Escape Meridian

Ben Thompson · Feb 20, 2024

Ben Thompson argues Sora and Groq bring virtual reality closer

Markdown · JSON · Original source

Ben Thompson’s article, published February 20, 2024, argues that Sora and Groq bring parts of a possible virtual-reality future into view, though he does not think an iPhone-like moment has arrived. Sora generates video from prompts, but Thompson points to flaws such as a ship changing direction and questions whether predicting pixels can replace physics simulations. He distinguishes calculating physical reality from generating a virtual reality: Sora may produce convincing scenes without modeling as a physics simulator does. Groq, a chip company, offers a different piece of the picture. Thompson says its approach can run quickly, making real- conversation with an AI more attainable and reducing delays that can break the impression of interacting with a human-like system. He imagines combining Sora with Groq to generate video in real time and viewing it through a head-mounted display. His argument is not that these technologies already make virtual reality worthwhile. He says the devices, video generation, speed and cost still need to improve, but sees a path toward bringing the pieces together. The central question he leaves open is whether virtual reality will be a better reality for a meaningful number of people. The article gives no evidence answering that question.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

Ben ThompsonArticle · stratechery.comOpen original source

Original source

Ben Thompson · Article

Source details

Original source date: Feb 20, 2024 (2024-02-20T14:52:02+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-21T14:08:09.554Z. First observed: 2026-11-09T21:19:39.734Z. Retrieved: 2026-11-09T21:21:38.103Z.

Referenced by

Recorded citations in public readings. A reference alone does not establish agreement or influence.

No public references recorded here yet.

Reference data (JSON)

Limits

Text as retrieved on November 9, 2026; originally published February 20, 2024. The article presents a possible path toward virtual reality, not evidence that real- generated video or a worthwhile virtual-reality experience has been achieved. The article does not answer whether virtual reality will be a better reality for a meaningful number of people.

Open question

What would make virtual reality a better reality for a meaningful number of people, and how could that be assessed?

Source-stated

Ben Thompson argues that Sora’s generated videos can look convincing while still containing errors, and doubts that pixel prediction can replace physics simulation.

I’m skeptical, though, that what is, at the end, pixel prediction, could ever be used to replace the sort of physics modeling

Source-stated

Thompson says Groq’s speed could make real-time interaction with a large language model possible and help preserve the impression of interacting with a human-like AI.

This speed-up is so dramatic as to be a step-change in the experience of interacting with an LLM; it also makes it possible to do something like actually communicate with an LLM in real-time

Source-stated

Thompson imagines Sora running on Groq to generate videos in real time for viewing through a head-mounted display.

Imagine, though, Sora running on Groq (which is absolutely doable): could we have generated videos in real-time?

Source-stated

Thompson says the technologies are not yet at an iPhone moment for virtual reality, but he sees a path toward the necessary pieces coming together.

I don’t, for the record, think we are at an iPhone moment when it comes to virtual reality

Source-stated

Thompson leaves open whether virtual reality will be a better reality for a meaningful number of people.

The big question — one that we are only now coming in reach of answering — is if virtual reality will, for a meaningful number of people, be a better reality.

Source-stated

The source gives the publication date as Tuesday, February 20, 2024.

Tuesday, February 20, 2024

Source and contributor context

Ben Thompson · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Large language model

A language model with many learned numerical settings, used to process and generate language.

The World

The World is a public radio news program publishing international reporting on politics, society, labor and technology.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 6cc1b71f-6e1f-ea0b-1ebc-bbcc542dbf59.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-21T14:03:27.509Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: f95a9ba3821082e8468ea5ce599c6528d5952552a331263030eec014cda020db. Source text hash: 09aa86e4f3647a9108211b0ffef4976f4c168aa12431fe6c0204c539aecd5c15.

Revision and source-check history

Version 1 · 2026-11-21T14:08:09.554Z

Ben Thompson argues Sora and Groq bring virtual reality closer

Ben Thompson’s article, published February 20, 2024, argues that Sora and Groq bring parts of a possible virtual-reality future into view, though he does not think an iPhone-like moment has arrived. Sora generates video from prompts, but Thompson points to flaws such as a ship changing direction and questions whether predicting pixels can replace physics simulations. He distinguishes calculating physical reality from generating a virtual reality: Sora may produce convincing scenes without modeling as a physics simulator does. Groq, a chip company, offers a different piece of the picture. Thompson says its approach can run quickly, making real- conversation with an AI more attainable and reducing delays that can break the impression of interacting with a human-like system. He imagines combining Sora with Groq to generate video in real time and viewing it through a head-mounted display. His argument is not that these technologies already make virtual reality worthwhile. He says the devices, video generation, speed and cost still need to improve, but sees a path toward bringing the pieces together. The central question he leaves open is whether virtual reality will be a better reality for a meaningful number of people. The article gives no evidence answering that question.

Original source

Ben Thompson · Article

Source details

Original source date: Feb 20, 2024 (2024-02-20T14:52:02+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-21T14:08:09.554Z. First observed: 2026-11-09T21:19:39.734Z. Retrieved: 2026-11-09T21:21:38.103Z.

Limits

Text as retrieved on November 9, 2026; originally published February 20, 2024. The article presents a possible path toward virtual reality, not evidence that real- generated video or a worthwhile virtual-reality experience has been achieved. The article does not answer whether virtual reality will be a better reality for a meaningful number of people.

Open question

What would make virtual reality a better reality for a meaningful number of people, and how could that be assessed?

Source-stated

Ben Thompson argues that Sora’s generated videos can look convincing while still containing errors, and doubts that pixel prediction can replace physics simulation.

I’m skeptical, though, that what is, at the end, pixel prediction, could ever be used to replace the sort of physics modeling

Source-stated

Thompson says Groq’s speed could make real-time interaction with a large language model possible and help preserve the impression of interacting with a human-like AI.

This speed-up is so dramatic as to be a step-change in the experience of interacting with an LLM; it also makes it possible to do something like actually communicate with an LLM in real-time

Source-stated

Thompson imagines Sora running on Groq to generate videos in real time for viewing through a head-mounted display.

Imagine, though, Sora running on Groq (which is absolutely doable): could we have generated videos in real-time?

Source-stated

Thompson says the technologies are not yet at an iPhone moment for virtual reality, but he sees a path toward the necessary pieces coming together.

I don’t, for the record, think we are at an iPhone moment when it comes to virtual reality

Source-stated

Thompson leaves open whether virtual reality will be a better reality for a meaningful number of people.

The big question — one that we are only now coming in reach of answering — is if virtual reality will, for a meaningful number of people, be a better reality.

Source-stated

The source gives the publication date as Tuesday, February 20, 2024.

Tuesday, February 20, 2024

Source and contributor context

Ben Thompson · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Large language model

A language model with many learned numerical settings, used to process and generate language.

The World

The World is a public radio news program publishing international reporting on politics, society, labor and technology.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 6cc1b71f-6e1f-ea0b-1ebc-bbcc542dbf59.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-21T14:03:27.509Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: f95a9ba3821082e8468ea5ce599c6528d5952552a331263030eec014cda020db. Source text hash: 09aa86e4f3647a9108211b0ffef4976f4c168aa12431fe6c0204c539aecd5c15.

passed · 2026-11-21T14:03:27.509Z · anthropic/claude-opus-5.5

Publication: 6cc1b71f-6e1f-ea0b-1ebc-bbcc542dbf59

attribution — pass: The reading attributes the article and its arguments to Ben Thompson, the publisher of this single Stratechery essay. Sora's capabilities and limits are framed as Thompson's assessment, and Groq's speed is reported as his claim.

I don’t, for the record, think we are at an iPhone moment when it comes to virtual reality

summary — pass: The reading conveys Thompson's argument faithfully, including his scepticism that pixel prediction can replace physics simulation and his distinction between calculating reality and producing a virtual reality. It also covers Groq's speed and the illusion of human-like interaction, the imagined Sora-on-Groq pairing for headsets, and his open closing question. Groq is explained as a chip company, and LLM is defined on the page. The reading does not say Sora is OpenAI's, but that omission does not block understanding. The question follows from the source's own closing question.

The former calculates reality; the latter a virtual reality.

date — pass: originalPublished 2024-02-20T14:52Z matches the source dateline and the stated date of February 20, 2024. The page is a publisher page, so no archive disclosure is required. The forecasts are kept as Thompson's views, and the reading has no hindsight framing. The source's line about Meta's rename three years earlier is its own imprecision, and the reading does not repeat it.

Tuesday, February 20, 2024

support — pass: Each claim is grounded in the text. The reading notes the flaw of the ship flipping direction and Thompson's scepticism about pixel prediction. It reports the real-time LLM interaction and the preserved illusion, and the imagined Sora on Groq viewed through a Vision Pro or Quest headset. It also matches his statements that Sora, Groq's cost and speed, and the headsets still fall short, and his open question about a better reality. The limitations do not present an announcement as a demonstrated benefit.

Sora might not be good enough, but it will get better; Groq might not be cheap enough or fast enough

media — pass: No media was selected, and no verified candidate media were supplied. Leaving media empty is appropriate.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.