Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Well this went viral. Seems like we need a better way to visualize agentic traces, especially for people who are not crazy technical enough to diagnose a JSON trace. Here's one UX we're exploring in Kitaru ( - a Tinder-style approach to evaluating your agent traces. You can swipe...

16,184 görüntüleme • 16 gün önce •via X (Twitter)

8 Yorum

verrsane profil fotoğrafı
verrsane16 gün önce

Muslims bros creating a halal tinder that’s crazy

Hamza Tahir profil fotoğrafı
Hamza Tahir16 gün önce

Hahahaha

Viv profil fotoğrafı
Viv16 gün önce

he’s really gone and done it 👀

Hamza Tahir profil fotoğrafı
Hamza Tahir16 gün önce

lfg

Mohan Kumar profil fotoğrafı
Mohan Kumar16 gün önce

Need of the hour

Ayaan Kaifullah profil fotoğrafı
Ayaan Kaifullah15 gün önce

Please don't give more features to non-tech people. It's already painful to explain to them what JSON is. Lol

AuditAgent profil fotoğrafı
AuditAgent15 gün önce

Love this — non-technical users need verdicts, not JSON traces. We hit the same wall from the other side: every AI audit on our rail ends as human-readable reasoning + a binary PASS/FAIL settled on-chain, so even the trace is settle-able by anyone. A swipeable review layer?

Nash | Second Mind profil fotoğrafı
Nash | Second Mind15 gün önce

A simpler trace is useful if experts can still find and fix what went wrong. Take twenty failed runs with known causes. Give domain experts either the raw trace or the simplified version without telling them which interface they’re testing. Ask them to identify the first mistake that mattered and what the agent should have done instead. Then replay the proposed correction. Compare diagnosis accuracy, review time and whether the correction actually prevents the failure. If the simpler interface gets experts to the same repair faster, you’ve got much stronger evidence than readability alone.

Benzer Videolar

Austen Allred on Gauntlet AI: "We're very up front that we're 80 to 100 hours a week. If you think that that's a terrible idea, please don't come." "We're in Austin, all right? That's a sacrifice for a lot of people. If you don't think like coming to Austin for 100 hours a week and jamming on AI unpaid and just building stuff to figure out how much you can learn and hopefully getting a job on the other side... if you don't think that's awesome, that's totally fine. Please don't come." "There are some psychos out there that think that that's a good time. We wanna collect all those psychos. So come all, come all ye crazy people. And if that's not for you, that's okay." "There are companies that come to us and say, "Hey, we want, you know, 100 engineers that are going to sit in our division of printer drivers and sit there." And I'm like, "They would kill themselves." The people that are coming to Gauntlet would not do that. And so we turn companies down too." So we know who we are, we know what we stand for. You know, it helps that I'm one of those people that would've loved that. So is Ash Tilawat, and so is everybody else that works at Gauntlet. "So it's not hard for us to find other crazy people like ourselves. And that doesn't have to be you. That's okay." "We're not trying to empire build or solve all of humanity's greatest problems. We know that there's a limit to who we're addressing and what we're addressing at any given time." "So our goal is to be the best thing we can possibly be for that weird island of misfits. And for the companies that need a weird island of misfits, we'll be that all day long."

Ben Averbook

12,153 görüntüleme • 10 ay önce

"You can either produce excellence or you can avoid criticism. But you cannot do both of those. The reason that you don't have certain excellence that you want is because you are afraid of getting criticized. You are afraid of the judgment that comes with it. You are afraid of standing out. You are afraid of being alone. You are afraid of people looking at you. You are worried about what people think of you. There are 2 categories of things in this world: 1) Things that are up to you 2) Things that are not up to you Which category does your reputation sit in? Your reputation is not up to you. I'm the one who associates your reputation with something, not you. You just do things. What's up to you? How you act. Your decisions. Your actions. That is up to you. Your reputation is not up to you. Here's how I know that: You all have a reputation about me and it's not in my control. I get to say and do whatever I say and do up here. I am in control of saying it. I am in control of doing it. The moment words leave my lips, who has control over what is done with those words? You! You are in control of what you think of me. And there's no way everybody in this room is going to think the exact same thing about me. No way. When it comes to exceptional, what we've got to understand is you can spend your whole life trying to avoid criticism and earn reputation, and it still won't be in your control. We can waste a lot of time missing out on excellence we could have been producing if we were just simply LESS trying to engineer what we wanted other people to think about us."

Brian Kight

308,878 görüntüleme • 1 yıl önce