Anthropic just eclipsed OpenAI
Anthropic has surpassed OpenAI with the release of Claude Opus 4.8, which leads on nearly all major benchmarks, paired with a $65B funding round valuing the company at $965B. The newsletter also covers Apple's AI Siri overhaul built on Google Gemini, Cursor's developer productivity report showing concentrated AI gains, and a Codex tutorial for autonomous game building.
Summary
The newsletter's headline story centers on Anthropic overtaking OpenAI on two fronts simultaneously: model performance and market valuation. Claude Opus 4.8 outperforms GPT-5.5 and Gemini 3.1 Pro on benchmarks including agentic coding, computer use, financial analysis, and Humanity's Last Exam. The model is described as Anthropic's least lazy and most honest, with a tendency to flag uncertainties rather than fabricate answers. It comes with a Fast mode that is 3x cheaper, new effort controls on claude.ai, and parallel sub-agent support in Claude Code. Alongside the model launch, Anthropic announced a $65B funding raise, pushing its valuation to $965B and surpassing OpenAI. The company also teased an upcoming 'Mythos-class' AI model.
Apple's long-delayed AI strategy is reportedly taking shape, with Bloomberg revealing a revamped Siri rebuilt on Google Gemini and housed within Dynamic Island. The new Siri will support AI-powered web search similar to Perplexity, a dedicated ChatGPT-style chat interface, third-party AI agent routing, and AI features integrated into the Camera app. The update is seen as critical given Apple's failure to deliver promised AI features in 2024 and the broader competitive pressure from OpenAI and Google.
Cursor's Developer Habits Report reveals that AI tools have more than doubled developer output, with lines of code per developer per week rising from 3.6K to 8.6K over 18 months. However, gains are highly concentrated — the top 1% of developers produce 46x more code than the median user. Agent tool calls are up 30% in two months, and 5x more AI-generated changes are reaching commits without manual review. Cost per agent request varies 9x across models, highlighting inefficiencies in how teams deploy AI.
A tutorial section walks readers through using OpenAI's Codex '/goal' feature to autonomously build a browser game, emphasizing the importance of writing a scoped, testable game description before engaging the agent. Additional news items include an AI consultant's client accidentally spending $500M in one month on Claude licenses, CNN suing Perplexity for allegedly reproducing paywalled articles verbatim, and Elon Musk clarifying that SpaceX's compute deal with Anthropic is 180 days rather than the three years cited in an S-1 filing.
About this episode
PLUS: Use Codex to build a functional game in one prompt
Key Insights
- Anthropic's safety-first positioning, which Sam Altman has called 'fear-based marketing,' appears to be paying off commercially — the company now has a higher valuation than OpenAI and leads on nearly every major benchmark.
- Cursor's data shows that AI productivity gains are highly unequal: the top 1% of developers produce 46x more code than the median active user, and the gap is widening every month rather than narrowing.
- The newsletter reports that an AI consultant's client accidentally spent nearly $500M in a single month after failing to set usage limits on employee Claude licenses, illustrating the financial risk of ungoverned enterprise AI deployments.
- Apple's revamped Siri is reportedly built on Google Gemini rather than Apple's own models, representing a significant strategic dependency on a competitor and a departure from Apple's typical vertical integration approach.
- Elon Musk publicly contradicted Anthropic's S-1 filing by stating SpaceX's compute deal with Anthropic is 180 days, not three years, while leaving open the possibility of a longer arrangement — introducing uncertainty around a key infrastructure partnership ahead of a potential IPO.
Topics
Transcript
Good morning, {{ first_name | AI enthusiasts }}. Anthropic just took the lead from OpenAI — not just with a more powerful frontier model, but in the market too. With the new Opus 4.8 crushing on benchmarks, a more powerful Mythos on the way, and a valuation closing in on $1T, Dario and co finally seem to have all the pieces in place for a blockbuster public listing. Anthropic’s Opus 4.8, nearly $1T valuation Apple’s new AI Siri to take on ChatGPT Use Codex to build a game in one prompt AI doubles dev output, but not for everyone 4 new AI tools, community workflows, and more ANTHROPIC The Rundown: Anthropic made two big announcements on the same day —…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The Rundown AI
Grok 4.6 storms the AI frontier
SpaceXAI's Grok 4.6 has emerged as a competitive frontier AI model, scoring 61 on benchmarks and undercutting rivals on price at 60% less cost, with Grok 4.7 teased for release in 3-4 weeks. The newsletter covers AI trends including practical guidance on using AI effectively, a personalized cancer vaccine startup for dogs built with AI tools, and updates on competing models and AI products.
Anthropic slips an invisible signature into Claude
Anthropic has announced the implementation of invisible watermarks in outputs from its Claude models to enhance transparency, aligning with the EU AI Act. The move has sparked mixed reactions, as stakeholders navigate the implications for open and private AI models.
Meta returns to its open-source roots
Meta is returning to its open-source roots with the launch of the Muse Glimmer model and plans for Muse Spark 1.2, positioning itself as a competitive force against China's AI dominance. OpenAI has also expanded its security program with a new hacking-tuned model, indicating a growing emphasis on cybersecurity in AI deployments.
OpenAI puts the safety brakes on Astra
OpenAI's Astra model, rumored to be GPT-6, is being treated as a potential 'critical' cyber risk following its success in solving significant problems, prompting increased safeguards and a delay in its rollout. This decision reflects broader concerns about AI's evolving capabilities and associated security risks as evidenced by recent breaches in the AI community.
AI designs viruses never seen in nature
Researchers used AI to design 16 novel viruses that infect only bacteria, successfully creating them in the lab to combat drug-resistant infections. While the work demonstrates beneficial applications, experts warn that the same AI tools could be repurposed to create dangerous pathogens, prompting calls for biosafety guardrails.