AI news daily: what changed on October 21, 2026

1. Judge dismisses Sam Altman's sister's sexual abuse lawsuit

Judge dismisses Sam Altman's sister's sexual abuse lawsuit was reported by The Verge AI. See the original report for details.

Source: The Verge AI

2. Anthropic pitches AI safety bill that may block a model if 50+ people could be killed

Anthropic proposed an AI safety policy that would allow federal agencies to block a model's release if it could cause 50 or more deaths. The policy aims to prevent AI systems from causing harm but includes a provision requiring judicial approval before blocking any model. Federal agencies would not be permitted to halt a model's deployment without a court's authorization. The proposal outlines specific criteria for evaluating risks but shifts decision-making authority to the judiciary. The bill's language focuses on balancing safety concerns with legal oversight. No details about implementation timelines or additional safeguards were provided. The policy's effectiveness depends on judicial interpretation of AI risk thresholds.

Source: TechCrunch AI

3. Hugging Face CEO warns that Netflix's OpenAI hack documentary isn't accurate

Hugging Face co-founder Clem Delangue criticized the Netflix documentary "AI Gone Rogue," stating it presents an inaccurate portrayal of AI risks. The film, which features Delangue, has been described by him as adopting "full doomer mode," a term implying excessive pessimism. He emphasized the documentary does not reflect accurate information about AI development. The summary highlights his concern over the film's depiction of AI challenges. No specific details about the documentary's content or Hugging Face's stance on AI risks were provided. Delangue's remarks focus on the film's credibility rather than technical aspects of AI. The warning urges viewers to approach the documentary's claims with caution.

Source: TechCrunch AI

4. Google's more personalized Gemini AI is heading to your car

Google is introducing Personal Intelligence to Gemini AI for use in vehicles. This update enables Gemini in Android Auto and cars with Google built-in to access data from Google apps like Gmail, Calendar, Maps, Photos, YouTube, and Search. The feature allows the AI to use information across these services to provide personalized responses. For example, users can ask about flight status while driving. The update focuses on integrating user data from multiple apps to enhance in-car interactions. Specific release details or technical specifications were not provided. The change aims to improve how Gemini assists drivers with context-aware information. No further steps or timelines were mentioned in the announcement.

Source: The Verge AI

5. Anthropic's cloud partnerships are getting messy

Anthropic is facing challenges in securing cloud computing resources as providers increasingly view it as a competitor. The company's infrastructure spending is rising ahead of its planned November IPO. Cloud partners are reportedly hesitant to support Anthropic due to competitive concerns. No specific numbers or models are mentioned in the report. The situation highlights tensions between Anthropic and major cloud providers. The outcome could impact its readiness for the IPO. The article does not detail further steps or resolutions.

Source: TechCrunch AI

6. SpaceX's AI business is beating expectations

SpaceX reported a net income of $4.4 billion in the third quarter, exceeding expectations. The company's AI initiatives were highlighted as a key contributor to this profitability. No specific AI products or models were detailed in the report. The results mark the first time SpaceX achieved profitability during the quarter. The company did not outline further steps or additional financial metrics related to its AI efforts. The focus remained on overall financial performance rather than technical advancements. No pricing or performance benchmarks for AI systems were mentioned. The summary emphasizes revenue growth over specific AI capabilities.

Source: TechCrunch AI

7. OpenAI acquires startup Orchestra Bio to beef up its biotech business

OpenAI has acquired Orchestra Bio, a startup, to enhance its biotech initiatives. The company will integrate Orchestra's team and technology into its Rosalind AI model for life sciences research. Rosalind Workbench, a tool designed to assist researchers in planning experiments and analyzing results, will also receive these resources. The acquisition aims to advance OpenAI's efforts in developing AI applications for biological studies. No specific timelines or technical details about Rosalind's capabilities were provided. OpenAI did not disclose financial terms of the deal. The move follows previous reports of OpenAI expanding into biotechnology sectors.

Source: TechCrunch AI

8. Google has joined Amazon in attacking Nvidia's stronghold with AI app developers

Google and Amazon have introduced a developer tool aimed at simplifying the migration of PyTorch applications from Nvidia GPUs. The tool is designed to help users transition their AI workloads to other hardware platforms. Both companies are targeting Nvidia's dominance in the AI chip market by offering alternatives. The release highlights growing competition in the AI infrastructure space. Details about specific performance metrics or adoption rates were not provided. The initiative reflects broader industry efforts to diversify GPU dependencies. No timeline for wider availability was mentioned. The tool's primary focus is on improving compatibility and flexibility for PyTorch developers.

Source: TechCrunch AI

9. ChatGPT now lets you lock your chats

ChatGPT now lets you lock your chats was reported by TechCrunch AI. See the original report for details.

Source: TechCrunch AI

10. Meta pauses Muse sharing after researchers find AI chats on the web

Meta paused a feature called Muse that allows users to share AI assistant conversations after researchers discovered URLs for these interactions were publicly accessible online. The security firm Claroc identified over 6,800 Muse links through public DNS records, according to a report. The issue arose because shared conversation links were not properly restricted, exposing private data. Meta has not specified when the feature will be restored. The company emphasized it is investigating the vulnerability. Claroc's findings highlight risks associated with AI chat sharing tools. The incident underscores challenges in securing user-generated content in AI platforms.

Source: The Verge AI