Figure shows off real autonomy: CEO denies the robot was teleoperated

Figure livestreamed its humanoid robot sorting packages in real time. Millions watched on YouTube and X. CEO Brett Adcock rejects criticism that the robot was teleoperated, saying it ran entirely on AI with no operator intervention. The stream drew a wave of skepticism, with critics suggesting an operator with a controller was working behind the scenes. Adcock gave Bloomberg Tech an interview in which he laid out his case: full autonomy, a live feed without suspicious lag, and an architecture built on GPT and computer vision.

AI-processed from Bloomberg Tech; edited by Hamidun News
Figure shows off real autonomy: CEO denies the robot was teleoperated
Source: Bloomberg Tech. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceBloomberg Tech ↗
Editor-in-Chief
Key metric #1Full autonomy — the robot uses real-time vision and advanced AI to decide how to move objects
Key metric #2A live broadcast without suspicious delay — the livestream had no noticeable lag that would point to remote control
Key metric #3An architecture based on GPT and computer vision — the robot sees the objects on the table, understands their shape and size, then makes the decision on its own

Figure ran a mass livestream in which its humanoid robot sorted packages in real time. The video drew millions of views on YouTube and X, but it also met a wave of skepticism: critics suggested an operator with a remote control was working behind the scenes.

Millions watched the livestream at once

The event was huge for the AI community. In the video, the Figure humanoid performed a simple but impressive task: it picks packages up from a table, analyzes their dimensions, sorts them into categories and places them in different containers. All of this happened in real time. Millions of people watched simultaneously on YouTube and X. For Figure, which is fighting for recognition in the industry, it was a big win — direct proof that its robot can handle complex logistics.

Why the skepticism arose

Almost immediately, a heated debate broke out on social media, Reddit and technical forums. Skeptics argued that the video could be a deepfake, or that the humanoid was being controlled by an operator over the internet — so-called teleoperation. The question was a fair one: the industry is full of cases where startups and large companies pass off remotely controlled robots as autonomous. Tesla, Boston Dynamics and other companies have been caught more than once using hidden remote control.

"We showed our robot's real capabilities, no teleoperation, no fraud," says

Figure CEO Brett Adcock.

How the CEO answered the criticism

In an interview with Bloomberg Tech, Brett Adcock gave a clear answer to the accusations. His main arguments:

  • Full autonomy — the robot uses real-time vision and advanced AI to decide how to move objects
  • A live broadcast without suspicious delay — the livestream had no noticeable lag that would point to remote control
  • An architecture based on GPT and computer vision — the robot sees the objects on the table, understands their shape and size, then makes the decision on its own
  • Honesty as a strategy — hiding the truth would be a disaster for Figure's reputation

Adcock stressed one key point: to earn trust, a technology has to be shown as it really is. Spoiler: if teleoperation had been involved, millions of viewers would have noticed — engineers in the chat would quickly have spotted latency artifacts.

What this means for the market

Figure is trying to turn around the narrative in the humanoid robot industry. After the high-profile failures and inflated promises of other companies, any demonstration of real, honest autonomy is a big plus for trust. If Figure has truly solved the teleoperation problem and shown a working AI robot, this could be a turning point for the entire market. Investors and partners would finally be able to tell real progress from marketing.

Story

Key Questions Answered

What are the key details on Millions watched the livestream at once?

The event was huge for the AI community. In the video, the Figure humanoid performed a simple but impressive task: it picks packages up from a table, analyzes their dimensions, sorts them into categories and places them in different containers. All of this happened in real time. Millions of people watched simultaneously on YouTube and X. For Figure, which is fighting for recognition in the industry, it was a big win — direct proof that its robot can handle complex logistics.

Why the skepticism arose?

Almost immediately, a heated debate broke out on social media, Reddit and technical forums. Skeptics argued that the video could be a deepfake, or that the humanoid was being controlled by an operator over the internet — so-called teleoperation. The question was a fair one: the industry is full of cases where startups and large companies pass off remotely controlled robots as autonomous. Tesla, Boston Dynamics and other companies have been caught more than once using hidden remote control. "We showed our robot's real capabilities, no teleoperation, no fraud," says Figure CE

How the CEO answered the criticism?

In an interview with Bloomberg Tech, Brett Adcock gave a clear answer to the accusations. His main arguments: Full autonomy — the robot uses real-time vision and advanced AI to decide how to move objects A live broadcast without suspicious delay — the livestream had no noticeable lag that would point to remote control An architecture based on GPT and computer vision — the robot sees the objects on the table, understands their shape and size, then makes the decision on its own Honesty as a strategy — hiding the truth would be a disaster for Figure's reputation Adcock stresse

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…