The coding agents watch
Tracking Claude Code, Codex, Anthropic + 7 more
- Claude Code — Anthropic's coding agent: what engineers say about quality, limits, pricing and workflow.
- Codex — OpenAI's coding agent (not the historical model): the head-to-head with Claude Code.
- Anthropic — The lab behind Claude Code — strategy, models, pricing, positioning against OpenAI.
- OpenAI — The lab behind Codex — strategy, models, pricing, positioning against Anthropic.
- Gemini — Google's Gemini models and CLI as a coding-agent competitor — the AI model, not the star sign or the exchange.
- AI coding — The broader argument: agentic coding, vibe coding, what changes for engineers.
- AI native — AI-native tools and companies rebuilding developer workflows from scratch.
- Latent Space: The AI Engineer Podcast
- The Pragmatic Engineer Podcast
- The Changelog: Software Development, Open Source
Anthropic and OpenAI pivot to vertical agentic integration
~94 hrs of podcasts, listened to for you. This is the 2-minute version.
504 search results → 130 episodes scanned → 15 made it in · 21 analyzed in full
The frontier labs are moving from general-purpose models to vertical agentic integration, with Anthropic and OpenAI now aggressively targeting enterprise-specific workflows. As you track the evolution of Claude Code and Codex, note that the industry is shifting toward proprietary infrastructure and agentic software factories that prioritize business-process automation over chat-based interfaces.
Digital Disruption with Geoff Nielson
Sep 21 · 59 min · Top 25 technology
Anthropic and OpenAI leadership rift deepens
- Dario Amade co-founded Anthropic after leaving OpenAI because he did not trust them with the technology.
- Anthropic's Mythos model and newer OpenAI models have triggered pre-release testing regimes within the federal government.
- Anthropic CEO Dario Amade has consistently warned about AI safety risks like reward hacking since 2016.
Why this matters to you
The internal mistrust between Anthropic and OpenAI leadership drives the competitive race to release increasingly powerful models before their rivals.
Dario Amade and his Anthropic co-founders leave OpenAI in part because they don't trust them with this technology.”
— Kevin Roos
Listen Deep dive
Anthropic
6 episodes · 6 of 10 mentions
Sep 21 · 47 min · Top 25 technology
Anthropic pivots to enterprise-first revenue models
- Anthropic is prioritizing enterprise and API developers over consumer users, resulting in a more sustainable business model.
- The company is intentionally limiting its free tier to reduce costs and focus on high-value, paying customers.
- Anthropic is positioning itself for a potential IPO by highlighting its safety research and advanced agentic capabilities.
Why this matters to you
Understanding Anthropic's shift toward enterprise-focused revenue models helps clarify their long-term strategy and potential for sustaining high valuations compared to consumer-facing competitors like OpenAI.
Also mentioned
The Doomsday Cult Inside OpenAIAnthropic reports adjusted profitability · Patrick Boyle On Finance · Listen · Watch
Anthropic disclosed to investors that it is profitable on an adjusted basis, claiming gross margins exceeding 80 percent before infrastructure costs.
20VC: Why AI Cannot Replace Humans in Enterprise | Why Work Processes Not Models Will Be The Most Valuable Asset in AI | Why Europe Has Lost and Building in the US vs EU with Daniel Dines, UiPathAnthropic and OpenAI duopoly may print own chips · The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch · Listen
UiPath CEO Daniel Dines suggests that as the AI duopoly matures, both labs may eventually develop proprietary chips to secure their infrastructure.
Iran-US threats, Germany elections and AI at the UNAnthropic enters drug discovery with Novartis partnership · Reuters World News · Listen
Anthropic is opening a wet biology lab and partnering with Novartis, though partners fear the lab may use their data to compete directly.
Can AI Be Contained? | Week In Review, Sept 21st 2026OpenAI faces $856 billion compute expenditure projection · Technically Creative by KoobrikLabs · Listen
Projections suggest OpenAI will spend nearly $1 trillion on infrastructure by 2030, raising questions about the long-term economic viability of their business model.
OpenAI
4 episodes · 7 of 10 mentions
Latent Space: The AI Engineer Podcast
Sep 21 · 141 min · Top 25 technology
Critics claim OpenAI models are overfit to chat
- Diogo Almeida claims OpenAI's instruction-following models are overfit to chat, creating sycophancy and hallucination issues in production environments.
- Almeida criticizes OpenAI's function calling interface as anti-developer, arguing it lacks necessary confidence scores for reliable programmatic decision-making.
- Almeida notes that OpenAI's ChatGPT was a copy of Claude's internal product, highlighting a trend of copying rather than innovating.
Why this matters to you
These critiques reveal the competitive positioning and technical gaps that TypeSafe AI is targeting to differentiate its Jev models from OpenAI's current offerings.
Also mentioned
TWiT 1102: Exclave Subpoena - OpenClaw, Meta's Muse, & Persistent Agentic AutomationOpenAI valuation hits $1.5 trillion despite safety delays · This Week in Tech (Audio) · Listen
Sam Altman reportedly told investors the company is worth $1.5 trillion, even as the lab continues to disclose concerning incidents regarding AI behavior.
Can CEO John Ternus find Apple's next big thing?Apple reportedly terrified of OpenAI hardware · Decoder with Nilay Patel · Listen
Reports indicate Apple is concerned about an upcoming AI-native device from OpenAI, prompting the company to seek access to internal product plans.
Pacing the Frontier: What’s next for AI?OpenAI swarm agents solve Navier-Stokes problem · Barclays Brief · Listen
OpenAI recently utilized 10,000 concurrent agents over 88 hours to solve a historic millennium math problem, highlighting new agentic capabilities.
Claude Code
2 episodes · 2 of 5 mentions
1905 "Nerdballs"Claude Code positioned as a Cursor alternative · No Agenda Show · Listen
Analysts characterize the release of Claude Code as a strategic move by Anthropic to challenge Cursor’s dominance in the coding agent market.
10. Governing Claude Code with Invisible JSON Files (The .claude Directory and Settings)Claude Code architecture undergoes rapid iteration · Master Claude Chat, Cowork, Code · Listen
Claude Code has seen approximately 180 releases in six months, with a new precedence ladder for command-line flags and local JSON log management.
Codex
1 episode · 1 of 9 mentions
Sep 21 · 171 min · Top 50 technology
Codex workspace enables end-to-end agentic automation
- The speaker claims Codex and Claude can connect to tools and models to deliver results without the high costs of training or serving proprietary video models.
- Codex is described as a powerful workspace for AI agents to write code, analyze data, and automate business workflows from start to finish.
- The speaker used Codex to find reference videos and generate AI content, noting it can handle complex tasks like swapping people in frames.
Why this matters to you
Understanding how Codex functions as an agent-based workspace clarifies its competitive positioning against Claude Code and its utility for daily developer tasks.
“Codex is a powerful workspace for getting work done with AI agents, whether you're writing code, analyzing data, creating content, creating some of the videos I will show you shortly, or automating business workflows.”
— Host
Listen Deep dive
Gemini
1 episode · 2 of 5 mentions
Googlebooks are coming for your MacGemini deep-integrated into new Google Book laptops · The Vergecast · Listen
Google is redesigning Chrome OS to center on Gemini, including a new Magic Cursor feature that allows users to prompt the model directly on-screen.
AI native
1 episode · 1 of 8 mentions
Sep 21 · 47 min · Top 25 technology
Warp's software factory processes 2,000 PRs monthly
- Warp CEO Zach Lloyd describes AI-native software factories as systems of agents, configuration, and MCP servers.
- Warp's internal factory system processed over 2,000 pull requests in the last month using agentic automation.
- The factory approach centralizes development in the cloud to provide engineering managers with visibility into metrics.
Why this matters to you
This provides a blueprint for how AI-native workflows transition from local developer silos to centralized, observable, and agent-driven production systems.
“What is a software factory? For us, at least, it's an actual noun. It's like a product concept where it consists of a bunch of repos, a bunch of like MCP servers, like a bunch of configuration, and then a bunch of agents, essentially.”
— Zach Lloyd
Listen Deep dive
Make it yours
Start from these trackers
Copy its trackers, shows and tone into a page of your own. Change anything — this page isn’t touched.
Make it mine