AI News for September 1, 2026
Browse any day on a calendar, install it to your home screen, read offline.Read in the app →-
1TechCrunch23:15
Nvidia’s $3.5B MediaTek bet reveals its plan for tackling Big Tech's AI chip buildout
Nvidia is investing $3.5 billion in Taiwanese chipmaker MediaTek, deepening a partnership that spans custom silicon and system design. The move comes as Google, Amazon and Meta expand their in-house accelerator programs, and signals how Nvidia intends to remain embedded in AI infrastructure even where it no longer supplies the main compute die.
Read original ↗ -
2Ars Technica21:41
ChatGPT and Reddit now face EU's toughest online safety rules
ChatGPT has crossed the EU user threshold that brings it under the Digital Services Act's rules for very large online platforms and search engines, alongside Reddit. OpenAI will be required to assess and mitigate systemic risks including harm to minors, mental health effects and the spread of illegal content, and to submit to independent audits. The Verge also reported on the designation.
Read original ↗ -
3TechCrunch04:13
The Pentagon now has its own version of ChatGPT and Grok
Government-tailored versions of OpenAI's ChatGPT and xAI's Grok are being added to the Pentagon's central AI portal, where Google's Gemini is already available. The rollout gives defense staff access to multiple frontier models through one gateway and marks a further step in the US military's procurement of commercial AI.
Read original ↗ -
4Ars Technica02:10
“Zlibrary my beloved”: Anthropic staff chats extolling piracy cited in Sony suit
Sony's music publishing suit against Anthropic quotes internal staff chats praising shadow libraries such as Z-Library, using them to argue the company knowingly trained on pirated material. The complaint links that sourcing to lost income for songwriters as AI-generated tracks climb the charts, and adds to a growing set of copyright cases over training data provenance.
Read original ↗ -
5VentureBeat11:26
OpenClaw 2.0 is here, ushering in the era of 'multiplayer' AI coding: What it means for enterprises
Peter Steinberger and OpenClaw's co-developers released version 2.0 of the open-source agent harness, which turns language models into autonomous workers reachable through Telegram, iMessage, WhatsApp and Discord. The update centers on shared, multi-user sessions, raising practical questions for enterprises about access control and auditability of agent actions.
Read original ↗ -
6TechCrunch08:13
Apple shares 'shocking evidence' against former employee accused of stealing company data for OpenAI
Apple filed evidence that a former employee accused of taking confidential data to OpenAI destroyed material after learning he was under investigation. The case is one of several recent disputes over engineers moving between Apple and AI labs, and turns on what counts as protected internal work.
Read original ↗ -
7The Verge21:00
Instagram cracks down on AI accounts pretending to be human
Instagram will limit the reach of AI-generated profiles that do not disclose what they are, and is renaming its "AI creator" label to "AI-generated profile" for clarity. The change follows user complaints about AI influencer accounts that have become harder to distinguish from real people. TechCrunch also covered the policy update.
Read original ↗ -
8The Verge23:34
Debian won't ban AI code from its Linux distribution
Debian developers voted to permit AI tools in development, maintenance and documentation work on the distribution. The resulting policy treats generative AI as neither exempt from nor subject to rules beyond Debian's existing contribution standards, leaving licensing and quality obligations with the human contributor.
Read original ↗ -
9TechCrunch00:00
Clipto uses AI to search terabytes of video and is now valued at $250M
Clipto, a three-year-old startup that uses AI to make large video archives searchable, raised $15 million at a $250 million valuation. The company says it had already reached $15 million in ARR and was profitable before the round, an unusual profile among media-AI startups.
Read original ↗ -
10TechCrunch02:35
Harvard Law dropout raises $6M for Blue Voice to build a 'Harvey for police officers'
Blue Voice raised $6 million to build a legal assistant for police officers, positioning itself as a Harvey equivalent for law enforcement. The product is trained on department-specific statutes, local ordinances and internal protocols that general-purpose models cannot retrieve from the open web.
Read original ↗ -
11TechCrunch20:56
Meeting note-taker Circleback adds a free tier to attract more customers
Circleback, which generates AI summaries and action items from meetings, introduced a free tier alongside new paid plans starting at $14 per month. The move puts it in more direct competition with note-taking features bundled into Zoom, Google Meet and Microsoft Teams.
Read original ↗ -
12Ars Technica18:00
Pocket's AI made my game ideas real. Now Meta controls the results.
A hands-on look at Pocket, Meta's tool for generating small interactive mobile "gizmos" from prompts, finds the creation process fast and genuinely usable. The catch is distribution: the resulting projects are difficult to export or share outside Meta's own platform, leaving creators dependent on it.
Read original ↗ -
13The Verge22:00
New York Governor Kathy Hochul thinks AI should be ‘less evil’
In an election-year interview, New York Governor Kathy Hochul lays out positions on AI regulation, data center siting and related tech policy fights, including the state's dealings with Meta. The conversation is a useful read on how state-level AI rules may diverge from federal policy.
Read original ↗ -
14VentureBeat02:52
Software engineers' new job isn't writing code — it's designing the boundaries AI agents can't break
An argument that with Cursor, Claude Code and agentic workflows embedded in IDEs and containers, producing a first implementation is no longer the bottleneck in data platform work. The author contends the scarce skill is now designing constraints, interfaces and invariants that agent-generated code cannot violate.
Read original ↗ -
15VentureBeat22:00
Identity and permissions aren’t enough to govern AI agent behavior
A sponsored piece from Box arguing that identity and permission models govern what an enterprise agent can reach but not how it acts once running autonomously. It advocates a layered approach that adds behavioral monitoring and runtime controls on top of access management.
Read original ↗
