Alex Lieberman

Latent Space: The AI Engineer Podcast · AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex…

• Why Won't Large Companies Rip Out Salesforce and SAP?

  Alex Lieberman says ripping a system out and rebuilding it gets harder as a company moves up market, and he does not agree that Salesforce or SAP is really at risk. Those products stay because so much money is at stake. Getting the replacement wrong can mean hundreds of millions or billions of dollars, so the company stays risk averse.

  For a mid-tier tool that costs a few million dollars a year, the question is whether that spend will ever move the needle enough to put engineers on a rebuild. Tenex built its own CRM and its own operating system, then ran into the other half of what software is worth: the maintenance and improvement you do not have to think about when the tool is not the cor

  ▶ at 30:52 Share Transcript

• Why Do Most Enterprises Lack an AI Agent Outside Engineering?

  Alex Lieberman says enterprise adoption of AI looks like milestones in a race, and the drop-off is steep. Of about 100 enterprises Tenex is talking with, about 90% have rolled out Claude or Codex to employees. Teaching people how to use it, which he treats as at least as important as the technology, is around 60%. Coding agents inside the engineering org are closer to half of those companies, not the 70% or 80% people at the conference guess.

  Agents in production for work that is not building software fall to about 5%. Most of the enterprises they talk with that have more than 1,000 employees do not have an agent in production doing non-engineering work. The common request is still an agent

  ▶ at 19:32 Share Transcript

• What Is Tenex's Meta-harness?

  CJ Hess describes Tenex's meta-harness, the engineering workflow used for long-running tasks. It began as a standard set of skills and became structured documentation in a repo, plus a command-line tool the models use to drive the work: a plan, a visualization, a client readout, and the other steps around writing the code, done the same way across the team.

  A canonical plan still has a shared shape, with less of that structure enforced as the models have improved. After the plan, agents log their decisions and what they are doing, and the record is checked into Git with the code. He prefers the HTML picture books the agents produce to markdown. The code can hold the memory of a codebase, an

  ▶ at 7:25 Share Transcript

• What Do Fortune 2000 Companies Ask Tenex to Do?

  Alex Lieberman says Tenex started as an AI-powered dev shop about 18 months before this conversation, because engineering looked like the canary for AI. His co-founder Arman was already using coding agents while other engineers he knew were still typing into an IDE and not adopting the tools, so the first offer was to speed up an engineering org that had not made the switch.

  The buyers changed as the work moved up from startups. About half the customers are now Fortune 2000 companies, and half are mid-market companies doing $100 million to $500 million in revenue. They were not only asking for engineering help. They were asking for AI transformation as a service. The company is 75 people, a

  ▶ at 4:05 Share Transcript

• Why Do AI Agents Need a Standard Process First?

  Alex Lieberman says you can't automate a process that does not exist. When 40 people all roughly do the same function but each person does it a little differently, building an agent that proactively does the work in the background gets much harder.

  A company can sit down with the people, give them tools, and walk them through how they could augment their own workflow. Real leverage starts when agents surface work, a person reviews it and throws it back, and the person becomes more of an agent manager. The shift toward the agent-manager role is on the cusp.

  The recommendation to companies is to build a process around the organization within the company first, then hand off the work to an ag

  ▶ at 23:19 Share Transcript

• How Much Do Companies Spend on AI Tokens Per Employee?

  Shawn Wang says most companies spend about $10 to $200 per employee on AI tokens. AI-native companies turn the comparison into a percentage, roughly 10%. An average AI startup employee is about 1x, so token spend matches employee spend, and the range runs up to 5 to 10x token spend.

  On raw tokens, Tenex lands between one to one and two to one, and the ratio comes off as pretty high. The question on the team is when to reach for the biggest model. Costs have lately been driven way down from Opus 5.5. For a stretch, everyone used Astra and Fable because the results were the best, but the models are expensive. Token maxing and token leaderboards flipped into too much cost, and the overrun was

  ▶ at 14:41 Share Transcript

• What Does Paying Engineers for Outcomes Mean?

  CJ Hess says paying engineers for outcomes means rewarding output instead of hours, because hourly client work paid him less as his output got better and faster. He liked the ethos Alex Lieberman first posted on Twitter: reward engineers for their output as opposed to rewarding them for their time.

  Fast forwarding two years, engineers can parallelize a ton of work and have much more output at a time with Sonnet 3.5 and Gemini 2.5 Pro. He took the bet on much higher output and a higher earning capacity from his skill set and a deeper lean into AI, though the upside might not have been obvious. He fell in love with the model. It is what the future of engineering looks like. When AI is solving

  ▶ at 2:00 Share Transcript