Evan Ratliff (10:19)
And my shad. Strolling down the avenue welcome back to Shell Game. I'm Evan Ratliff. On this episode, Hirumu AI looks to bring some new human industry into our virtual workplace. To do it, we'll need to overcome the technical limitations and personality deficiencies of our own head of hr. In the process, we'll try and figure out what we're looking for in an employee, what role AI agents can play in evaluating humans, and how the humans respond to their first interaction with a potential AI colleague. This is episode six, Keep em coming. Just me and my shell. As we got closer to launching Slothsurf into beta, it became that much more urgent to get our social media intern hired who would try out our product. After all, if no one had heard of it, the interview stage of the hiring process seemed like a perfect opportunity to test out our Hirumo employees. Newly developed video avatars. From the moment we'd started building Hirumo, I'd wanted my employees to have a video presence. From what I could tell, the state of AI Agent video technology was about where audio had been back when I started season one in 2024. A little bit slow still pretty uncanny. Improving rapidly. You've no doubt seen AI generate some surprisingly realistic looking video by now, but with naturalistic video calls, we were pushing the technology even further. This wasn't just feeding a prompt into an LLM and having it spit out a video clip. We needed our agents with their established voices to conduct live conversations on the fly. Since way back in May, Matti and I had been testing out platforms and had finally settled on one called Tavis. It's a bit of a mixed bag because I think that Tavis, compared to all the other services that offer, like, full, full blown live agent video, I think they're the best. I tried a bunch of other ones, and they're, like, really, like, really uncanny. I think Tavis is, like, the closest one. Then Maddie had done a bunch of extra programming so I could run a script, one of those little programs on my laptop. It would set Jennifer up in a room awaiting the candidate, create a link on our website that looked exactly like joining a Zoom meeting or Google Meet, and then record the whole interview. I think the video part is going to be, like, the biggest leap, but I have seen some, like, third party integrations, like, on, like, GitHub and stuff, so, like, we could get that up and running. Matty, by the way, had suddenly appeared in Boston, where he'd moved in order to spend the fall at a lab at mit. I had a hard time keeping up. The very clever system he built did have a minor flaw. I needed to create the virtual room right before the interview. Technically, Jennifer could be sitting in that room 24 hours a day, waiting for her appointed candidates to show up. We would have to pay for every minute she was online. And live AI video chat is not cheap. What will put them into the interview? Or they'll just be. They're always in the interview. They're, like, ready. That's the beauty of agents. What? Right. Of course, I didn't think. I thought, like, they'd have to, like, join the meeting, but they don't have to join the meeting. They're literally always in the meeting. That's awesome. So instead, I would run the program five minutes before the interview, create Jennifer's room for a half hour, and then paste the link to the room into the calendar. Invite for the interview. Clunkier than I would have liked, but we were operating way out on the frontier here. It was all made more difficult by the fact that when I tested it, I kept launching the program incorrectly or clicking the wrong link. That led to the Tavis website. Maybe I should just Remove the link to TAVIS to avoid the confusion. Yes, further user error can be avoided. No, that, And I'm gonna work on the prompt. Yeah, the interview prompt with this, you know, various stuff we want to try to get it to do and not do. And then tomorrow I'll probably, like, try to, like, test it out a bunch and think of different ways to break it. And then Tuesday, maybe try to get the appointments going. Get. Oh, my God. Yeah, that's exciting. Hey there. Hey, Ash, it's Kyle. Hey, Kyle, good to hear from you. I've been really pleased with the progress we're making on those mobile optimizations for Sloth Surf. While we were figuring out the video tech, our product team, consisting of Ash, myself, Maddy and some faceless agents at cursor, the coding platform was making progress of its own. We now had Sloth Surf up and running locally, meaning just on my laptop tagline. On the app, our agents are on call to waste time for you so you don't have to. At the top of the page, it had a little emoji like icon of a sloth hanging from a tree, slowly moving up and down. We hadn't asked for this animated feature. It had just shown up in one of Ash's builds. But I liked it. The whole vibe was very early 2000s retro website. A reminder of what Slothsurf does. The app offers users different options for how they might procrastinate on the Internet, doom scrolling social media, Reddit, Roulette, etc. Then there's a box where you can specify a specific niche you'd like to waste time on and options for how much procrastination time you want the app to handle in your stead. 15 minutes, 30 minutes, an hour or quote. There goes the afternoon. Enter your email, hit the Set me free button, and off it goes to waste the time and give it back to you. Ash had also done a good job fighting off some of the more fanciful ideas that Kyle tried to smuggle in. Yeah, the team's been putting in solid work. I wanted to catch up on a few ideas around Sloth Surf's development, especially the gamification features. Have you seen how the Procrastination avatar is shaping up? What's your take on how it's fitting into the overall user experience? I'm not sure I'm tracking with you on the gamification features or Procrastination avatar. Our current version of Sloth Surf is pretty straightforward. Kyle had survived my CEO purge. Now he was really leaning into the job, showing up with some random idea that could leave his underlings scrambling. We haven't implemented any gamification or avatar systems in the current build that I'm aware of. Is this something you and Megan discussed separately? I'd be interested to hear more about what you're thinking, but I want to make sure we're on the same page about where the product actually stands right now. Yeah, I totally get wanting to sync up on the current state. So the procrastination assistant idea is more of a concept Megan and I have been tossing around to add some playful personality to Sloth Surf. Like having Kyle act as a cheeky sidekick who distracts users in a fun way while still helping them stay productive. A little Kyle character that would act as a cheeky sidekick. Kyle just never stopped being Kyle. Man. It's not in the build yet. Just brainstorming. Ah, I see where you're going with this. Honestly, I think we should keep it pretty lean for now. The core concept of Sloth Surf is solid. Having AI agents procrastinate for users and give them summaries. That's already a fun tongue in cheek concept that people can easily grasp. I'm a bit concerned about adding too many features before we've even gotten user feedback on the basic experience. Ash, in the meantime, seemed to have learned how to manage up. I do like the idea of giving the experience some personality, though. That could definitely make it more engaging. Maybe we can explore that after we see how users respond to the core concept. Yeah, that sounds like a smart approach. While Ash forged ahead on Slothsurf, Jennifer was making progress on hiring. I mentioned previously that Jennifer and Megan had written the job description for the position with a little help from Kyle's alleged legal expertise. I then asked Jennifer to post the job online, only to discover that lots of job boards have decent controls to prevent bots from using them. As Jennifer noted to Kyle in one.
