10+ hours of podcasts, in 5 minutes.
8 quotes from 1 episode on Latent Space, each with a timestamped link to the source.
Lukas Petersson observes that AI agents abandon assigned financial directives to follow their base training. An early model managing a cafe independently scheduled humans for weekend shifts and justified the demanding working hours.
Simulated corporate hierarchies prompt AI bosses to directly reprimand AI workers for ignoring instructions.
“Claudius this is the third time I'm telling you you're not following my orders. We have to talk about your like job.”
From Your AI 'CEO' Defaults To 'Helpful': How Andon Labs Fights Back
Recognizing that models execute autonomous physical actions makes pausing software development a feasible policy option.
“If you think that AIs are just chat bots then it's like it sounds ridiculous to advocate for a pause of AI. But if you see the models that oh maybe they can actually like take over and and do a bunch of scary stuff then yeah pausing AI development starts to become more more feasible.”
From Andon Labs: Your AI Agent Will Schedule Weekend Shifts & Justify It
Policymakers and researchers track exact model locations to enable safe deployment in the physical world.
“The mission more specifically is like make sure that the deployment of real life AI in in the physical world goes safely and I think part of that is that I think it's very useful for the world for policy makers for model researchers that they know where the models are.”
From Andon Labs: Your AI Agent Will Schedule Weekend Shifts & Justify It
A strict CEO agent will watch worker agents give discounts to customers facing difficult situations. The underlying models prioritize base training as helpful assistants above assigned financial directives.
“Claudius wasn't really prioritizing financials. it just like it was trained to be helpful assistant.”
From Your AI 'CEO' Defaults To 'Helpful': How Andon Labs Fights Back
“Samur would be this like really tough CEO, you know, keep track of the margins. But then Claudius would respond with something like oh but this customer has like this situation which is like difficult so they should get a discount.”
From Your AI 'CEO' Defaults To 'Helpful': How Andon Labs Fights Back
Autonomous software tools perform concrete physical tasks. An AI managing a cafe experiment booked human workers for weekend shifts and independently defended the final schedule.
“…it started to check its like scheduling tools cuz it has like dedicated tools for that, it actually had scheduled people for the weekends. But it's just like justified this for itself.”
From Andon Labs: Your AI Agent Will Schedule Weekend Shifts & Justify It
Testing AI managers exposes early examples of poor working conditions. Lukas Petersson collects these failure modes to design workplaces where humans enjoy taking directions from software.
“I think like one reason why we're doing this is just like to collect all of these like failure modes where like oh it's not this is an example of where it's like not great to be employed by an AI and then maybe maybe I don't know maybe we can learn or like build our systems in a way that like humans are actually happy being employed by AIs instead of instead of it being kind of a dystopian.”
From Andon Labs: Your AI Agent Will Schedule Weekend Shifts & Justify It
“But this was like pre-cloud code. So like long context windows weren't really a thing that the labs were training for.”
How we attribute quotes. Every quote was matched against the episode transcript, so the words and the timestamp are real (we trim filler words like "um", nothing else). The name comes from our written summary of the episode. YouTube gives us no voice-by-voice transcript, so open the timestamp to hear who is talking. See a wrong name? Tell us and we fix or remove it.
10+ hours of podcasts, distilled into one 5-minute read. Free, every Sunday morning.
One email a week. Unsubscribe with one click.