On October 19, 2026, Elon Musk posted, lamenting the experience of renting AI/SI computing power, and attached a video without further explanation. There are currently no subsequent developments or official responses, and the status of the event remains at the posting of that post itself.
Matched fieldsSummary
On October 18, 2026, blogger AYi published content comparing the 3D modeling outputs of DeepSwarm V4.1 Flash and Claude Opus 5.5 using the same prompt. The content includes a comparison video, and the author questions in the post whether the two are at the same level in terms of lighting and floating island details. The available information does not provide further test conditions or quantitative results, nor does it offer any follow-up developments.
Matched fieldsSummary
PixVerse promotes the in-app F1 Chaos gameplay. On October 17, 2026, the official PixVerse account posted a chase-themed demo video and prompted users to search for "F1 Chaos" within PixVerse to create the same effect. Currently, this gameplay is showcased to the public through the official account's demo video, and users can search for and use it in the app as instructed.
Matched fieldsSummary
On October 17, 2026, during a conversation with Lex Fridman in Taiwan, Jensen Huang expressed opposition to computing power export controls, advocating a "flywheel" logic in place of blockade. He argued that blockades make sense for physical goods such as rare earths and bullets and bombs, but that the more widely the United States sells computing technology, the safer it becomes: sales generate wealth, wealth is invested in more R&D, and this cycle keeps the United States ahead, because the safest countries in the world tend to be the wealthiest. The full video was published on Lex Fridman's YouTube channel. This claim is currently reported by a single source, with no subsequent developments of fact.
On October 17, 2026, during a conversation with Lex Fridman in Taiwan, Jensen Huang expressed opposition to computing power export controls, advocating a "flywheel" logic in place of blockade. He argued that blockades make sense for physical goods such as rare earths and bullets and bombs, but that the more widely the United States sells computing technology, the safer it becomes: sales generate wealth, wealth is invested in more R&D, and this cycle keeps the United States ahead, because the saf
Matched fieldsSummary
On October 16, 2026, Thomas Wolf posted that "it is time to invent Mathematics 2.0," along with a video. In the post, he quoted a post from SAIR, which announced the launch of Open Math OS, describing it as "the infrastructure for Mathematics 2.0," offering open models and AI agents, 24/7 cloud computing power and 100+ tools, with the community controlling its own data and approved teams able to use computing power for free. At present, this claim is supported only by the above post and the quoted content as evidence, with no subsequent developments.
On October 16, 2026, Thomas Wolf posted that "it is time to invent Mathematics 2.0," along with a video. In the post, he quoted a post from SAIR, which announced the launch of Open Math OS, describing it as "the infrastructure for Mathematics 2.0," offering open models and AI agents, 24/7 cloud computing power and 100+ tools, with the community controlling its own data and approved teams able to use computing power for free. At present, this claim is supported only by the above post and the quot
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Tencent has launched the "Lüpaopao" short drama mini-program within the WeChat ecosystem, without creating a standalone app. It focuses on "free good dramas, watch with one tap," and features a "Friends are watching" social recommendation section on its homepage. Its core performance metric is average playback time per user. According to a TMTPost·AGI report on October 18, 2026, during the gray-scale testing period, the AI-simulated-actor short drama "Film Queen Yaya" surpassed hundreds of millions of views within 3 days of launch, with peak concurrent online viewers exceeding 2 million and advertising revenue reaching several million yuan. The report considers short dramas to be a scenario for technical validation and commercial monetization of Tencent's Hunyuan video generation model. There is currently no subsequent development supported by evidence.
Tencent has launched the "Lüpaopao" short drama mini-program within the WeChat ecosystem, without creating a standalone app. It focuses on "free good dramas, watch with one tap," and features a "Friends are watching" social recommendation section on its homepage. Its core performance metric is average playback time per user. According to a TMTPost·AGI report on October 18, 2026, during the gray-scale testing period, the AI-simulated-actor short drama "Film Queen Yaya" surpassed hundreds of milli
Matched fieldsSummary
Matched fieldsSummary
She Zongming proposed a judgment, arguing that comparisons of the uses of AI computing power between China and the US such as "US AI boosts productivity, Chinese AI creates bread and circuses" are a false premise. A Beijing News commentary article questioned this comparison: the claim that "55% of China's Tokens are used for AI short dramas" comes from a listed company heavily invested in AI short dramas, while "58% of the US's is used for code" has no authoritative source, and Token usage does not equal where computing power is directed. The article also pointed out that video generation is a high-Token-consumption-density scenario: a 10-minute high-definition AI short drama requires 80 million to 120 million Tokens, while an R&D team of 8 to 10 people uses about 3 million to 5 million Tokens per week; as of March this year, domestic daily average Token calls had exceeded 140 trillion. At present, this judgment is carried by the above-mentioned commentary article, and there are no subsequent developments.
She Zongming proposed a judgment, arguing that comparisons of the uses of AI computing power between China and the US such as "US AI boosts productivity, Chinese AI creates bread and circuses" are a false premise. A Beijing News commentary article questioned this comparison: the claim that "55% of China's Tokens are used for AI short dramas" comes from a listed company heavily invested in AI short dramas, while "58% of the US's is used for code" has no authoritative source, and Token usage does
Matched fieldsSummary
Higgsfield AI announced the launch of Higgsfield Academy. A report published on October 17, 2026, said the Academy is described as the only AI course needed to start creating in 2026, covering everything from writing your first prompt to directing your own AI film, with no experience required, and is now live on Higgsfield. Another report the same day said Higgsfield Academy has officially launched, promoted with a video titled "handing Higgsfield to a monkey." There are currently no subsequent developments.
Higgsfield AI announced the launch of Higgsfield Academy. A report published on October 17, 2026, said the Academy is described as the only AI course needed to start creating in 2026, covering everything from writing your first prompt to directing your own AI film, with no experience required, and is now live on Higgsfield. Another report the same day said Higgsfield Academy has officially launched, promoted with a video titled "handing Higgsfield to a monkey." There are currently no subsequent
Matched fieldsSummary
On October 17, 2026, Elorian released its first AI model, Glimpse. According to Testing Catalog, Glimpse can, during its thinking process, return to a specific point in a video, zoom in on details, or replay half a second of video at a higher frame rate before answering. Elorian's own testing shows that Glimpse outperforms GPT-6 Astra in reading scientific charts and describing short videos; in a simulated task of planning for a robotic arm to place cream cheese into a bowl, Glimpse cost $0.035, compared with $0.249 for GPT-6 Astra; in a demonstration, Glimpse solved 5 mazes of increasing difficulty, while Gemini 3.8 Flash solved 1/5. Another report the same day said that Elorian positions Glimpse as the first visual reasoning model, used to compare scene changes at different points in a video, and calls this frame-based reasoning visual thinking; during its reasoning process it can return to the source video, crop regions, extract specified frames, or run code to measure and compare what it sees, a mechanism called visual scratch space; co-founder Andrew M. Dai said it has frontier performance competitive with Astra on key public visual reasoning benchmarks. At present, there have been no further evidence-supported developments since the model's release.
On October 17, 2026, Elorian released its first AI model, Glimpse. According to Testing Catalog, Glimpse can, during its thinking process, return to a specific point in a video, zoom in on details, or replay half a second of video at a higher frame rate before answering. Elorian's own testing shows that Glimpse outperforms GPT-6 Astra in reading scientific charts and describing short videos; in a simulated task of planning for a robotic arm to place cream cheese into a bowl, Glimpse cost $0.035,
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary
Matched fieldsSummary