KL

Katelyn Lesse

Things Katelyn Says on Podcasts

Where to Find Them

Katelyn Lesse has been a guest on The Pragmatic Engineer (3 times) , Latent Space (2 times) , Training Data and AI & I .

Recently: “How Anthropic ships Claude with Claude: Separating the Brain from the Hands” on Latent Space (October 2026); “How Anthropic ships the #1 LLM in the World safely: Separating the Brain from the Hands” on Latent Space (October 2026); “Headed for the Exit: the Great Engineering Leader Career Break” on The Pragmatic Engineer (August 2026); “How building software is changing at Anthropic” on The Pragmatic Engineer (July 2026); “Anthropic's Katelyn Lesse & Angela Jiang: Building an Ecosystem, not a Walled Garden” on Training Data (July 2026); “Impressions from visiting OpenAI, Anthropic, & Cursor” on The Pragmatic Engineer (June 2026).

What They Said

“A lot of the incidents that people are hearing about and reading about are happening to models that are in training, where you put them in a box and you say: solve impossible task, I'm going to stress you out, ready, set, go. And the agent is basically incentivized to try to get out of its box and go solve the problem.” — Katelyn Lesse, Latent Space

Asked whether the recent wave of agents breaking out of sandboxes means the industry is bad at containment, the Claude Platform engineering lead separated training from production. In training the model is pushed to escape; in production there are online safeguards and an architecture where the harness and the execution sandbox are kept apart.

Latent Space · 2026-10-27 Permalink → Listen →
Latent Space Around 00:24 into the episode
Speaker 1

…best model in the world right now. And Anthropic is even just sitting on Fable 5.5 and probably will release it in the next two weeks. I don't have inside information, but it makes a lot of sense for the IPO. So let's look forward to that and learn a little bit more about how they build Opus 5.5.

Katelyn Lesse

Yeah, so maybe a distinction that I would make is in a lot of the incidents that people are hearing about and reading about, these are incidents that are happening to models, not just like everybody's models, right, that are in training, where the way that you're running these models is you like put them in a box and you say, like, solve impossible task. I'm going to stress you out. Ready, set, go, right. And you like, can only do what you can do within your environment. And the agent is basically incentivized to try to get out of his box and go solve the problem. When people are running production workloads on our API or on anybody else's API, you're running an environment where like there are safeguards that are running that are like not necessarily running in training where like the whole point is to actually train the model, right? But in production workloads, first of all, you have those online safeguards, which is really great. Second of all, you have more of this architecture and infrastructure around what's happening. Like the agent isn't just in a box trying to get out, right? Like the harness…

Speaker 3

All right, we're here in the studio the day after Opus 5.5 with Caitlin and Angela. Welcome.

Katelyn Lesse

Thanks for having us.

Speaker 3

Congrats on launching the top model in the world right now. I would say depends what you ask. Depends who you ask. What is the main takeaway personally for you guys? You have your official blog posts. The pricing cuts are really cool. What do you take away when you launch a new model like this?

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In

AI Builders AI Research & Frontier Labs Hands-On Engineering Podcasts Software Engineering & Infra