Ai_Mobility_News: "Tesla's Secret Optimus Training Strategy Just Revealed!". Facts on this page cite the primary sources below, not the video.
The claim: Tesla has a secret army of people in camera suits teaching Optimus around the clock.
The proof: Tesla's own job ads describe a real, multi-site operation, but a more ordinary one. As of Nov. 9, 2026, 31 Optimus data-collection postings were open in 15 places, according to RobotTraining.jobs, which reads Tesla's careers site every day. Where Tesla states pay, hourly rates run from $22.72 to $48.30 an hour. The job: wear motion-capture equipment, handle objects, perform set movements and walk set routes so Optimus can learn from the recordings. The "40 cameras" and "three shifts around the clock" details come from Robert Scoble's Oct. 28 post and have not been verified.
Editorial Thesis
Optimus learns from people before it learns on its own. Tesla's job ads are the clearest public record of that: real roles, real pay, real physical demands. That is more concrete than most Optimus headlines. But hiring people to record motion is an input, not an output. The ads tell you Tesla is building a training-data pipeline. They do not tell you how well the robot uses that data.
Honest Verdict
- Confirmed (Tesla postings via RobotTraining.jobs, read Nov. 9): 31 open postings in 15 places across three roles: Data Collection Operator, Data Collection Operations Lead and Data Collection Supervisor, all labelled Optimus.
- Confirmed (stated pay, 12 of 31 postings): Operators $22.72 to $34.50 an hour; Leads up to $48.30 an hour; Supervisors $92,000 to $138,000 a year. Tesla mostly states pay where state law requires it, so the missing postings are not lower pay.
- Confirmed (postings): walking 7 or more hours a day while carrying up to 30 lbs of equipment; a height between 5'7" and 6' or 6'3" in most postings, because the work uses motion-capture equipment; Fremont runs 12-hour day or night shifts.
- New this week (job-board copies of Tesla listings): a Data Collection Operations Lead in Littleton, Colorado dated Nov. 8, and an Austin Technical Program Manager listed Nov. 6 to "own the creation, scaling, and maintenance of SI datasets that train Tesla's Optimus."
- Unverified: Scoble's Oct. 28 description of 24-pound suits with up to 40 cameras and three shifts around the clock in Palo Alto. Humanoids Daily found the listings support a broader collection effort, not those specifics.
- Not shown: how many people do this work today, how much data Tesla has, or any measured improvement in Optimus from it.
What the jobs involve
The operator role is physical. People wear motion-capture gear while they pick things up and move in set ways, walk planned routes at Tesla sites and in public, start and stop the recorders, upload the data and file a daily report. Most postings ask for a height range because the suits have to fit and the recorded body has to be close to the robot's size.
Leads run a shift of collectors and check that the work follows the protocol. Supervisors plan collection across locations, approve timecards and audit the data. The newest Austin listing, a program manager role, is about the datasets themselves: quality, quantity, variety and fixing errors fast. It describes them as "SI datasets," the same "SI" (superintelligence) wording Musk now uses for SpaceX SI.
Why this matters for Optimus
Tesla's cars learn from millions of drivers. Optimus has no fleet of customers yet, so Tesla has to make its own training data. Musk said on the Q3 call that Tesla needs "on the order of 10 or 20,000" robots practising in what it calls Optimus Academy. These job ads show the human side of that same problem: before robots can practise, people have to show them what good movement looks like.
What this does not prove
- It does not prove the 40-camera suits or around-the-clock Palo Alto shifts.
- It does not prove the data is making Optimus better at open-ended tasks. Musk said in October the robot's AI still needs work.
- It does not give a headcount. Open postings are not the size of the team.
- It does not change any date. Outside deliveries are still targeted for around the end of 2027.
What to watch next
- Whether the count of open postings keeps rising, and in which cities.
- The unscripted Optimus task video Musk said could come "in a month or two."
- Any Tesla disclosure of Academy robot counts or training-data volume.
Bottom Line
Tesla is openly hiring people at $22.72 to $48.30 an hour to wear motion-capture gear and teach Optimus how to move. That is a real, checkable fact, and more useful than most Optimus hype. It is proof of an effort, not proof of a capable robot.
Source notes
Tesla's careers search returned HTTP 403 to our server, so listings were read through RobotTraining.jobs (which reads Tesla's careers site daily) and job-board copies of Tesla postings dated Nov. 6 to 8. Scoble's claims via Humanoids Daily (Oct. 30). Counts change daily; re-check on publish day.
Sources
- RobotTraining.jobs (read Nov 9 2026; counts read from Tesla's careers site daily): 31 open postings in 15 places across 3 jobs; $22.72-$48.30/hr where Tesla states pay (12 of 31); Operator $22.72-$34.50/hr; Lead $23.99-$48.30/hr; Supervisor $92,000-$138,000/yr; duties, 7+ hours walking, up to 30 lbs, height 5'7"-6' or 6'3"
- Job-board copy of Tesla's 'Data Collection Operations Lead, Optimus' in Littleton, Colorado, dated Nov 08 2026: Support I-IV bands $22.72-$43.47/hr; Sunday-Thursday 8:00 AM-4:30 PM; applications accepted until Dec 31 2026
- Job-board copy of Tesla's 'Technical Program Manager, Data Collection, Optimus' (Austin, listed Nov 6 2026): 'own the creation, scaling, and maintenance of SI datasets that train Tesla's Optimus'
- Humanoids Daily (Oct 30 2026): Scoble's Oct 28 claims of 24-pound suits with up to 40 cameras and three shifts are unverified; Tesla listings confirm motion-capture work, overnight schedules in Austin and dual shifts at Fremont Factory and Milmont Academy
- Tesla careers search (primary; returned HTTP 403 to our server on Nov 9, so listings were read via the mirrors above)