ARC-AGI, every mention
45 scenes, the whole family · ← back to ARC-AGI
tap a year for its mentions
every year anyone Mike Knoop 26John Coogan 26Jordi Hays 3Will Brown 1
Verbatim, from the transcripts: the passages where ARC-AGI comes up
Astra Reactions, Jobs Print, Cybercab Launch | Diet TBPN
- ▶ 0:15 unnamed speaker We talked about Arc AGI. 2 times in the scene
Model Mayhem, GPT-6 Astra, Why Nvidia Bought Hugging Face | Diet TBPN
- ▶ 20:59 Jordi Hays Arc AGI three, 98.6. 2 times in the scene
- ▶ 27:08 John Coogan The team over at Arc AGI is working on VIV. 2 times in the scene
Leopold Stays in the Game, Big Tech Earnings, OpenAI Slashes GPT-5.6 Prices | Diet TBPN
- ▶ 24:33 unnamed speaker Uh, anyway, uh, we also, I don't think we touched on this, but Arc AGI VIII, the leading labs have been going back and forth. 6 times in the scene
Model Mayhem: OpenAI’s 5.6 and Meta’s Muse Spark 1.1 | Diet TBPN
- ▶ 2:58 unnamed speaker The most interesting benchmark to me has always been Arc AGI V three. 5 times in the scene
AI Profits Surge, $70B Capex Surge Raises Investor Concerns, Automated Skin Exams | Diet TBPN
- ▶ 29:14 unnamed speaker Put out an update on ArcGi VIII, which, Tyler, are you still on the human leaderboard for that? 4 times in the scene
Meta’s AI Comeback Moment, Claude Mythos | Diet TBPN
- ▶ 7:57 unnamed speaker Um, it underperforms on Arc AGI too, as you mentioned, but this maybe is the, the bull case here is that they have at least moved on from the culture of like optimizing for the benchmarks.
Novartis Buys Excellergy for $2B, Anthropic Vs. Pentagon, The Mansion Section | Diet TBPN
- ▶ 9:36 unnamed speaker I want to know how it will do an Arc AGI V three, because all of the models are under one percent completion on that benchmark.
Benchmark’s Ship of Theseus, OpenAI Kills Sora, SpaceX $2T IPO Buzz | Diet TBPN
- ▶ 29:50 John Coogan I mean, the, the, the, the true frontier, the, the models that are, are, are dominating Arc AGI, those command a premium, a price premium.
Anthropic’s Trust Nuke, OpenAI’s new releases, Google Claims AI Crown | Diet TBPN
- ▶ 14:58 unnamed speaker Uh, yeah, I think it was, uh, Arc AGI v two.
xAI Raises $20B, Anthropic’s $350B Valuation, OpenAI Launches ChatGPT Health | Diet TBPN
- ▶ 13:32 John Coogan If you're wondering whether saturating Arc AGI one or two means we have AGI now, I refer you to what I said when we launched Arc AGI two last year, which is also the same thing I said when we announced Arc AGI two was coming. 4 times in the scene
- ▶ 13:49 John Coogan The Arc AGI series is not an AGI threshold. 2 times in the scene
GPT-5.2 Reactions, Jacob Elordi vs AI, Disney x OpenAI Deal Breakdown | Diet TBPN
- ▶ 18:33 unnamed speaker Arc AGI one has been out for six years. 5 times in the scene
Google Gemini 3 Reactions, Google Antigravity, Anthropic-Nvidia-Microsoft Deal | Diet TBPN
- ▶ 0:36 unnamed speaker Obviously the big news, at least in, from my reading on it is that Gemini three performs very well on Arc AGI V two, uh, a huge jump twice the performance of the previous state of the art. 2 times in the scene
- ▶ 6:38 unnamed speaker Arc AGI is up at 31%, over 13, 17. 2 times in the scene
Weekly Recap: Apple Sitting on $100B, Disney Enters The AI Race, Marc Andreessen, All About GPT-5
- ▶ 1:48:19 John Coogan Arc AGI. 4 times in the scene
- ▶ 1:48:56 unnamed speaker ArcGIV one and V two, right on both V three V three. 4 times in the scene
The State Of AI in 15 Minutes
- ▶ 7:11 unnamed speaker Arc AGI. 10 times in the scene
Weekly Recap: Casey Neistat, OpenAI Cracks Math, The Future of ChatGPT, Apple x F1, Intel Layoffs
- ▶ 11:55 Jordi Hays The three K task arc AGI, 80% could probably be 30 dollars if we cared to optimize it.
OpenAI Just Cracked the World’s Toughest Math Challenge — Here's How
Weekly Recap: CEO Affair at Coldplay Concert, Elon Musk's AI Girlfriend, Cognition Acquires Windsurf
- ▶ 12:23 John Coogan They're like, look, Arc AGI, the insiders, inside, the most insider AI test you can possibly have.
ELON MUSK's Grok New Vulgar Anime Mode
- ▶ 1:11 unnamed speaker They're like, look, Arc AGI, the insiders inside, the most insider AI test you can possibly have.
Weekly Recap: Grok 4 Launch, Texas Floods, Web Browser War, Top Signals, Meta Smart Glasses
- ▶ 1:18:55 unnamed speaker So, uh, Mike Newp over at Arc AGI says, zooming out on Arc Progress, I'd say OpenAI's O-series progression on V-one is a bigger deal than Grok's progression on V-two so far. 3 times in the scene
- ▶ 1:33:41 unnamed speaker Arc AGI.
- ▶ 1:43:14 unnamed speaker People are really pushing this idea of like, okay, like, like we are accelerating, like the, the AG, uh, the Arc AGI leaderboard is accelerating, but I keep seeing this and, and feeling deceleration. 2 times in the scene
ELON MUSK's New Grok 4 Is Here! - We Break it Down
- ▶ 4:05 unnamed speaker So, uh, Mike Noop over at ArcGi says zooming out on Arc progress, I'd say open AIs 3 times in the scene
- ▶ 18:52 unnamed speaker Arc AGI.
- ▶ 28:28 unnamed speaker Like, the, the AG, the Arc AGI leaderboard is accelerating, but I keep seeing this and, and feeling deceleration. 2 times in the scene
Weekly Recap - Elon Vs Trump, Ukraine's Drone Attack, Cluely Update & OpenAI CRO
- ▶ 1:19:30 John Coogan Obviously opening eyes made incredible progress there, but 4 times in the scene
- ▶ 1:55:02 John Coogan Uh, talk to me about ArcGGI. 2 times in the scene
Why Humor Is the True Test of AI Intelligence | Will Brown on TBPN
- ▶ 9:48 Will Brown I would have, I did some tests on like RKGI problems where you give the screenshot to O three and it can like basically figure it out in 10 minutes of like zooming in and looking at the thing and then writing some code to see if it like…
This Company is Changing Cybersecurity | Alex Dhillon on TBPN April 10th
- ▶ 10:51 unnamed speaker Uh, you could imagine Arc AGI puzzles becoming the next CAPTCHA, uh, or, or, or in your view, do we need just to stay, is it, is it a, uh, probabilistic on probabilistic combat from here on out?
Mike Knoop (Arc Prize) on Why Scaling AI Won’t Get Us to AGI
- ▶ 1:26 Mike Knoop Uh, this is how I rediscovered, uh, Francois Cholet's, uh, ARC benchmark that was originally published in 2019. 3 times in the scene
- ▶ 2:37 Mike Knoop Especially ArcGGI one. 8 times in the scene
- ▶ 5:34 Mike Knoop Right now, I think ARC AGI two that we just introduced this week on Monday, uh, is pretty direct evidence that we need new ideas still to get to AGI. 4 times in the scene
- ▶ 10:37 Mike Knoop And even there, like within the sort of same efficiency and, uh, you know, accuracy specs, we still haven't seen any frontier AI system beat, uh, RKGI two level and RKGI one.
- ▶ 13:20 Mike Knoop Um, and this was kind of what led me to sort of originally getting a bit interested in arc is, uh, you know, it's, it's this reliability problem. 4 times in the scene
- ▶ 16:32 Mike Knoop Uh, one of the things that surprised me about Arc AGI two was the relative degree of ease by which we were able to create it.
- ▶ 17:43 Mike Knoop That's what RKGI shows. 2 times in the scene
- ▶ 27:50 Mike Knoop Uh, you know, I think one of the reasons I personally find things like arc more useful, um, is they provide a direct way to do sense finding on this stuff. 3 times in the scene
The Future of Artificial Intelligence
- ▶ 15:28 unnamed speaker Um, but it's on Arc AGI, SWE Bench, and Frontier Math, Advanced Mathematics, like, things that most consumers might not even care about the benchmarks on, but are potentially good proxies.
DeepSeek Update, Market Crash, Timeline in Turmoil, Is VC Cooked, Zero Cope Policy
- ▶ 19:53 John Coogan And it just thinks for hours and hours and hours basically, and it was able to break arc that, uh, that AI, AGI eval.
Year of Tech in Review (ep25)
- ▶ 1:03:02 John Coogan Let's go to O three, Mike Noop who runs arc, 4 times in the scene
Who's Getting Rich Off Ozempic? (e12)
- ▶ 3:11:02 John Coogan Uh, a product called Keras, uh, was deeply involved in machine learning, and also had a great appearance, star appearance, great performance on the Dwarakash Patel podcast, and, uh, and is also involved in a very interesting AI benchmark…