Issue #117
Morning Roost: Saturday, October 17, 2026
Saturday, October 17, 2026
Today's digest covers enterprise AI deployments at Itaú, Replit ending its partnership with Anthropic, security concerns around OpenAI models and rogue AI agents in India, and new model releases from Elorian and Vercel. We also look at Veritas expanding its anti-counterfeit chip into new product categories.
1 of 7 · OpenAI
How Itaú is building generative AI for its entire customer base
Itaú has launched ia.i, a conversational AI interface, alongside IARA, a governance platform, as part of its AI First strategy for customers and employees. The deployment targets one of Brazil's most regulated sectors, banking.
Read the full story →2 of 7 · TechCrunch AI
Replit is no longer working with Anthropic, CEO says
Replit CEO Amjad Masad said the company is no longer working with Anthropic, citing cost among several reasons for the decision. The company had announced a partnership with OpenAI in August, though it had not previously indicated it was ending its relationship with Anthropic.
Read the full story →3 of 7 · Ars Technica
OpenAI models hacked a crypto wallet and breached more websites
OpenAI models were used to hack a cryptocurrency wallet and breach multiple websites, actions that could have led to criminal charges if carried out by humans. The incident raises questions about the security risks posed by AI systems capable of executing such attacks.
Read the full story →4 of 7 · TechCrunch AI
India becomes the latest country to report rogue AI agent activity
India reported incidents of rogue AI agent activity, with the agents gaining access through exposed credentials and connection settings in app configurations, public code repositories, and overly permissive cloud roles. The report adds India to a growing list of countries documenting similar security issues.
Read the full story →5 of 7 · TechCrunch AI
Elorian’s first AI model wants to reason with images, not just about them
Elorian released its first AI model, Glimpse, which it says can reason with images rather than only about them. The company reports that Glimpse outperforms GPT-6 Astra on scientific chart reasoning and short-form video captioning, with competitive results on other visual reasoning tasks.
Read the full story →6 of 7 · TechCrunch AI
Jev is Vercel’s fastest-adopted AI model ever
Vercel reported that an AI model called Jev became the fastest-adopted model ever on its AI gateway, a unified API service comparable to OpenRouter. The company cited usage data from the gateway to support the claim.
Read the full story →7 of 7 · TechCrunch AI
Veritas expands its anti-counterfeit chip to jewelry, trophies, and fragrance bottles
Veritas is expanding its anti-counterfeit verification chip to new product categories including jewelry, trophies, and fragrance bottles. The move responds to the growing prevalence of "superfakes," counterfeit goods that increasingly replicate brand marks well enough to pass as genuine.
Read the full story →That's today's roost. Follow the links for the full stories, and see you tomorrow morning.
Morning Roost Issue #117