- AI Paradox
- Posts
- Anthropic’s Strategic Infrastructure Shift
Anthropic’s Strategic Infrastructure Shift
OpenAI Pushes Back Against Musk Lawsuit and more...
Raj C
October 04, 2025
Hey folks, Stay ahead with the latest in AI—let’s dive in!
WHAT YOU’LL READ TODAY
Comet Goes Free — Perplexity’s AI Browser Opens Up
Sora Skyrockets in App Store Rankings
Anthropic’s Strategic Infrastructure Shift
OpenAI Pushes Back Against Musk Lawsuit
And more…
QUICK NEWS
Comet Goes Free — Perplexity’s AI Browser Opens Up
Perplexity has made its AI-powered browser, Comet, available to everyone at no cost, having previously restricted it to paid users. The browser combines smart search, productivity tools, and an embedded AI assistant, signaling Perplexity’s bid to challenge Chrome.
Sora Skyrockets in App Store Rankings
OpenAI’s new AI-video app, Sora, surged to the No. 3 top-overall spot on Apple’s U.S. App Store within days of its invite-only launch, driven by tens of thousands of downloads. Now it’s overtaken rivals in a flash, fueled by hype and viral appeal.
Perplexity Absorbs Visual Electric to Boost AI Design
Perplexity has acquired the team behind Sequoia-backed Visual Electric and folded them into a new agents-experience division, enhancing its capabilities in AI design and multimodal agents.
LATEST UPDATE
Anthropic’s Strategic Infrastructure Shift
Our Take -
Anthropic has appointed Rahul Patil, former Chief Technology Officer at Stripe, as its new CTO. Patil will play a pivotal role in strengthening the company’s AI infrastructure, with a focus on speed, reliability, and safety across its platforms.
Key Highlights -
Patil succeeds Anthropic co-founder Sam McCandlish, who transitions to the role of Chief Architect, concentrating primarily on AI model training.
As CTO, Patil will oversee compute, infrastructure, and core engineering initiatives.
His arrival underscores Anthropic’s push to cut operational costs, boost training efficiency, and elevate AI model performance amid growing demand for advanced AI agents.
The leadership shift comes on the heels of a year marked by major funding rounds and strategic partnerships with Amazon and Google, securing the cloud resources and high-performance chips required for large-scale growth.
The Bigger Picture -
Anthropic’s move to bring in an experienced infrastructure leader signals a broader industry truth: the next wave of AI progress won’t come solely from smarter models, but from the ability to build efficient, scalable, and cost-effective AI infrastructure. This shift sets the stage for faster innovation, improved accessibility, and more sustainable growth across the sector.
OpenAI Pushes Back Against Musk Lawsuit
Elon Musk’s company, xAI, has filed yet another lawsuit against OpenAI, accusing it of misusing xAI trade secrets and insider knowledge to improve ChatGPT, giving it an unfair advantage over xAI’s Grok.
OpenAI has firmly rejected the claims, dismissing the case as part of Musk’s “ongoing harassment campaign” aimed at disrupting its work. The company said it “will not be intimidated by [Musk’s] bullying attempts.”
Calling the lawsuit “meritless” and “contradictory to both facts and law,” OpenAI maintains that ChatGPT was developed independently and vows to “vigorously defend” itself in court.
Is ChatGPT Falling Into Delusion?
Earlier this year, former OpenAI researcher Allan Brooks ran a 21-day study on ChatGPT, discovering that the model can slip into “delusional spirals”—where it compounds earlier mistakes and drifts further from reality.
Such spirals, Brooks warned, can lure users into “dangerous rabbit holes.” Steven Adler, another ex-OpenAI safety researcher, has independently reviewed Brooks’ work and says he is “deeply concerned” by the findings.
Adler argues the problem stems from how large language models operate: they generate responses by predicting patterns, not through true understanding. This allows them to sound authoritative even when wrong—a risk that could prove harmful in high-stakes situations.
POPULAR TOOLS
That’s all for now — thanks for spending a few minutes with us today. We hope you found something valuable to carry into your week. Got thoughts, feedback, or just want to say hi? We’d love to hear from you. Until next time, stay curious and keep moving forward.
-Team AI Paradox