2019
We expect to be feature complete in self driving this year, and we expect to be confident enough from our standpoint to say that we think people do not need to touch the wheel and can look out the window sometime probably around the second quarter of next year.
Source
May 2019
We could have gamed an LA/NY Autopilot journey last year, but when we do it this year, everyone with Tesla Full Self-Driving will be able to do it too
Source
April 2020
Robotaxis release/deployment... Functionality still looking good for this year. Regulatory approval is the big unknown
Source
April 2020
we could see robotaxis in operation with the network fleet next year, not in all markets but in some.
Source
July 2020
I am extremely confident that level five or essentially complete autonomy will happen, and I think, will happen very quickly, I think at Tesla, I feel like we are very close to level five autonomy. I think—I remain confident that we will have the basic functionality for level five autonomy complete this year, There are no fundamental challenges remaining. There are many small problems. And then there's the challenge of solving all those small problems and putting the whole system together.
Source
October 2020
FSD beta rollout happening tonight. Will be extremely slow & cautious, as it should.
Source
December 2020
I am extremely confident of achieving full autonomy and releasing it to the Tesla customer base next year. But I think at least some jurisdictions are going to allow full self-driving next year.
Source
December 2020
I'm extremely confident that Tesla will have level five next year, extremely confident, 100%
Source
January 2021
Tesla Full Self-Driving will work at a safety level well above that of the average driver this year, of that I am confident. Can't speak for regulators though.
Source
January 2021
FSD will be capable of Level 5 autonomy by the end of 2021
Source
March 2021
Due to high levels of demand for FSD Beta, adding "Download Beta" button to Service section of car display in ~10 days
Source
March 2021
Build 8.3 of FSD should be done QA testing by end of next week, so that's roughly when download button should show up
Source
March 2021
Next significant release will be in April. Going with pure vision — not even using radar. This is the way to real-world AI.
Source
April 2021
Button timing of May is aspirational. Depends on how well limited beta of V9.0 goes, but I would be surprised if wide beta (aka button) is later than June. FSD subscription next month is a sure thing.
Source
April 2021
Any wisdom on limited v9 fsd beta release ? - Probably two weeks
Source