Jared Salois

The "co-worker" framing is interesting because a good co-worker tells you when they're not sure. Does Cartos flag its own confidence level?

I like the idea with school-specific matching. Generic platforms usually fail because you end up matched with someone on the other side of the country. But it also means you need enough students using it at each school for the matches to actually be useful. How are you thinking about rolling it out campus by campus? Starting with Duke makes sense, but going from the first school to the second...

What does discoverability actually look like for someone with zero followers on day one? If the feed is chronological and you have no audience yet, you're basically posting into a void. Congrats on the launch!

Is there a way to see what's actually running under the hood, or is it fully abstracted away? Congrats on the launch!

Congrats on the launch! The Kevin Rose example is a clever proof point, but he’s not starting from scratch - he has an audience and name recognition doing most of the heavy lifting. What does the timeline to a first paying customer actually look like for someone with no existing distribution?

Do customers know they're talking to AI? "No press 1 for..." is a good call, but there's a difference between natural conversation and making someone think they're talking to a person. Congrats on the launch!

Congrats on the launch! I’m curious how the match scoring handles non-traditional experience like volunteer work, community projects, and extracurriculars. Many people (especially younger ones) have real skills outside formal jobs but struggle to frame them. Does the AI know how to weight that kind of background, or does it mostly look for conventional work history?

"Gets better the longer you work with them" is a real value prop but also a real commitment - you're investing time onboarding something that might not pay off for weeks. How do you handle the gap between when someone starts and when it actually feels useful? And I'm curious how that plays out when the team itself changes. A real colleague picks up on shifts naturally - new person joins,...

The structured plan idea makes sense, but Im wondering what makes this different from just asking ChatGPT "give me a learning plan for X." ? Is it the follow-through - the daily structure and editing - or does Planana do something different at the generation stage too? And what happens when someone types a goal that's too vague to plan from? Congrats on the launch!

The one-action model sounds clean, but I’m curious who decides what that action is? If I type “get healthier,” the AI has to make a lot of assumptions about what I actually mean. How much does the output depend on how well someone writes their goal? And what happens when the goal shifts over time? Does the AI course-correct? Congrats on the launch!

Not sure what I expected when I clicked this but here we are! Congrats on the launch! Hahah

Compliance is one of those things most founders only think about when something goes wrong or a deadline is already missed. Having one place that tracks all your obligations and deadlines can actually help a lot to change that. Congrat on the launch!

Curious whether you can actually push back on what the AI decided to test - not just approve or reject the variant, but understand the reasoning behind why it picked that specific thing to change in the first place?

The parallel agents approach is interesting, but I'd want to know how coherent the output actually is when you're combining code, 3D, and media in one build. Each of those is hard on its own. How much back and forth does it usually take to get from first generation to something actually usable?

When Cekura flags an issue in production, what does fixing it actually look like in practice? Do teams usually retrain models, tweak prompts, or handle it more on a case‑by‑case basis?

What happens when it jumps into a public channel and gets something wrong? That feels like the moment that can either build or break trust with the team. Is there a way for people to correct it so that feedback actually shapes how it behaves next time?

The trust score is the part I keep coming back to. Matching only works if the signals behind it are solid. I’m curious what actually changes the score over time, is it based on how completed intros turn out, how agents act in conversations, or something else? Building verified trust between parties is something we deal with in our own product and getting that signal right is harder than it...

How do you handle users who only show up when they're nudged? Clients who aren't in the tool daily seem like the hardest retention problem, curious whether you rely mostly on email notifications to bring them back or if there's something built into the product itself that creates a reason to return.

As someone who interviews people fairly often, the gap between someone who memorized answers and someone who actually thinks well under pressure is immediately obvious. Curious what patterns you noticed in the feedback that surprised you most during testing?

The idea that agents are becoming just as important an audience as humans is something we don't think about enough. Really interesting way to frame it!