DeepSeek, every mention
142 scenes (2025), the whole family · ← back to DeepSeek
tap a year for its mentions
every year 2025 anyone John Coogan 186Jordi Hays 40Jordan Schneider 7Will Brown 4Tae Kim 4Jordy (Geordie) 4Dwarkesh Patel 4Bill Bishop 3Sholto Douglas 2Marc Andreessen 2
Verbatim, from the transcripts: the passages where DeepSeek comes up
OpenAI Sets Sights on $750 BILLION Valuation
Disney Partners with OpenAI, Breaking Down SpaceX’s IPO Plans, Oracle Slips | Diet TBPN
- ▶ 22:40 John Coogan This is the new main catch of DeepSeq 3.2.
Nvidia’s China Shift, Meta AI Drama, Paramount's Pitch | Diet TBPN
- ▶ 4:14 John Coogan The deep seek teams figuring out how to get some blackwell apparently.
Ilya Sutskever on Dwarkesh Patel Reaction, NVIDIA vs Google, Trump Unveils Genesis | Diet TBPN
- ▶ 12:37 unnamed speaker The TPUs equal bad for Nvidia Take is up there with the dumbest, maybe worse than DeepSeek, as it completely misses what actually happened in the last six weeks.
Thursday's Diet TBPN
- ▶ 12:35 unnamed speaker In a month-over-month change in total visits to leading Gen AI tools, ChatGPT is at the bottom of a list that includes Gemini, DeepSeek, Perplexity, Grok, Claude, Copilot, and Meta AI.
Microsoft CEO Satya Nadella Live on TBPN
- ▶ 27:25 unnamed speaker Well-timed Jevons Paradox post, by the way, back during the, the Deep Seek moment.
Weekly Recap: Matthew McConaughey on AI, Nvidia x OpenAI
- ▶ 43:34 Jordi Hays Deep Seek.
Weekly Recap: Sam vs. Elon, Cluely's Crazy Ad, UFC x Paramount, Apple's AI Robot
- ▶ 1:01 John Coogan First, in January of 25, DeepSeek reached number one overall in the App Store, similar to how Ramp is our number one sponsor on this show.
- ▶ 9:21 Jordi Hays Musk's Apple antitrust claim is undermined by apps like DeepSeek and perplexity, reaching number one in 2025.
BREAKING: Sam Altman VS Elon Musk
- ▶ 1:00 John Coogan First, in January of 25, DeepSeek reached number one overall in the App Store, similar to how Ramp is our number one sponsor on this show.
- ▶ 9:20 unnamed speaker Musk's Apple antitrust claim is undermined by apps like DeepSeek and perplexity, reaching number one in 2025.
Weekly Recap: Apple Sitting on $100B, Disney Enters The AI Race, Marc Andreessen, All About GPT-5
- ▶ 5:37 John Coogan And I also want to be, be able to use deep seek and I want to be able to call all these APIs within my, within my product.
- ▶ 1:01:36 Jordi Hays Uh, it was reported yesterday by the misinformation that, uh, reflection, reflection, uh, the misinformation reported that reflection, uh, one-year-old startup, uh, we've had, uh, the CEO on before he's in talks to raise one billion to… 2 times in the scene
- ▶ 1:02:53 unnamed speaker The model's way better than Deep Sea Guard one.
- ▶ 1:25:01 Marc Andreessen What they don't have is they don't have open data, right?
Marc Andreessen on Why Apple Might Miss the Next Platform Shift
- ▶ 19:18 Marc Andreessen And then what you don't, what you don't have, what, what most, what most of the open source, uh, open weights models don't have, including, you know, deep seek specifically,
Microsoft’s Cloud Outpaces Its AI Growth
- ▶ 5:32 unnamed speaker And I also want to be, be able to use deep seek.
Weekly Recap: Casey Neistat, OpenAI Cracks Math, The Future of ChatGPT, Apple x F1, Intel Layoffs
- ▶ 17:20 Jordi Hays Remember they, after, what did they release after the deep seek moment?
OpenAI Just Cracked the World’s Toughest Math Challenge — Here's How
- ▶ 15:44 unnamed speaker Remember they, after what did they release after the deep seek moment?
Weekly Recap: CEO Affair at Coldplay Concert, Elon Musk's AI Girlfriend, Cognition Acquires Windsurf
- ▶ 2:00:03 John Coogan I won't attribute to him because I might botch it, but it was basically that, uh, China banned high frequency trading and the dividend of that was deep seek and all this brilliant AI research.
DWARKESH PATEL on the Biggest Problem of AI
- ▶ 17:54 Dwarkesh Patel Uh, and, you know, who knows what the labs are doing, but we do have open source research from companies like Deep Seek, which does seem to be at the frontier or close to the frontier.
Weekly Recap: The Soham Parekh Drama, Meta Attacks OpenAI, Trump's Mega Bill
- ▶ 2:20:05 John Coogan Oh, deep seek.
- ▶ 2:34:18 John Coogan And when you talk to Dylan Patel, I was asking him about deep seek and how deep seek was, was, was doing all these innovations and like moving to floating .8, like smaller, uh, smaller, basically smaller numbers, storing the weights. 4 times in the scene
- ▶ 3:00:46 Will Depue Of course, from the like research open source world, there was a deep seek as like the RL craze taking off, but like, I mean, I'm, I work on RL, so I like obsess over it and like think about it a lot, but I do think we're really starting…
Apple Considers New AI partnership
- ▶ 7:02 John Coogan But I mean, I, I think the, I think the fact of the matter is like, uh, two billion dollars is enough to do a GPT four level run to do a deep seek style run.
- ▶ 7:11 John Coogan And so you, if you take all the best practices from deep seek V three and R one, like the reasoning model, and you spend the money to go do the training and you pull all the data together and, and maybe you're not completely stealing the…
- ▶ 7:11 John Coogan And so you, if you take all the best practices from deep seek V three and R one, like the reasoning model, and you spend the money to go do the training and you pull all the data together and, and maybe you're not completely stealing the…
Meta Poaches OpenAI Engineers
- ▶ 34:14 unnamed speaker And, and when you talk to Dylan Patel, I was asking him about Deep Seek, and how Deep Seek was, was, was doing all these innovations, and like moving to floating .8, like smaller, uh, smaller, basically smaller numbers storing the weights. 3 times in the scene
Weekly Recap: Tesla Robotaxi, Mr. Beast's AI Tool, Yacine, The Social Network 2, Neuralink Update
- ▶ 36:20 Yacine Mtibaa Like even if both of them are down, like I can hit deep seek over open router, right?
Weekly Recap - Elon Vs Trump, Ukraine's Drone Attack, Cluely Update & OpenAI CRO
- ▶ 1:31:05 Sholto Douglas Um, I think you have like deep seek V three and this kind of stuff like R one, um, which means that with that's like at least two rooms just to get to the scale of GPT four and GPT four was two years ago.
- ▶ 1:31:05 Sholto Douglas Um, I think you have like deep seek V three and this kind of stuff like R one, um, which means that with that's like at least two rooms just to get to the scale of GPT four and GPT four was two years ago.
We Went to DC to Interview Tech's Power Players
- ▶ 8:05 unnamed speaker I think there's incredible kind of capex and development, like the open source, deep seek and all that stuff will be great.
Is America Losing the Tech War with China? | Jordan Schnaider on TBPN
- ▶ 7:36 unnamed speaker And you have Deep Seek, which is basically an American model, or Manus, which is clawed, like, vended into China.
Why Humor Is the True Test of AI Intelligence | Will Brown on TBPN
- ▶ 5:33 Will Brown Um, I don't know, like, one I think is, like, people should do some math and math about, like, how long until big GPUs are readily available, and it's, it's gonna, gonna be a while before people can really run, like, even DeepSeek R-One…
- ▶ 15:21 John Coogan What, uh, now that we're like a couple of months out from the deep seek moment, uh, what is your takeaway? 3 times in the scene
- ▶ 15:42 Will Brown Like, I think a lot of their open source releases have really, like, like, since R-one, they've released a lot of code and details on their inference stack and training stack.
- ▶ 16:20 Will Brown Like, that's, whereas, like, the Deep Seek chat is literally free anywhere.
- ▶ 18:25 Will Brown Deep Seek did early, and then OpenAI has been doing lately, where they're really kind of betting, like, OpenAI seems to be essentially betting on scaling up RL as the path, um, and that it's not just longer reasoning, but it's reasoning…
How Tech Can Save America | Delian Asparouhov on TBPN
- ▶ 4:33 Delian Asparouhov DeepSeq becoming the, like, default open source model that the rest of the world uses that can't afford the, like, you know, you know, chat GPT and can't afford, you know, sort of alpha fold.
Building GPUs Is Easier Than You Think | Aaron Ginn on TBPN
- ▶ 19:44 unnamed speaker Deep Seek. 3 times in the scene
Google's AI Comeback in Their Own Words - Logan Kilpatrick
- ▶ 13:58 unnamed speaker Obviously Google's already doing the TPU, but in terms of just, uh, maybe even just squeezing extra, uh, performance out of inference, we saw this with deep seek. 2 times in the scene
Why Industrialization WON'T Create New JOBS | Samuel Hammond on TBPN April 8th
- ▶ 23:10 unnamed speaker Uh, how are you thinking about, uh, the Deep Seek versus Meta's Lama strategy? 6 times in the scene
The Most UNDERRATED Use Case in AI | Jeff Huber on TBPN April 8th
- ▶ 4:58 Jeff Huber Now, obviously like, you know, the DeepSeq R-ONE launched a few months back, like totally took everybody from surprise.
What AI and LEGO Have in Common | Flo Crivello on TBPN April 7th
- ▶ 10:34 unnamed speaker Uh, what are you, what was your reaction to Deep Seek?
Are AI AGENTS Learning to DREAM? | Shawn Wang on TBPN April 4th
- ▶ 21:59 Shawn 'swyx' Wang Um, but for example, one of the R one replications that did
Mike Knoop (Arc Prize) on Why Scaling AI Won’t Get Us to AGI
- ▶ 6:01 Mike Knoop Um, and basically it resets the baseline back to zero percent language models, uh, pure LLM systems are scoring like zero percent now, uh, again on, uh, on arc B two, um, single COT systems like R one and O one score like one percent.
Dwarkesh Patel (Dwarkesh Podcast) on The Scaling Hypothesis, AI and China
- ▶ 2:53 Dwarkesh Patel When DeepSea came out, they were like, oh, it's gonna be Jevin's Paradox, and, uh, we'll be using them way more in that they're cheaper.
- ▶ 10:15 Dwarkesh Patel And it sounds like after the deep seek moment, that sort of changed, at least in the AI, uh, at least in AI, because now the state, you know, the city funds and whatever are more willing to push in. 2 times in the scene
Aidan McLaughlin (OpenAI) On ChatGPT, model welfare, and path to AGI
- ▶ 2:56 Aidan McLaughlin Um, but the funny thing was like the, you know, when deep state came out, it was kind of this like really interesting moment where a lot of people hadn't tried or like, you know, our most recent models, a lot of people hadn't tried our…
Cristóbal Valenzuela (Runway) Talks About The Future of Media and Hollywood
- ▶ 19:20 unnamed speaker Uh, and then there's like the deep seek approach, which was like, 2 times in the scene
Jordan Schneider (ChinaTalk) On the US-China Power Struggle
- ▶ 10:04 Jordan Schneider I gotta, I gotta take for you that I'm parroting from Alvin Wong who gave it to me the today at breakfast America created deep seek. 6 times in the scene
- ▶ 25:19 Jordan Schneider I, I think it's kind of like inhuman almost to cut that off, but the sort of hope and expectation is that like America is just a better place to live and folks will want to stay here and more like America will gain more from that exchange…
Scott Wu on Cognition, Devin, and the Evolution of AI Agents.
- ▶ 13:27 unnamed speaker How did you process the DeepSeek moment? 2 times in the scene
Joshua Steinman talks cybersecurity and the vulnerabilities of a digital world.
- ▶ 8:23 Jordi Hays Uh, what, what's been your read on our response to Deep Seek broadly?
- ▶ 16:32 John Coogan I think DeepSeek's taken one of those. 2 times in the scene
Michael Mignano on Building Anchor, Social Podcasting, and the Future of Oboe.
- ▶ 23:51 Michael Mignano I mean, I think, you know, I think a few weeks ago, a lot of people wanted to have the debate about whether or not, you know, deep seek meant that models were commodities and nobody should ever train foundation models again.
Sam Lessin on Meta, Artificial Intelligence, and Investing in Creators.
- ▶ 7:39 Sam Lessin So do you think, uh, and if they're, and if it's not a capital game, it's a deep seek open source thing, they're also going to lose.
How Does GROK Compare? (Full Analysis)
- ▶ 0:12 John Coogan Across math, science, and coding benchmarks, Grok III beats Google Gemini, DeepSeq VIII, Anthropic Cloud, and GPT's four
- ▶ 2:41 John Coogan He says the impression overall I got here, I got here is that this is somewhere around O one pro capability and ahead of deep seek R one. 2 times in the scene
- ▶ 5:18 John Coogan That's not surprising, giving, given it is what scaling laws predicted, but it is comforting as there was a question as to scaling, as to whether scaling laws had hit a wall, and people were saying this during the, uh, deep seek, uh, kind…
- ▶ 12:25 John Coogan On the other hand, this is why deep seek matters.
- ▶ 20:20 John Coogan DJI, Unitree, Tesla, Anderle, DeepSeek, Jane Street, Palantir, social media platforms, uh, Meta, vertical AI companies, XAI, DeepSeek, Google, 2 times in the scene
- ▶ 28:02 John Coogan Like if open AI had not launched deep research and like the reasoning model, uh, would deep seek have been able to come up with it?
- ▶ 31:36 John Coogan Uh, with Deep Seek, and it was born again on February 17th, less than a month.
- ▶ 33:22 John Coogan The top OpenAI thinking models, O-one Pro get it too, but all of DeepSeek R-one, Gemini, two point O-flash thinking and Claude do not. 2 times in the scene
Elon Musk vs. Sam Altman Continues (WHO’S WINNING?)
- ▶ 3:13 unnamed speaker It's like, okay, deep seek launches. 2 times in the scene
The BEST Super Bowl Commercial (Saquon Barkley)
- ▶ 37:42 John Coogan That's of course, deep seek the free to use model was built at a fraction of the cost. 4 times in the scene
- ▶ 45:07 Jordy (Geordie) They may anthropic made a clear bet around AI safety and a lot of other firms ignored that deep seek one of them. 4 times in the scene
What is PALANTIR? (And Why it's RIPPING!)
The Truth About Chinese Manufacturing (De Minimis)
- ▶ 14:44 Jordi Hays So over the week, over the last week, it's been, if you talk about deep seek, you're gonna get a lot of, um, views.
The Future of Artificial Intelligence
- ▶ 0:00 unnamed speaker So, uh, Dylan Patel writes, the Deep Seek narrative takes the world by storm. 6 times in the scene
- ▶ 1:24 unnamed speaker Is that deep seek is so efficient that we don't need more compute. 6 times in the scene
- ▶ 4:15 unnamed speaker uh H 800 has the same compute as the H H 100 but less network bandwidth uh the misleading six million figure was widely circulated as the cost of v three but that doesn't include R and D hardware depreciation uh repeated experiments actual… 3 times in the scene
- ▶ 5:30 unnamed speaker And so in terms of model performance, DeepSeq V 3 times in the scene
- ▶ 5:32 unnamed speaker R three is the base model while R one is the reasoning model. 2 times in the scene
- ▶ 9:40 unnamed speaker So, you know, V three surpasses GPT four Oh, but GPT four
- ▶ 10:19 unnamed speaker And so let's, let's break down the direct comparison between R one and O one, uh, deep seeks model versus chat GPT. 7 times in the scene
- ▶ 10:19 unnamed speaker And so let's, let's break down the direct comparison between R one and O one, uh, deep seeks model versus chat GPT. 13 times in the scene
- ▶ 15:54 unnamed speaker And so, uh, well, there was a frenzy of hype for R one, a 2.5 trillion dollar us company released a reasoning model a month before for cheaper. 6 times in the scene
- ▶ 18:57 unnamed speaker Uh, none of this distracts from DeepSeek's remarkable achievements. 4 times in the scene
- ▶ 19:50 unnamed speaker These training pre and post V three utilizes multi-token prediction at a scale not seen before. 2 times in the scene
- ▶ 20:43 unnamed speaker People just went through the deep seek paper and everything that they were doing, they assumed was novel. 5 times in the scene
- ▶ 22:21 unnamed speaker Uh, in terms of R-one, it benefited immensely from having a robust base model, V-three. 2 times in the scene
- ▶ 27:44 unnamed speaker And so, a hundred and forty billion subsidy, AI subsidy, after meeting DeepSeek's founder. 3 times in the scene
- ▶ 29:57 unnamed speaker DeepSeq lags top labs, but closes gaps swiftly with new paradigm reasoning. 3 times in the scene
- ▶ 30:21 unnamed speaker Do something to bring this to the masses so you can cut, stem the bleeding of people installing the Deep Seek app.
What Open AI Got WRONG
- ▶ 0:04 unnamed speaker Um, for those that don't know, this is an American company that competes with, um, a Chinese company called deep seek. 5 times in the scene
- ▶ 7:58 unnamed speaker What's interesting is like, so he goes on to talk about the O-one chain of thought and how they, when they launched O-one in, in comparison to R-one, R-one tells you everything it's thinking and just dumps it out and it's open source. 5 times in the scene
- ▶ 9:23 unnamed speaker That was the, the, the only UI change that people cared about or talked about with the deep seek long launch was that exposed chain of thought.
- ▶ 14:29 unnamed speaker With DeepSeek, if I recall correctly, a lab in Berkeley read their paper and duplicated the claimed results on a small scale within a day. 4 times in the scene
- ▶ 19:13 unnamed speaker One of the big R one takeaways is that you can infuse reasoning into normal LLMs. 3 times in the scene
The Character AI Controversy (Role Play)
- ▶ 4:17 unnamed speaker We see that with the deep seek thing where like the models are very mature, very robust, certainly good enough to have a conversation with you as a romantic partner, or even if they just want to role play as George Washington, right?
- ▶ 17:10 unnamed speaker Like, people are gonna be able to fine-tune DeepSeek on this stuff very easily and create just, like, magnet links out there on the internet where you just download it, run it locally, and
DeepSeek Update, Market Crash, Timeline in Turmoil, Is VC Cooked, Zero Cope Policy
- ▶ 0:04 John Coogan We are staying on Deep Seek.
- ▶ 1:02 John Coogan This went out on January 25th and, uh, was a very, very large deep dive on, um, how DeepSeq and their R-One model might change the demand for GPUs, specifically Nvidia GPUs, and, uh, we have a summary article here, but we'll take you…
- ▶ 1:02 John Coogan This went out on January 25th and, uh, was a very, very large deep dive on, um, how DeepSeq and their R-One model might change the demand for GPUs, specifically Nvidia GPUs, and, uh, we have a summary article here, but we'll take you…
- ▶ 12:24 John Coogan He calls it COT models, reasoning models introduced in the past year, most notably in open AI's flagship O one model, but very recently in deep seeks new R one model, which we will talk about later in much more detail. 2 times in the scene
- ▶ 12:24 John Coogan He calls it COT models, reasoning models introduced in the past year, most notably in open AI's flagship O one model, but very recently in deep seeks new R one model, which we will talk about later in much more detail. 2 times in the scene
page 1 of 2 · 100 scenes per page · newest episode first next →