0:00 An agent called Rosply just launched on Product Hunt that can watch your screen and control any desktop app for you, completely hands-free.0:09 In episode 86, we talked about sifting signal from the silence in daily launches.0:14 Well, today the signal isn’t quiet at all — it’s a full-on roar of AI agents, and they’re no longer just toys.0:21 They’re specializing.0:23 The main theme today, June twenty-third, is this Cambrian explosion of agents.0:28 They’re splitting into three distinct categories.0:31 You’ve got your infrastructure plays, like Conduit, which focuses on secure, local orchestration for coding teams.0:39 Then you have a whole new class of "AI Chiefs of Staff." Two big ones launched today: AirJelly and readywhen.0:46 Both of them plug into your work apps, watch for commitments you make in Slack or email, and then automatically create reminders and follow-ups.0:56 They’re trying to solve the "who's got the ball?" problem without you ever opening a project manager.1:02 Superhuman AI is making a similar claim on the email front, saying it saves users four hours a week.1:09 That’s a bold number.1:10 Even design is getting an agent, with Buddy Figma Agent automating large-scale canvas operations.1:17 Then you have the workflow accelerators.1:19 Glide just dropped a version that turns spreadsheets into full-blown apps with AI automations built-in.1:26 And for anyone in computer vision, T-Rex Label is a huge deal.1:30 It uses zero-shot learning to auto-annotate datasets, which is one of the most tedious, expensive parts of building a vision model.1:39 And of course, the voice agent space is just solidifying.1:43 ElevenLabs has a four-point-nine out of five rating from nearly two hundred reviews.1:48 That's not a launch, that's a utility.1:51 The one place to be cautious?1:53 A tool called Alai two-point-oh, an AI design partner.1:57 It was number three yesterday, but the platform has access issues.2:01 So for now, that’s just noise.2:03 Okay, let's go deeper on the two launches that really define the shift happening right now.2:09 It's the difference between what's practical today versus what's coming tomorrow.2:14 First is Happycapy.2:16 This one doesn't have the flash of a desktop-controlling robot, but it might be more important.2:22 The problem with most DIY agents has been that they need to run on local files.2:27 You're constantly uploading documents, feeding them context, and it's a mess.2:32 Happycapy changes that.2:34 It's a platform to build and launch agents that operate directly on your cloud data.2:39 It has a built-in database and analytics.2:42 What this means is you can build an agent that, say, permanently monitors your customer support database and automatically generates weekly insight reports without you lifting a finger or uploading a single CSV.2:56 It moves agents from being a clever script on your laptop to being a persistent, automated part of your company's infrastructure.3:05 This is the quiet, foundational work that enables everything else.3:09 And then you have Rosply.3:11 This is the complete opposite.3:13 It's all flash, but the implications are massive.3:16 By using computer vision, it can see and understand any application on your desktop, whether it's a modern SaaS tool or some ancient piece of enterprise software that doesn't have an API.3:28 It learns by watching you work, and then it can replicate those workflows.3:33 Think about all the repetitive, click-heavy tasks you do every day.3:38 Exporting reports, copying data from one system to another...3:42 Rosply is designed to automate that, not by plugging into an API, but by literally using the app just like a human would.3:50 It’s a robot for your screen.3:52 So on one hand, you have Happycapy making agents scalable, secure, and data-native.3:57 On the other, you have Rosply giving agents eyes and hands to work with the messy reality of the software we all actually use.4:06 This isn't just about more tools.4:08 It’s a fundamental change in what we should expect from software.4:12 The era of the single, all-knowing AI assistant is over before it even began.4:17 What's emerging is a specialized, distributed workforce of agents.4:22 Some work deep in the plumbing of your database, others sit on your shoulder and handle your email.4:28 The question is no longer if you should use an agent, but which of these new, purpose-built tools solves your single most expensive problem.
