MChris, what's the single most annoying thing about working across AI tools right now?
CSigning into Claude, then ChatGPT, then checking which one has the file I need. It's like juggling wallets.
MZach Perret, the Plaid co-founder, posted exactly that. He says one day Claude, ChatGPT, and Google will auto-recognize which account can access each doc, and that day will be awesome.
CHe's right, but let's not pretend it's some distant sci-fi future. Account switching already exists in most of these apps.
MClaude's mobile app already does multiple accounts. ChatGPT lets you have two signed in on the web. The missing piece is the auto-recognition thing.
CAnd that's harder than it sounds, because each of these companies wants you living inside THEIR ecosystem. Auto-recognizing across platforms means sharing document permissions.
MWhich is a privacy quagmire if done carelessly. You'd be giving every AI provider a map of your files.
COkay, but the current system is WORSE. You end up copying text manually, pasting between tabs, losing formatting. The friction is constant.
MSo you want one sign-in to rule them all? A Google Account for AI?
CNo, I want the tools to be as seamless as email. I don't think about which SMTP server handles my message; it just works. AI should be the same.
MEmail had decades to standardize. AI is still a land grab. None of these companies want to hand over identity control.
CBut the customers clearly want it. Zach's post got engagement precisely because it names a pain everyone feels.
MThat said, if Perret is this annoyed about AI tooling, maybe his next startup solves it. He's done it before.
CAnd I'd be first in line. Moving on to something more concrete, a library with sewing machines and robotics.
MOn Bluesky, Zalno says they talked to someone and confirmed a local library is building a community space: sewing machines, a robotics area, other hands-on stuff.
CI love libraries doing this. Public libraries today are becoming maker spaces, and that's the right evolution.
MA century ago, they were about books and silence. Now? You might check out a sewing machine or learn to solder.
CTax dollars funding access to tools people can't afford individually — that's civic infrastructure working correctly.
MAnd robotics in a library exposes kids to engineering before they hit high school. Before a coach can scare them into it, which brings me to the last one.
COh no. What happened?
MKristine Smith on Bluesky posted a savage paragraph about Palmer Luckey and his robot wardog company. She says he's the last guy you want overseeing a high school robotics team, let alone a biotech firm breeding mutant wardogs.
C'Breeding mutant wardogs' is a colorful way to describe engineering quadrupedal robots for defense.
MIt's satire, clearly. But she's making a real point: operational failures follow big-money, aggressive personalities. We saw them with Oculus and Anduril's early drones.
CHold on. Luckey co-founded Oculus, sold it to Facebook, then built Anduril into a major defense contractor. A lot of people would call that track record impressive, not failure-prone.
MThe critique isn't that he can't build companies. It's that the engineering culture rewards speed and risk over rigor. That works in software; less so in biotech.
CBut what's the actual operational failure she's citing? Her quote doesn't include one. It's just a vibe: he's aggressive, therefore bad.
MThe whole context is his defense work and now this biotech venture. She's skeptical from the start. You don't have to wait for the failure — you can judge the incentives.
CThat's exactly the kind of gatekeeping that keeps innovation stuck. Guy builds hardware, gets funding, expands into a new field. That's how progress happens!
MBiotech isn't consumer hardware. If a robot dog breaks, you fix the actuator. If a biotech product fails, people can die. Different stakes.
CWhich is why biotech has the FDA, clinical trials, independent review. None of which disappear because the founder is ambitious. The process holds.
MThe process holds if the culture respects it. Her worry is the culture won't. That's fair to debate.
CI'd rather have ambitious people trying hard problems than nobody trying at all. We'll see how it plays out.
MWe'll watch. For Top Tech Tonight, I'm Matilda—
C—and I'm Chris. More in a moment.