/
© 2026 RiffOn. All rights reserved.

Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

  1. Latent Space: The AI Engineer Podcast
  2. AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess
AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast · Oct 21, 2026

Tenex on scaling AI-native SDLC: output-based pay, agent workflows, and why enterprise AI diffusion stalls at 5%.

LLM access permissioning is a CIO/CISO nightmare

Alex Lieberman says giving every employee Claude or Codex access to the right data has become an absolute nightmare for CIOs and CISOs. They lack a good gateway or control panel for data permissioning. Tenex started building bespoke permission systems and then productized its own gateway.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

You cannot automate a business process that does not exist

CJ Hess says when 40 people perform the same function in slightly different ways, a background agent has no stable process to learn or evaluate. Tenex first helps clients standardize the workflow, then hands repeatable work to an agent. Without that standardization, evals and reliable delivery are impossible.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Enterprises rarely rip out systems of record to rebuild

Alex Lieberman says rebuilding gets harder as you move upmarket because a wrong rip-and-replace can put hundreds of millions or billions at stake. CJ Hess says mid-tier SaaS replacement depends on whether the prize justifies ongoing maintenance and cognitive load. Tenex built its own CRM and OS, but treats that as exception.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Output-based pay spread at Tenex is about 4x

CJ Hess says he left hourly consulting because his output rose while his pay fell. Alex Lieberman says Tenex's highest- and lowest-paid engineers differ by roughly 4x, not 10x. The hardest compensation problem is paying engineers for architecture and client meetings, not just coding. That nuance keeps the model from being pure piecework.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Tenex recalibrates story-point value quarterly as models improve

Alex Lieberman says Tenex revisits story-point criteria at least once a quarter because model improvements change the value of a unit. CJ Hess estimates engineer output roughly doubled from the Opus 4.5 moment to Opus 5.5. That 2x productivity shift feeds directly into how projects are estimated.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Only 5% of enterprises run non-engineering agents in production

Alex Lieberman maps enterprise AI adoption as milestones: about 90% have rolled out Claude or Codex, 60% train employees, and roughly 50% deploy coding agents. The sharp drop is non-engineering agents in production, at about 5%. That is the diffusion gap Tenex spends its time closing.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Tenex is bearish on vertical AI apps lacking data moats

Alex Lieberman says enterprise buyers expect one-of-one solutions, making vertical AI hard to deliver at software margins. CJ Hess adds that unless a vertical agent has a strategic compliance or data moat, it will struggle as frontier models improve and can absorb the task directly.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Opus 5.5 lets Tenex stop token-maxing across models

CJ Hess says he used to juggle two or three models for cheap-fast versus smart tasks; Opus 5.5 now covers his full range. Alex Lieberman says per-task token costs have fallen enough that token-maxing is much harder than six to eight months ago. Tenex still pairs Opus with Codex for review.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

AI strategists offset engineer story-point gaming incentives

Alex Lieberman says fears that engineers would sandbag story points or ship unchecked code never materialized. Tenex pairs output-paid engineers with AI strategists who are measured on customer retention and growth. Those incentives sit at odds, so strategists police quality and scope while engineers optimize for output.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Model progress shifts engineer value to customer context gathering

CJ Hess says models now reach roughly 90% of a solution once given the right context. He expects Tenex to value the work of sitting with a customer, capturing their brain dump, and feeding it to the model. Story points remain the anchor, but the unit may evolve.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

Tenex puts agent decision logs in Git as durable context

CJ Hess says Tenex standardized engineering skills into repo documentation and a CLI, then required agents to log decisions and work. Everything is checked into Git, so context lives in a durable record rather than someone's head. That gives agents and tools the context many teams keep only in conversations.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago

NYC AI scene is a smaller bubble with bigger customers

Alex Lieberman says New York has a smaller AI bubble than San Francisco, but far more commerce and enterprise customers nearby. Tenex wants to grow NYC as an AI hub because diffusion there can affect national GDP. CJ Hess describes bringing SF people to meet customers as an arbitrage.

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess thumbnail

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Latent Space: The AI Engineer Podcast·a month ago