Today's AI brief, summarized in minutes.
Today's 6 highest-signal stories across 2 verticals, curated by DeepSignal.
last refreshed 90 min ago
Codex usage surged over 10x in six months, reaching 7M users, while Claude Code remains at 2M. Prime Intellect's Verifiers v1 enhances RL efficiency with O(n) trace growth, marking a significant shift in agentic workflows.
On July 13, Jumpshare launched the STEPX brand, introducing the world's first agentic phone, STEPX Neo, and the Step AOS operating system, designed for AI-native interactions. This ecosystem aims to redefine human-machine collaboration, featuring low-latency local inference and a new safety framework for intelligent agents.
The recent launch of Jumpshare's STEPX brand, featuring the STEPX Neo, the world's first agentic phone, and the Step AOS operating system, marks a significant advancement in AI-native interactions and human-machine collaboration, as detailed in 雷峰网 AI (article). Concurrently, Uber is diversifying its offerings beyond ride-hailing by introducing hotel bookings and concierge services while developing AV Labs for autonomous vehicle data collection, as reported by TechCrunch (article). These developments highlight a trend towards integrating advanced AI technologies and enhancing service ecosystems, suggesting that builders and investors should focus on innovative applications of AI in consumer services and mobility solutions.
The AI landscape is witnessing significant advancements with Codex usage skyrocketing over 10 times in just six months, reaching 7 million users, while Claude Code remains at 2 million, indicating a shift in user preference towards Codex as noted in AINews. Additionally, the launch of Step AOS, the world's first agent-native operating system, enhances the capabilities of personal AI agents like Amoo, facilitating seamless task execution and interaction, as highlighted by 雷峰网 AI in their coverage of the release here. Furthermore, Vercel AI has improved user experience by introducing a Subagents tab in Agent Runs for eve projects, allowing detailed monitoring of subagent activities, which can be seen in their latest update here. What this means for builders/investors is that there is a growing demand for advanced AI tools and infrastructure that support more complex agent interactions and monitoring capabilities.
![[AINews] Codex usage up >10x in 6 months to 7M users, +1M in the past ~day; did Codex overtake Claude Code??](https://substackcdn.com/image/fetch/$s_!cqvt!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09c078c3-d47d-4ab1-91e5-b09ad5d082dd_1388x902.png)
Codex usage surged over 10x in six months, reaching 7M users, while Claude Code remains at 2M. Prime Intellect's Verifiers v1 enhances RL efficiency with O(n) trace growth, marking a significant shift in agentic workflows.
The surge in Codex usage to 7M users, significantly outpacing Claude Code, indicates a strong market preference for Codex's capabilities, which could influence builders to prioritize integration with Codex for development tools. Additionally, the introduction of Prime Intellect's Verifiers v1 suggests advancements in reinforcement learning, potentially enhancing the efficiency of AI workflows for product managers and investors looking for scalable solutions.


On July 13, Jumpshare launched the STEPX brand, introducing the world's first agentic phone, STEPX Neo, and the Step AOS operating system, designed for AI-native interactions. This ecosystem aims to redefine human-machine collaboration, featuring low-latency local inference and a new safety framework for intelligent agents.
Jumpshare's launch of the STEPX brand and the STEPX Neo phone represents a significant advancement in AI-native technology, enabling low-latency local inference for smarter interactions. This development signals new opportunities for builders and PMs to create applications that leverage real-time AI capabilities, while investors should consider the potential market disruption from intelligent agent devices.

Uber is expanding beyond ride-hailing and delivery, introducing hotel bookings and concierge features, while developing AV Labs for autonomous vehicle data collection. The company aims to enhance its financial services for drivers and merchants, but remains cautious about becoming an all-encompassing app like Asian super-apps.
Uber's expansion into hotel bookings and concierge services signals a strategic diversification that builders and PMs should note as it opens new avenues for user engagement and monetization. For investors, this development indicates Uber's intent to strengthen its ecosystem while avoiding the pitfalls of overextension, potentially leading to more sustainable growth.

Tech entrepreneurs like Tom Blomfield and Chamath Palihapitiya are re-engaging with AI, driven by the fear of missing out on its transformative potential. Blomfield joins Anthropic's compute team, while Palihapitiya takes on a CEO role at 8090 Labs, reflecting a broader trend of established figures diving back into hands-on technical roles.
The re-engagement of successful tech entrepreneurs like Tom Blomfield with AI roles, such as his position at Anthropic, signals a renewed focus on AI innovation and development. This trend indicates to builders, PMs, and investors that established leaders are prioritizing hands-on involvement in AI, which could lead to new opportunities and advancements in the field.

Step AOS, the world's first agent-native operating system, was launched by Jieyue Star on July 13, 2023, in Shanghai. It features a unique memory structure and decision-making capabilities, enabling seamless interaction and task execution for personal AI agents like Amoo, built on a robust model matrix that excels in benchmark tests.
The launch of Step AOS, the world's first agent-native operating system, signifies a major advancement in AI infrastructure, enabling more efficient and effective personal AI agents. Builders and PMs can leverage its unique memory structure and decision-making capabilities to enhance user experiences, while investors should consider its potential to drive innovation in AI applications and services.

Vercel AI has introduced a new Subagents tab in Agent Runs for eve projects, allowing users to monitor subagent activities. This feature displays each subagent's prompt, duration, and failures on a shared timeline, providing detailed insights similar to parent runs, including tool calls and token usage.
Vercel AI's introduction of the Subagents tab in Agent Runs allows builders and PMs to gain detailed insights into subagent activities, including prompt performance and failures. This transparency can enhance debugging and optimization processes, making it easier for teams to refine AI workflows and improve overall project efficiency.