Latest / Elon Musk Podcast / Can Tesla Optimus Robot actually perform surgery?

Transcript

Can Tesla Optimus Robot actually perform surgery? · Elon Musk Podcast

• 0:00
  This is the Elon Musk Podcast, your daily hit of what is really
• 0:04
  going on at Tesla, SpaceX, XAI, and the rest of the Musk
• 0:08
  universe. I'm your host Will Walden, and I
• 0:10
  have covered Elon Musk for more than five years, spent a year on
• 0:13
  the ground at SpaceX, Starbase during early Starship
• 0:16
  development. And before this, I spent my
• 0:18
  career as a software developer working with billion dollar
• 0:21
  companies. I've also built and sold my own
• 0:25
  businesses. And now I make content and help
• 0:28
  other people grow their companies.
• 0:29
  And on this show, I use that experience to break down the
• 0:32
  news, filter out all the noise, and give you clear context you
• 0:37
  can actually use. Recently, Elon Musk has laid out
• 0:44
  a concrete goal for the companies Optimist Robot.
• 0:47
  He says Optimist could one day perform surgery with high skill,
• 0:51
  giving every patient access to top tier care.
• 0:55
  And of course, that is if they could pay for it.
• 0:57
  He links that vision to a broader shift to test the toward
• 0:59
  technology that reaches into daily life.
• 1:02
  And he points out that a shortage of skilled surgeons and
• 1:06
  suggest robots can close that gap.
• 1:09
  And as we know, people have been using telemedicine, Tele
• 1:12
  robotics to perform surgeries for a while.
• 1:15
  But if the fingers can be nimble enough in an optimist robot,
• 1:20
  Elon Musk thinks that yes, Elon Musk's optimist robot could
• 1:25
  perform heart surgery on people. But can a general purpose robot
• 1:29
  earn enough trust in the operating room?
• 1:32
  Tesla plans a new Optimist showcase in early of 2026, and
• 1:36
  the company is building the manufacturing line it says will
• 1:38
  scale toward 1,000,000 units a year.
• 1:41
  Now, in this episode, I'll explain what Optimist can do
• 1:44
  today, what Tesla says will change by the next model, and
• 1:47
  what it would take for any robot to approach autonomous clinical
• 1:50
  work. Also, breakdown the regulatory
• 1:52
  wall that stands between a demo and a cleared medical device.
• 1:56
  And we're gonna talk about what this plan signals to investors.
• 2:00
  And we will get right into that after this very, very short
• 2:05
  break. Now, Tesla positions Optimus as
• 2:09
  a pathway from clean energy and cars into everyday robotics.
• 2:12
  The company frames this as part of Master Plan Part 4, an
• 2:16
  expansion from electrification to general technology that
• 2:19
  tackles skilled tasks in the physical world.
• 2:22
  That reset moves the discussion from quarterly vehicle numbers
• 2:26
  to whether a new class of hardware and AI can take over
• 2:28
  fine motor work. Now today's optimist can walk,
• 2:34
  lift light items and handle fragile objects like eggs
• 2:37
  without breaking them. Those are non trivial steps
• 2:40
  because they require balance, object reorientation and grip
• 2:45
  control that does not crush or drop that egg.
• 2:49
  Tesla says the next iteration, expected for early 2026, will
• 2:52
  include a handmade up of far more small components to enable
• 2:57
  finer manipulation. That change, if it ships as
• 3:01
  promised, would aim at the dexterity ceiling that separates
• 3:04
  a warehouse picker from a surgical suture.
• 3:09
  Now, Tesla says it is building manufacturing lines so the robot
• 3:12
  can be produced at scale. The ambition here is up to
• 3:17
  1,000,000 units per year once a full build is underway.
• 3:20
  Mass production matters because volume spreads costs and pushes
• 3:25
  suppliers to standardize parts around the design.
• 3:28
  And volume also invites a software platform approach where
• 3:31
  one hardware base runs different skills for homes, factories, and
• 3:36
  eventually clinics. Now my analysis is that a large
• 3:39
  installed base would let Tesla train and improve manipulation
• 3:43
  skills faster, but it would not by itself grant clinical trust.
• 3:47
  Elon Musk and Tesla have to earn that trust now.
• 3:52
  Tesla's leadership connects optimist to healthcare by naming
• 3:54
  surgery as a future target. The pitch is access.
• 4:00
  Everybody could have a robot surgeon if a robot could perform
• 4:03
  precise procedures reliably. Rural hospitals in undersupplied
• 4:07
  regions could schedule care without waiting for scarce
• 4:11
  specialists. The claim assumes safe autonomy
• 4:15
  or near autonomy and high stakes settings.
• 4:18
  Now my analysis is that even if Optimist progresses quickly, the
• 4:22
  bar for surgical work is not a simple extension of warehouse
• 4:25
  dexterity because surgical tasks couple fine motion with
• 4:30
  millisecond level reactions to living tissue and changing
• 4:35
  vitals. These are real people that the
• 4:37
  Optimist robot is working on. This isn't a test.
• 4:41
  These are people's lives on the line and this robot has to act
• 4:45
  like a human, but better. It has to have the dexterity of
• 4:48
  the top surgeons in the world, but better.
• 4:52
  Now regulation sits at the center of that gap.
• 4:55
  Though most medical robots in use today operate as tools
• 4:59
  guided by human surgeon in the room and they went through
• 5:02
  device specific testing and also those people have been trained
• 5:07
  for decades. A robot that acts with broad
• 5:10
  autonomy would face a higher standard stepwise validation and
• 5:14
  multi year trials that match specific indications now.
• 5:18
  Tesla would need to document safety at each stage, prove
• 5:22
  failure handling for edge cases, and show repeatability across
• 5:25
  patient autonomy, anatomy, and hospital environments.
• 5:30
  Now, I think that even with strong engineering, the review
• 5:34
  timeline alone stretches any aggressive schedule.
• 5:39
  The thing can barely pick up an egg right now.
• 5:42
  It doesn't break an egg. That's great, but can it handle
• 5:45
  a tool like a scalpel? Can it pick up a scalpel and
• 5:50
  manipulate it, roll it in its fingers and also use it to
• 5:55
  precisely cut into a person's body?
• 6:00
  Only then, And that's one small thing that it has to do.
• 6:04
  It has to do so many more things to become a successful surgeon,
• 6:11
  for it to become a successful healthcare provider.
• 6:16
  Now, I'm not counting Tesla and this robot out.
• 6:20
  Optimus is a work of art. It's an amazing piece of
• 6:24
  engineering, but when Elon says, and we all know Elon Musk time,
• 6:27
  right when he says something's going to happen in two years, we
• 6:32
  know not to hold our breath because it's probably not going
• 6:35
  to happen in two years. Just like Tesla's full self
• 6:40
  driving. It really hasn't happened yet.
• 6:42
  Full autonomous driving hasn't happened yet.
• 6:44
  He's been promising this since the beginning of Tesla.
• 6:49
  Also, something like the Super heavy, the Starship super heavy,
• 6:53
  he has said that it should be on the moon by now and of course
• 6:58
  that didn't happen and you know why?
• 7:01
  There is a lot of different reasons.
• 7:02
  One of those reasons is regulatory strangleholds which
• 7:06
  happened at Star base when I was there.
• 7:08
  When I was on the side of the road on Hwy. 4 and I was
• 7:11
  reporting on the Starship, the FAA was put in place to hold up
• 7:20
  Tesla. At that point.
• 7:22
  There was a lot of stuff going on, environmental reviews.
• 7:24
  They had to review that whole environment so SpaceX could
• 7:28
  launch the Starship. And because of that, they were
• 7:30
  behind about nine months. Now, with my time there, I know
• 7:35
  that they do work aggressively, and if the Optimist robot team
• 7:41
  are working just as aggressively as SpaceX was on the Starship, I
• 7:46
  know they're working hard and I know that they want this thing
• 7:49
  to succeed as fast as possible. Now, regulations, now that's
• 7:53
  another thing. There's going to be a lot of red
• 7:55
  tape. Even if Optimist can do certain
• 7:58
  things, how do you get it into an environment and test it and
• 8:05
  make sure that everything is OK with regulations and also with
• 8:08
  the robot and also with the patient.
• 8:10
  It's going to take years, maybe a decade from now before this
• 8:14
  thing actually does real work on a person.
• 8:17
  I'm not saying like stitch somebody or put a Band-Aid on
• 8:21
  somebody. I'm saying like this thing can
• 8:24
  you walk into an operating room and you see a team of optimist
• 8:28
  robots. Each team has a specialty.
• 8:30
  Each person on the team has a specialty and that robot itself
• 8:34
  knows everything about that specialty and can do that job
• 8:38
  better than a human. That's not going to happen
• 8:39
  anytime soon. I, I hope it happens in the next
• 8:42
  20 years because that would be an amazing feat.
• 8:45
  And if these robots are as affordable as Elon says they're
• 8:47
  going to be, he says 20 thousand, $30,000, something
• 8:52
  around that. Then yes, the hospital in rural
• 8:57
  America, the hospital where I grew up in the middle of
• 9:00
  nowhere, in the middle of the forest could have an optimist
• 9:03
  robot. And that optimist robot could
• 9:05
  save hundreds of lives. It's a car crash, people
• 9:10
  bleeding out the optimist robot. You could have a dozen optimist
• 9:14
  robots ready to go at any time. Instead of having a specialist
• 9:19
  come in from a city that's two hours away, instead of having a
• 9:23
  specialist come in from someplace that's far away, then
• 9:26
  specialist is already asleep. 2 hour drive.
• 9:29
  But also you have to wake the specialist up.
• 9:32
  You have to get them ready. That takes 15 or 20 minutes.
• 9:34
  These people are dying in your operating room.
• 9:37
  You could have a dozen Optimist robots ready to take over when
• 9:41
  that accident happened and these people show up at the ER and
• 9:46
  they are ready to be worked on. Now, if that's going to be the
• 9:49
  case in the next 20 years, this is going to revolutionize
• 9:52
  healthcare. This is going to save so many
• 9:54
  people, and the population is going to go up as well.
• 9:58
  So what is the overall outcome of this?
• 10:02
  People will survive longer, more people will be alive for a
• 10:06
  longer amount of time, and also the costs for surgery will go
• 10:12
  down tremendously because surgeons will be getting cut out
• 10:16
  of the whole process. Do you think that's going to
• 10:19
  happen easily? Do you think surgeons are going
• 10:22
  to be replaced by robots overnight?
• 10:25
  I don't think so. Surgeons and also hospital
• 10:29
  workers, you know what? They have Ways and Means to
• 10:35
  fight this. They have unions.
• 10:38
  And this is not going to happen without a huge fight from the
• 10:41
  surgeons. Those people that paid 100,000
• 10:43
  two, $100,000 to get that degree to be the best surgeon, the best
• 10:47
  brain surgeon in their area. Of course they're not going to
• 10:50
  go down without a fight. I mean, they're not going to,
• 10:56
  the hospitals are not going to take these Optimus robots in for
• 11:01
  high end surgeries for a long time anyway.
• 11:04
  And by that time, the little things, the triage, you know,
• 11:09
  getting people into the system and seeing them if they have a
• 11:15
  cut or a scrape, something like that, or stitches, that's cool.
• 11:19
  But if they need them for high end surgery, that's going to be.
• 11:23
  And I don't, I hate to be this person that's always, I'm not
• 11:27
  always like this. I'm never, I don't want things
• 11:29
  to happen far down the line because I want to see them
• 11:33
  happen. I want to see this robot and I
• 11:35
  want to see other robots in this space do surgeries because I
• 11:41
  think it's going to help so many people and I want to help a lot
• 11:44
  of people. So I don't want to be the
• 11:46
  naysayer here, but I know how technology works and I know what
• 11:50
  it's like to get any sort of practical application for
• 11:57
  healthcare from a robot. Even even Apple, when they made
• 12:03
  the, the, the Apple Watch and they started putting health
• 12:08
  gadgets on it, there was a lot of regulatory tape.
• 12:11
  They have to go through all that and Optimist can't even like
• 12:14
  sometimes they can barely walk. It dances around a little bit,
• 12:18
  but we've all seen the Optimist robot stumble around.
• 12:22
  You can't have that in a surgical environment.
• 12:24
  You can't have that at a hospital.
• 12:27
  Now, the next 6 models, 7 models, sure they'll be better.
• 12:32
  And of course the robotics and the AI that is trained on it's
• 12:37
  going to be so much better in the next 3-4 years.
• 12:40
  And we have seen great growth by the optimist team and the robots
• 12:46
  have been getting so much better.
• 12:48
  But doing a doing surgery is going to be a a wild thing for
• 12:56
  this thing. A Tesla says it is still early
• 13:00
  and that work though and moving from general hand to a surgical
• 13:04
  capable system would require sterile field design materials
• 13:08
  that withstand cleaning protocols, biocompatibility,
• 13:11
  sensor redundancy, fail safe controls that degrade gracefully
• 13:16
  if a component slips outside its normal range.
• 13:19
  So if the dexterity in the finger stops working when it's
• 13:23
  to say they have somebody opened up and something happens in the
• 13:26
  dexterity the the hand fails, what happens then?
• 13:30
  Do they send in a real surgeon, a person surgeon to take over
• 13:34
  and they have to have fail safe controls that it just doesn't
• 13:38
  happen You know if a component slips outside its normal range
• 13:42
  it needs to have a backup, needs to have redundancies.
• 13:44
  Hospitals would need training pathways for staff too if this
• 13:47
  happens. Procedures for pre op checks and
• 13:50
  incident reporting automation would be a huge factor for these
• 13:55
  types of robots. And 3rd, you know, those are
• 13:59
  baseline requirements before a pilot even touches a patient,
• 14:03
  before any sort of optimist robot touches a person.
• 14:07
  Now, engineering for hands is the next step.
• 14:14
  Tesla says the next hand will have more small parts to expand
• 14:17
  Fine motor control. A surgical task such as tying a
• 14:21
  knot, placing A suture at a specific depth or handling
• 14:25
  delicate tissue requires fingertip for sensing, tactile
• 14:29
  feedback loops and high frequency control that dampens
• 14:33
  tremor while preserving speed. Cameras and depth sensors must
• 14:39
  resolve soft tissue without losing tracking in fluids, body
• 14:45
  fluids, and so a path exists for this, but only if software
• 14:53
  sensing and actuation progress is in lockstep.
• 14:58
  A stronger hand without robust perception will not deliver even
• 15:02
  one safe stitch. Now Elon and Tesla planned for
• 15:10
  one million robots a year. They are going to make the
• 15:14
  supply chain for actuators, sensors, and the compute plus
• 15:19
  test rigs that validate every joint before shipping these
• 15:22
  things out. It's like making vehicles but
• 15:26
  instead of making vehicles you make robots.
• 15:28
  It's all an assembly line. It also implies field service
• 15:31
  networks, replacement parts, logistics, and over the air
• 15:35
  update discipline that treats a hand controller bug as a safety
• 15:39
  issue and not just a delay. And also when do they do the
• 15:44
  over the air updates? Is it when optimist is docked?
• 15:48
  If there's a recall, when do they recall these?
• 15:53
  And if there's a big enough recall in there are optimist
• 15:57
  robots all over the country, all over the world performing
• 16:00
  surgeries and the recall happens in the the recall is for surgery
• 16:05
  robots immediately. What happens then if Tesla
• 16:09
  reaches volume in non medical environments?
• 16:11
  First, it could bank telemetry to tune manipulation models.
• 16:15
  So what that means is that they ship out a million robots.
• 16:22
  Every one of those robots, just like Tesla's vehicles right now,
• 16:27
  test the environment. They test every part of that
• 16:30
  robot until they get so much data that they'll be able to
• 16:34
  make the medical robots. This data would be valuable but
• 16:40
  would still need domain adaptation before they inform
• 16:44
  clinical moves. You know, I've been digging
• 16:49
  through the analysis of the show.
• 16:50
  Speaking of analysis, and notice that 37% of you are following
• 16:55
  this channel, which I love you, thank you and for you I am
• 16:59
  forever grateful. And the other 63% haven't hit
• 17:02
  the follower subscribe button yet.
• 17:04
  And I've been an independent journalist covering Elon Musk
• 17:07
  Space Flight in tech for the last six years.
• 17:09
  And I'm going to continue for the next 10 years.
• 17:12
  All I ask for you is one second of your time and you just hit
• 17:15
  the follow or subscribe button on the platform you're on, which
• 17:18
  you're watching or listening on right now.
• 17:20
  I'm extremely grateful for you and I'm blessed to have you in
• 17:23
  this community. So thank you for doing that.
• 17:27
  Now they have to be clinically ready as well.
• 17:32
  They have to prove sub millimeter accuracy on bench
• 17:35
  tests with standardized phantoms that mimic human tissue.
• 17:40
  That has to be in the works for years.
• 17:42
  Second, complete supervised procedures in simulation and
• 17:46
  animal models where the robot handles part of a task while the
• 17:49
  human controls the rest. Third, conduct early human
• 17:53
  trials for tightly defined steps like holding and positioning
• 17:57
  instruments while a surgeon, a human surgeon executes the
• 18:01
  critical cut. Now each milestone needs
• 18:06
  quantified failure modes, alarm thresholds, and intervention
• 18:10
  handoffs. Medical robots must earn the
• 18:15
  trust of real surgeons, little by little, every tiny step.
• 18:21
  Now humans will be in this loop for every step.
• 18:24
  They'll remain essential for years, maybe decades.
• 18:27
  Even advanced robots will face edge cases such as unexpected
• 18:31
  bleeding and allergic reaction or tissue differences that could
• 18:36
  confound pre op. Imaging systems must escalate to
• 18:39
  a human and accept immediate override without lag.
• 18:43
  People should be able to take over at any moment.
• 18:46
  Training will focus on teams, not just those machines.
• 18:50
  So anesthesiologist, nurses and surgeons act from a shared
• 18:56
  mental model and what the robot does, what it is sensing and
• 19:01
  when it will pause, we'll all be monitored by people.
• 19:06
  The most credible near term optimist use in a hospital would
• 19:09
  be non invasive logistics, sterile handling, or room
• 19:13
  turnover where error costs very very little and throughput still
• 19:19
  improves care. Now, would you want an optimist
• 19:23
  robot to operate on you or even dress your bandage to stitch you
• 19:28
  up? Now, the promise of universal
• 19:30
  access to top care is a huge thing, but people might not want
• 19:37
  it in a patient education must match the reality of early
• 19:41
  deployments for the trials. Stay transparent about what the
• 19:44
  robot controls and what a human controls.
• 19:47
  That's a good thing. Tesla must tell in all of its
• 19:51
  marketing that humans are still in control for a very long time
• 19:56
  and if there's any incident risks, this is going to set them
• 20:00
  back tremendously. They need steady, transparent
• 20:04
  reporting on task level outcomes because if they don't do that,
• 20:08
  people will not trust this robot and they will not want it to
• 20:12
  operate on them. You know, there will be some
• 20:15
  early adopters and that's great. We need those people.
• 20:19
  We did those people that really want a robot to be their surgeon
• 20:24
  and not a trained professional who's been going through school
• 20:26
  for 10 years. So I don't know the the Tesla
• 20:31
  road map for 2026 will be an early proof point hopefully.
• 20:37
  Do they do a great demo? A finer manipulation, and it
• 20:41
  would show whether the new hand delivers controlled, repeatable
• 20:45
  grasps that match tasks like threading, knot tying, or even
• 20:49
  instrument passing. Now, if Tesla pairs that demo
• 20:54
  with documentation, if they show off the robot, that's great.
• 21:00
  But if they show how they did it, they would be even better
• 21:03
  because people could see behind the scenes.
• 21:06
  And if the demo stays at the level of stage tasks without
• 21:10
  metrics, clinicians might not even look at this thing.
• 21:15
  They'll keep treating surgical ambitions of this robot as an
• 21:19
  open question. Now.
• 21:21
  Specifically, this will be the difference maker next year, and
• 21:27
  it's going to be the difference between hype, the Elon hype as
• 21:31
  usual, or people actually thinking this robot is going to
• 21:36
  be a viable healthcare professional in the future.
• 21:42
  Hey, thank you so much for listening today.
• 21:44
  I really do appreciate your support.
• 21:46
  If you could take a second and hit the subscribe or the follow
• 21:48
  button on whatever podcast platform that you're listening
• 21:51
  on right now, I greatly appreciate it.
• 21:53
  It helps out the show tremendously and you'll never
• 21:56
  miss an episode. And each episode is about 10
• 21:59
  minutes or less to get you caught up quickly.
• 22:02
  And please, if you want to support the show even more, go
• 22:05
  to patreon.com/stagezero and please take care of yourselves
• 22:10
  and each other and I'll see you tomorrow.