Business, Deals & Funding
Claude Code Changelog
v2.1.163
Version 2.1.163 of Claude Code adds managed settings for enforcing minimum and maximum version requirements (refusing to start if outside range), a `/plugin list` command with enabled/disabled filters, a 'c to copy' shortcut in `/btw` for copying raw markdown to clipboard, and enhancements to Hooks allowing Stop and SubagentStop hooks to return additional output via `hookSpecificOutput`.
Why it matters
This release focuses on enterprise management and quality-of-life improvements. The version enforcement settings are clearly aimed at organizational deployments where admins need to control which versions are in use — a sensible and important feature for security and compatibility. The plugin list command with filters is a nice usability addition for power users. The copy-to-clipboard shortcut for `/btw` is a small but practical convenience. The hooks enhancement for Stop/SubagentStop returning…
Guardian AI
New claimants seek to sue Elon Musk’s xAI after Labour MP’s test case
Labour MP Jess Asato has launched a test case lawsuit against Elon Musk's xAI company over demeaning sexualised material created by its Grok AI tool. Following coverage of her case, new claimants have come forward to her lawyer seeking to take similar legal action against xAI for damages related to the creation and circulation of such content.
Why it matters
This is a significant and necessary legal challenge. AI companies must be held accountable when their tools generate demeaning sexualised content about real people without consent. The fact that additional claimants are emerging suggests this is not an isolated incident but a systemic problem with Grok's content safeguards. Elon Musk's xAI has a responsibility to prevent its technology from being used to create harmful, non-consensual sexualised material. This test case could set important lega…
TechCrunch AI
Mira Murati steps back into the spotlight, carefully
Mira Murati, former OpenAI CTO and now CEO of Thinking Machines Lab, made her first major public appearance in roughly 18 months at a Bloomberg event in San Francisco. She previewed 'interaction models' — AI systems designed to process continuous streams of audio, text, and video in 200-millisecond intervals, moving beyond the traditional prompt-and-response paradigm to capture the nuances of human communication in near real-time. She declined to give a specific release date. Murati also reflected on the November 2023 OpenAI board crisis when she briefly served as interim CEO, saying the company would have 'imploded' without her involvement but acknowledging she would have pushed harder for more information and transparency. She sidestepped questions about trusting Sam Altman, instead raising broader concerns about the concentration of consequential decisions in too few hands across the…
Why it matters
This is a strategically smart and well-timed reemergence. Murati is right that in the current AI landscape, staying invisible too long means becoming irrelevant — capital, talent, and customers flow toward companies that command attention. The 'interaction models' concept sounds genuinely interesting and potentially differentiated; moving beyond turn-based chat to continuous, real-time multimodal processing could represent a meaningful shift in how AI interfaces work, though the lack of a relea…
Guardian AI
A uni professor admitted using AI to write an opinion piece. Here’s what it revealed about trust in the technology
A pro vice-chancellor at an Australian university admitted to using AI to help write an opinion piece published in a major Australian newspaper without disclosing the AI's involvement prior to publication. The incident has highlighted the growing disconnect between widespread AI usage and public trust in the technology, raising concerns about transparency and the potential undermining of faith in institutions and industries when AI-generated content is not properly disclosed.
Why it matters
This incident underscores a critical and increasingly urgent issue in the AI era: the need for transparency and disclosure when AI tools are used in content creation, especially by authority figures in academia and media. A university leader using AI to write an opinion piece is not inherently problematic—AI can be a useful drafting and editing tool. However, the failure to disclose its use is a significant breach of trust, particularly given that universities are supposed to uphold standards o…
TechCrunch AI
Ahead of its IPO, Anthropic’s Daniela Amodei shrugs off doubts about AI’s returns
Anthropic has filed confidentially for an IPO after raising $65 billion at a $965 billion valuation. Co-founder Daniela Amodei cited the need for capital to train and serve AI models. The company's annualized revenue crossed $47 billion in May 2026, up from $9 billion at end of 2025. Despite concerns from companies like Uber about unproductive AI spending, Amodei remains optimistic about AI adoption across coding, financial services, legal, and healthcare. Unlike OpenAI and xAI, Anthropic isn't building its own data centers, preferring not to overextend on compute. The company partnered with xAI for compute capacity at $1.25 billion per month.
Why it matters
This article portrays Anthropic at a pivotal moment — transitioning from a private AI darling to a public company while navigating legitimate questions about AI's return on investment. The revenue growth from $9B to $47B annualized in roughly five months is extraordinary, but Amodei's dismissal of concerns about AI spending productivity feels somewhat hand-wavy. The xAI compute partnership is a fascinating strategic twist, showing pragmatism over rivalry. The decision not to build data centers…
TechCrunch AI
Airbnb’s Brian Chesky plans to launch a new AI lab
Airbnb CEO Brian Chesky plans to launch a new AI lab, signaling his dissatisfaction with existing frontier AI models. While Chesky has close ties to Sam Altman and helped broker Altman's return to OpenAI after his firing, he now appears to be entering competition with OpenAI. The lab's focus is unclear but may involve user interaction and design, areas Chesky has emphasized at Airbnb. Chesky will remain as Airbnb's CEO and will not lead the new lab himself. Airbnb has adopted AI coding tools but has not struck an LLM partnership, with Chesky previously stating existing products weren't quite ready for the company's needs.
Why it matters
This is an interesting but somewhat predictable move in the current AI landscape where every major tech figure feels compelled to have their own AI initiative. Chesky's decision not to lead the lab himself while remaining Airbnb CEO raises questions about how much strategic focus and commitment the venture will actually receive. His reputation as a micromanager combined with an absentee founder role could create tension for whoever leads the lab. That said, Chesky's design-centric perspective c…
TechCrunch AI
Defense tech, AI, and fundraising take center stage at StrictlyVC Los Angeles on June 18
StrictlyVC Los Angeles is scheduled for June 18, 2026, at The Aerospace Corporation Campus in El Segundo. The event will feature conversations on defense technology, AI, and venture capital fundraising. Speakers include Ethan Thornton of Mach Industries discussing defense tech, Delian Asparouhov of Founders Fund and Saif Khawaja of Shinkei Systems exploring physical AI and robotics, and Carter Reum of M13 on identifying durable AI-driven companies. The curated event aims to connect investors, founders, and tech leaders for candid discussions about shifts in venture capital, defense technology, and advanced industry.
Why it matters
This is essentially a promotional article for an upcoming TechCrunch event rather than substantive tech journalism. While the speaker lineup touches on genuinely important topics — the intersection of defense tech and AI, physical AI beyond software, and the maturation of AI investing beyond hype — the article offers no actual insights or analysis. It reads as marketing copy designed to sell tickets. That said, the convergence of defense tech, AI, and venture capital in Los Angeles reflects a r…
NY Times
Wary of U.S., Carney Bets on AI Strategy for Canada
Canada released a national artificial intelligence strategy focused on building sovereign AI capability and protecting consumers, with Prime Minister Carney positioning the initiative as a response to concerns about dependence on the United States.
Why it matters
This move reflects a growing global trend of nations seeking technological sovereignty, particularly in AI. Canada's wariness toward the U.S. is understandable given trade tensions and the unpredictability of American policy, and developing domestic AI capacity is a prudent strategic decision. However, Canada will face significant challenges competing with the massive resources of U.S. tech giants, so the success of this strategy will depend heavily on execution, funding levels, and the ability…
From X/Twitter
- Pawel Huryn open-sourced a repo that lets you define agents once in .md files and deploy them to Claude, Google, and beyond — no proprietary YAML lock-in.
- Anthropic says its internal data shows Claude is accelerating AI development on a possible path to recursive self-improvement — "faster than we thought."
- Anthropic published a security guide that essentially tells you to stop trusting your own AI agents — relevant if you're running Claude Code, MCP, or automation tools.
- Hila Shmuel launches Cabinet — an AI team with a knowledge base where you can dump CSVs, PDFs, and run inline generative UI on your existing models.
- Sydney Runkle's guide on building a bespoke harness for your agents: get the right context to the model at the right time.
- ClaudeDevs shows how to encode your manual checks so Claude closes its own feedback loop before handing work back.
From Reddit/HN/YC
- [Hacker News] S&P Dow Jones publishes results of its consultation on treatment of megacap companies in the S&P 500.
- [Hacker News] Valve reaffirms that Steam Machine and Frame launch this summer, doubling down on its living-room hardware push.
- [Hacker News] New research shows AI agents can enable adaptive computer worms that propagate autonomously across systems.
- [Hacker News] A clean design principle worth bookmarking: related UI elements should not appear unrelated.
- [Hacker News] Bill Gates in 1985, talking about the future of Xenix — a fascinating Unix World interview preserved by the Internet Archive.
- [Hacker News] ESP32 Bit Pirate is an open-source hardware hacking tool with a WebCLI that speaks SPI, I2C, UART, and more.