0:00 A new app called Layered just launched, and it claims it can build a digital version of your entire wardrobe by analyzing your selfies.0:08 Not photos of your clothes.0:10 Photos of you.0:11 And this is why we do this every day—to sift through the dozens of launches on Product Hunt, to find the signal in the noise.0:19 To separate the genuine breakthroughs from the clever marketing, and figure out what’s actually changing for builders, for founders, for us.0:28 So let's get into it.0:29 Besides the selfie-stylist, we've got Chaterm, which is basically an AI co-pilot for the people who run the internet's plumbing—the Site Reliability Engineers.0:39 You tell it in plain English to diagnose a problem across a thousand servers, and it just… does it.0:45 Then there’s Origin, a development environment that’s so private, it uses special hardware chips to create a black box that even its own creators can't look inside.0:56 They're making a bet that privacy is about to become the single most important feature.1:02 And finally, there's Ray, a personal finance tool that lives in your computer's terminal.1:07 No fancy charts.1:08 Just a command line interface that acts like a CFO for your life, telling you what to do with your money instead of just showing you where it went.1:18 So we have an AI stylist, an AI engineer, a privacy fortress, and a command-line accountant.1:23 A pretty good snapshot of where we are right now.1:27 Everyone is trying to use AI to solve a very specific, very difficult problem.1:32 Okay, let's go deeper on the two that I think represent the fundamental push and pull of technology today: Chaterm and Origin.1:40 One is about giving AI the keys to the kingdom.1:43 The other is about building a kingdom AI can never enter.1:46 Let's start with Chaterm.1:48 The founder, Walter Duan, spent over a decade as an SRE, managing tens of thousands of servers.1:54 He knows the pain.1:55 The problem he’s solving is brutally real.1:58 When a system goes down, you have engineers digging through logs, running complex commands, trying to remember that one weird flag for Kubernetes that they used six months ago.2:09 It's high-stakes, high-stress, and it’s a huge bottleneck for most companies.2:14 Chaterm’s pitch is simple.2:16 You just type what you want.2:17 In English.2:18 "Diagnose the latency spike in the US-East-1 cluster and roll back the last deployment if it's the cause." According to the founder, Chaterm understands your infrastructure, plans the steps, and then executes them.2:32 Now, the first reaction from any engineer is, "No way.2:35 I am not handing that much control over to an AI." And that's the right reaction.2:40 But here’s the turn.2:42 Chaterm isn't trying to replace the engineer.2:45 It's trying to augment them.2:46 The really clever part isn't just the natural language.2:50 It's a feature they call "Agent Skills." When a team member figures out a new way to solve a problem, they can save that workflow as a reusable skill.2:59 So the team's collective knowledge gets codified.3:03 It turns tribal knowledge—the stuff that only lives in your senior engineer's head—into a tool everyone can use.3:10 That’s the part that isn’t hype.3:12 That’s a genuine solution to a massive, expensive problem in tech.3:16 It's not just about automation; it's about scaling expertise.3:20 And then you have Origin.3:21 It’s tackling the exact opposite problem.3:24 Not "how can we use AI more?" but "how can we build things without AI seeing what we're doing?" Every company I know is having this conversation right now.3:34 Developers want to use AI coding assistants, but leadership is terrified of their proprietary code ending up in some training model.3:42 It's a total stalemate.3:44 Origin's solution is hardcore.3:46 They're using something called Trusted Execution Environments, or TEEs.3:50 Think of it like a vault inside the computer's main processor.3:54 The maker, Marina Romero, puts it perfectly: "Everything inside...3:59 your code, prompts, files, everything...4:01 is encrypted and locked down at the hardware level." The encryption keys are generated inside the chip itself and never leave.4:09 This means even Origin, the company providing the service, cannot see your data.4:14 If the program crashes, the vault slams shut and the data is gone.4:18 Zero retention.4:19 This is not a software promise.4:21 It’s a hardware guarantee.4:23 And that’s a massive shift.4:25 For years, privacy has been a pinky-swear from tech companies.4:29 "We promise we won't look." Origin is saying, "We've built a system where we can't look, even if we wanted to." This isn't just noise.4:37 It’s a direct answer to a C-suite-level panic.4:40 As Marina says, "I think we are entering into an era in which people are starting to realise that privacy is as important as the ability to move fast!" She's right.4:51 And the first company that gets breached because their secret algorithm was leaked to a competitor via an AI prompt will prove her point in the most expensive way possible.5:02 So what about the other two?5:03 Where do they fit?5:05 Layered, the AI stylist, is the wild card.5:07 The idea of skipping the tedious task of photographing every t-shirt you own is brilliant.5:13 But it all hinges on one question: does the selfie-scanning tech actually work?5:18 If it's 95 percent accurate, it's magic.5:21 It's an instant digital wardrobe.5:23 But if it's 70 percent accurate and thinks your navy sweater is black, the whole experience breaks down.5:29 The maker says the most-used feature is now the travel packing—it builds a capsule wardrobe for your trip.5:36 That tells me they've hit a real user need.5:39 But the core tech is still a question mark.5:42 It's not noise, but it's pure, high-risk, high-reward execution.5:46 And Ray, the terminal-based CFO?5:48 This one is solving a real problem with a very sharp, specific tool.5:52 The creator, Clark Dinnison, said it best: with other finance apps, you "stare at a dashboard for 30 seconds, close it, and nothing about my behavior would change." Ray is designed to break that cycle by giving you directives, not just data.6:08 Running it locally, in a terminal, is a brilliant filter.6:11 It’s not for everyone.6:13 It’s for the person who’s comfortable with a command line and who values privacy and control over flashy interfaces.6:20 It’s a tool for a specific tribe, and for them, it could be a game-changer.6:25 When you zoom out, the picture from today is incredibly clear.6:29 The era of just slapping "AI-powered" on a product and calling it a day is over.6:34 Thank God.6:34 What we're seeing now is the second wave.6:37 These aren't just features; they're foundational bets.6:41 A bet that AI can be trusted with our most complex systems.6:44 A bet that we need fortresses to protect us from that same AI.6:48 A bet that computer vision is finally good enough to catalog your closet, and a bet that the best way to change financial habits isn't a pie chart, but a direct command.6:59 These are focused, opinionated tools built by people who have lived the problem.7:04 And that’s the difference between a product that makes noise and a product that makes a difference.
