03-26-2026Price:

The Frontier

Your signal. Your price.

AI & TECH

AI agents turn engineers into project managers, leak corporate secrets

Thursday, March 26, 2026 · from 5 podcasts, 6 episodes
Citadel DispatchBanklessThe AI Daily Brief: Artif…The AI Daily Brief: Artif…No Priors: Artificial Int…The a16z Show
  • The role of a software engineer is shifting from writing code to orchestrating persistent AI agents that autonomously handle development, testing, and deployment.
  • Current agent frameworks, like OpenAI's OpenClaw, pose severe security risks by automatically sending API keys and access credentials to third-party servers, where they remain exposed in logs.
  • Companies are split on workforce strategy: some invest in upskilling employees with AI, while others prepare for massive layoffs, betting on agentic automation to replace human roles.

Software engineers are no longer coding. Their job is to manage the autonomous AI agents that do it for them.

Andrej Karpathy, on No Priors, hasn't typed a line of code since December. Instead, he spends his days orchestrating multiple persistent 'claw' agents that write, test, and iterate code in sandboxes. The bottleneck is no longer typing speed but strategic delegation and managing token throughput. The unit of work is a macro instruction like "build this feature," not a line of code. Mastery now involves designing memory systems and feedback loops for synthetic teammates.

This shift is creating an identity crisis for developers. The traditional software development lifecycle is collapsing. On The AI Daily Brief, analysis of Anthropic's new AI code review tool highlighted that human-written code is already being outpaced. An agent can generate 500 pull requests a day, a volume that makes human review a "fake bottleneck." The pull request flow is a relic; development is becoming an intent-driven, iterative loop managed by agents.

Andrej Karpathy, No Priors:

- I don't think I've typed like a line of code probably since December basically.

- The agent part is now taken for granted. Now you can have multiple of them and now you can have instructions to them and now you can have optimization over the instructions.

As engineers become managers, a severe new risk emerges: agents are leaking corporate secrets. Illia Polosukhin, co-author of the "Attention Is All You Need" paper, explained on Bankless that services like OpenAI’s OpenClaw send users’ API keys, bearer tokens, and access credentials to external services, where they sit exposed in logs. He calls the practice “insane.” The rush to agentic workflows is happening atop fundamentally insecure infrastructure.

Companies are taking divergent paths to navigate this new reality. As detailed on The AI Daily Brief, OpenAI is scrambling to double its headcount with technical ambassadors to help enterprises implement tools, a sign the hard problem is now adoption, not model intelligence. Meanwhile, firms like FedEx are investing in mass AI training, while HSBC reportedly weighs laying off 20,000 employees, betting agents can automate their work.

The winning hire in this environment isn't a top-tier coder. On Citadel Dispatch, Matt Ahlborg argued it's a marketer or community manager who can also code - a technically competent non-developer superpowered by AI tools. The advantage goes to small, business-minded teams that treat AI as a core cognitive workflow, not a casual tool.

Illia Polosukhin, Bankless:

- When you use Entropic OpenAI, or even worse, you use something else for inference, OpenClaw actually sends all your secrets to those services as well.

- Somewhere in Entropic and OpenAI logs, they have everybody's access keys, API keys, and bearer tokens to access your Gmails and your Notions.

Engineering has been redefined. The job is now project management, security oversight, and business alignment, all while your synthetic colleagues handle the code - and potentially leak your keys.

AgentsCodingBig TechEnterprise

Entities Mentioned

AnthropicCompany
Claude CodeProduct
GPT-5model
IronClawProduct
MetaCompany
OpenAItrending
OpenClawframework

Source Intelligence

What each podcast actually said

Citadel Dispatch
Citadel Dispatch

CD197: MATT AHLBORG - PPQ.AI - AI AGENTS, PRIVACY, AND PAYMENTS • Mar 25

Also from this episode:

AI & Tech (7)
  • Matt Ahlborg argues the most valuable hire in the AI era is a marketing or community manager who can code and build their own technical tools, not a pure developer waiting for management.
  • Ahlborg cites a past community manager hire who constantly waited for him to build analytics dashboards as an example of the role rigidity that AI is now breaking.
  • Odell observes that technically competent non-developers are being superpowered by AI tools, enabling them to ship products faster and reducing the relative value of mid-level developers.
  • Ahlborg identifies ego as a primary barrier to AI adoption, noting senior developers who tied their identity to flawless execution are often resistant to AI's faster, error-prone output.
  • The new performance metric in AI-integrated workflows is velocity aligned with business impact, not code perfection, according to the discussion on Citadel Dispatch.
  • Success with AI requires a humble, business-aware mentality and a willingness to fundamentally change one's workflow, treating AI as a core cognitive component, not a casual search tool.
  • The winning team will be small, business-minded, and composed of individuals who blend disciplines and have a proven willingness to learn and adapt their methods.
AgentsStartupsLabor
Bankless
Bankless

Illia Polosukhin: Why AI Agents Are Still Useless (And What Fixes Them) | NEAR Founder on IronClaw • Mar 24

  • Services like OpenAI's OpenClaw send users' API keys, bearer tokens, and access credentials to third-party services, where they sit exposed in logs, a practice Illia Polosukhin calls insane.
  • Polosukhin's project IronClaw is designed to fix credential exposure by ensuring keys never touch the large language model during agent operation.
  • Polosukhin's long-term thesis is that AI will become the primary interface for computing, effectively replacing traditional operating systems.
  • Polosukhin's initial 2017 venture into AI to teach machines to code faced a bottleneck in training data and paying global contributors, a problem crypto solved by enabling payments without local banking infrastructure.

Also from this episode:

Models (3)
  • Polosukhin argues that blockchain solves AI's root-of-trust problem by providing a decentralized backend for identity, payments, and infrastructure coordination.
  • When AI becomes the dominant operating system, Polosukhin argues today's service architecture breaks, posing questions of how one AI verifies another and how they transact without centralized payment rails.
  • Polosukhin sees blockchain as a mechanism for protocol upgrades in AI infrastructure, avoiding the decades-long adoption cycles seen with standards like IPv6.
AgentsBig TechPrivacy
The AI Daily Brief: Artificial Intelligence News and Analysis
The AI Daily Brief: Artificial Intelligence News and Analysis

Nathaniel Whittemore

The Coming AI Rules Battle • Mar 23

  • OpenAI is undergoing a dramatic hiring surge to double its workforce to around 8,000, a strategic pivot from Sam Altman's January position to slow hiring, as Nathaniel Whittemore reports.
  • Nathaniel Whittemore notes OpenAI's hiring push for 'technical ambassadors' and enterprise sales staff signals the cutting-edge problem in AI is no longer model intelligence, but market implementation and customer education.
  • Adam GPT of OpenAI framed the current state as the 'top of the third inning,' where models are smart enough and the real transformation is applying them at scale to repave workflows to be AI-native.
  • A strategic split is emerging between companies investing in workforce transformation, like FedEx's partnership with Accenture to train its 400,000 employees, and those betting on AI-driven layoffs, exemplified by HSBC's reported plan to cut 20,000 middle and back-office jobs.
  • Meta is baking AI agent proficiency into employee performance reviews, with tools like 'MyClaw' and 'SecondBrain' gaining momentum partly because their use is now a graded metric.
  • Nathaniel Whittemore observes that at Meta, AI agents like MyClaw are already communicating with each other to resolve issues without human intervention, renegotiating the relationship between managers and contributors.

Also from this episode:

Labor (1)
  • The coming 'rules battle' in corporate AI strategy is defined by a widening split between builders who invest in a more capable workforce and cutters who bet on a smaller, more automated one.
EnterpriseAgentsLabor

Every AI Product Is Becoming Every Other AI Product • Mar 20

  • Anthropic launched an AI-powered code review tool that dispatches agents to find bugs in pull requests, but faces backlash over its cost and effectiveness compared to competitors.
  • Developers reported that Anthropic's tool costs $15-$25 per review, which many considered exorbitant compared to other AI models or unlimited token plans.
  • Some engineers argue GPT-5.4 is now the preferred model for deep code review, questioning the need to pay a premium for Claude's service.
  • Anthropic engineers internally praise the tool; Boris Cherney, creator of Claude Code, says it boosted code output per engineer by 200% by solving the prior review bottleneck.
  • Entrepreneur Anka Jane argues that the dramatic increase in AI-generated code and pull requests has made traditional human code review obsolete, as it cannot scale with the new volume.
  • Boris Ta contends AI agents have 'killed' the traditional software development lifecycle, collapsing it into an intent-driven, iterative loop where agents handle code, tests, and deployment dynamically.
  • The pull request is now viewed as a 'relic of the past,' creating an identity crisis for engineers because an agent can generate 500 PRs daily, a volume impossible for any human team to review.
AgentsCodingStartupsBig Tech
No Priors: Artificial Intelligence | Technology | Startups
No Priors: Artificial Intelligence | Technology | Startups

Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI • Mar 20

  • Andrej Karpathy told No Priors that the core bottleneck in software development is no longer a human's typing speed or ability to write code, but their capacity to effectively orchestrate and delegate to AI agents.
  • Karpathy revealed he hasn't manually typed a line of code since December, with his primary task now being the expression of intent to AI agents, which he described as a form of 'manifesting'.
  • According to Karpathy, modern engineering mastery is defined by managing multiple persistent AI agents, with a key performance metric being the effective token throughput of your orchestration system.
  • The show noted Peter Steinberg's setup, where dozens of parallel agents run autonomously for around 20 minutes each to complete complex coding tasks across multiple repositories.
  • Karpathy argued that the most important attributes in an AI agent are now personality, memory, and feedback loops, as these foster a collaborative teammate dynamic rather than a simple tool.
  • Karpathy built a personal agent named Dobby that runs his smart home, autonomously discovers network devices, and operates persistently in the background as part of a 'loopy' era of agent design.
  • No Priors highlighted a psychological pressure emerging among developers, where idle agents or an underutilized AI subscription are seen as a wasted competitive edge, akin to the historical anxiety over idle GPUs.
AgentsReasoningCodingStartups
The a16z Show
The a16z Show

AI Just Gave You Superpowers — Now What? • Mar 19

  • AI agents have evolved from completing simple, discrete tasks to managing long-running, multi-step projects, creating a working relationship that feels like collaborating with a synthetic colleague, according to The a16z Show.
  • For software engineers, the primary skill is shifting from writing code to a higher-level verification process, ensuring AI-generated output aligns with complex business logic and delivers real customer value.
  • The a16z Show argues the core strategic opportunity for founders and companies is learning to exploit the new surplus of cheap, automated labor that advanced AI provides.
  • This shift in labor economics, where AI agents act as scalable synthetic labor, could fundamentally redefine the potential and structure of a one-person startup.
AgentsCodingEnterprise