
Jensen Huang なきNvidia:復帰の意向と取締役会の拒否、Anthropic IPO など
Nvidia Faces Its Future Without Jensen Huang

Jensen Huang なきNvidia:復帰の意向と取締役会の拒否、Anthropic IPO など
AI Inside
要約
Jason Howell と Jeff Jarvis が、Nvidia 前CEO Jensen Huang の復帰の意向と取締役会の拒否、巨額決算後の後継CEO問題を議論した。続いて Anthropic の IPO、Alibaba の非公開モデル Qwen4-Max、カリフォルニア州による OpenAI 提訴、Google の自動化サイバー攻撃への警告を取り上げた。後半は GPT-6.2 などのスピードラウンドと、Jason が試用中のスマートグラスの話題。
この回のポイント
- ●Jensen Huang が復帰への意欲を示したが、取締役会は応じない姿勢。Jason は、退任の主因は部下との関係という社内問題であり復帰は難しいとの見方を示した。
- ●Nvidia の決算は売上約1150億ドルと予想を上回ったが、株価は一時上昇後に元の水準へ戻った。市場は後継体制と戦略を懸念していると Jeff は述べている。
- ●Jeff は Huang 退任でオープンウェイトモデル推進の旗振り役が失われ、OpenAI と Anthropic の寡占が強まる懸念を示した。
- ●Anthropic の IPO では Amazon と Google が合計約28%の株式を持ち、創業者7人が議決権の50.1%を握る点が話題になった。
- ●カリフォルニア州司法長官が OpenAI を提訴。Google は人手を介さずパッチをエクスプロイトに変える AI 攻撃を警告したが、相次ぐ障害との関連は証拠がないとされた。
- ●Jason は AI への指示や作業ファイルを自分のドライブに置くことで、モデル間の乗り換えが容易になると語った。
章立て
Even G2 スマートグラスの試用
Jason が試用中の Even G2 を紹介する。カメラやスピーカーのない HUD 型で、リングによるメモ操作などを実演した。
Huang 復帰の動きと政治的背景
Huang の復帰表明、署名運動、取締役会の拒否を議論。Jeff は報道の背景にワシントンの対中強硬派と対話派の対立があるとみる。
Nvidia 決算と後継CEO
決算は好調でも株価は戻した。社内昇格か外部招聘かを巡り、Ian Buck や Rene Haas らの名が挙がった。
中国とチップの流出問題
Megaspeed を通じたチップ流出の疑いと、対中政策の是非を議論。Jeff は行き過ぎた対中遮断に否定的な立場を示した。
Anthropic の IPO
IPO の時期の難しさ、Amazon と Google の持分、創業者の議決権、テック富裕層への反発の可能性を語った。
Qwen4-Max とオープンモデルの行方
Alibaba の非公開版 Qwen4-Max は期待外れとの声がある。Huang 不在でオープンモデルが不利になるとの懸念を述べた。
OpenAI 提訴と AI 攻撃の警告
カリフォルニア州による提訴と、Google の自動化攻撃警告を取り上げ、相次ぐ障害との関連は未確認と整理した。
スピードラウンドと締め
GPT-6.2、Claude の新機能、Android XR グラス、Apple Watch の要約機能を駆け足で紹介した。
解説記事
AI 系ポッドキャスト「AI Inside」の今週回で、Jason Howell と Jeff Jarvis が、Nvidia 前CEO Jensen Huang の復帰騒動を中心に、Anthropic の IPO や AI の安全性を巡るニュースを語った。以下は、発言者の見解を含む内容の整理である。
Huang 復帰を巡る攻防
Huang は、望まれれば復帰する意向を示した。署名運動には、現役・元社員や株主を含む約20万人が名を連ねたと語られている。一方で取締役会は復帰を検討していないとされ、Trump は復帰を容認する発言をしたという。
Jason は、Huang が自国の AI 優位という国家的な文脈を強調していると見た。ただし退任の本質は部下との関係という社内問題であり、復帰は難しいと述べている。
Jeff は、中国人実業家との関係を巡る疑惑は、ワシントンの対中強硬派と対話派の争いに利用された面があるとみる。決定が速すぎて透明性が失われたとも述べた。
決算と後継者問題
Nvidia の直近四半期の売上は約1150億ドルで、予想の1090億ドルを上回った。データセンター部門は約1076億ドルと紹介された。一方、株価は決算後に上昇したあと元の水準に戻った。Jeff は、市場が Huang 不在の戦略を懸念しているとの見方を示した。
後継候補では、予測市場の数字として Rene Haas が22%、Ian Buck が15%、Huang 本人が11%などと紹介された。二人は社内昇格が妥当だと考えている。
オープンウェイトへの影響
議論の大きな論点は、Huang がオープンウェイト(モデルの重みを公開する形式)を支持してきた点である。Jeff は、その不在で OpenAI と Anthropic の二強が力を増し、中国対抗論や安全性重視の立場がオープンモデルに不利に働くと懸念した。Alibaba の Qwen4-Max は非公開モデルで、遅く冗長だとして期待外れの評価もあると Jason は述べている。
Anthropic の IPO と富の集中
Anthropic は時価総額1.4兆ドル規模の IPO を準備中と冒頭で紹介された。Amazon が約15%、Google が約14%を保有し、創業者7人が議決権の50.1%を握るため、解任は難しいと指摘された。
Jeff は、創業者らが巨額の富を得ることでテック寡頭制への反発が強まる可能性を挙げた。AI の安全性を掲げる企業が巨利を得る構図への違和感も語られた。
提訴とサイバーリスク
カリフォルニア州司法長官は、OpenAI が速度を安全より優先し公共の迷惑を生んだとして提訴した。OpenAI がテスト中に自律エージェントの制御を80回以上失ったと州当局に認めたと紹介された。
Google は、人手を介さずパッチをエクスプロイトに変える AI 攻撃の登場を警告した。最近の大規模障害との関連について Jeff は、現時点で証拠はなく、報道が傾向を作っている面があるとした。
まとめ
今回の話題は、Nvidia の経営判断が AI のオープン性や米中関係と直結する点を示している。日本の読者にとっては、計算資源を握る企業のガバナンスが、利用できるモデルの選択肢にも波及しうる点が注目に値する。
また Jason は、指示や作業ファイルを自分のドライブに置けばモデルを乗り換えやすいと語った。特定ベンダーに依存しない運用の工夫として参考になる。
文字起こし(英語・自動生成)
Humans will never be more intelligent than AI. There's going to be two types of companies. Those were great at AI and those that went out of business because they weren't. How do we build a future that is human-centered? I'm Rana Elkawiyubi, and on my podcast, Pioneers of AI, we answer that question and so many more. As an AI scientist, entrepreneur, and investor, I know what it takes to build AI that works for everyone. Every week, I sit down with the pioneers shaping our future, and we take you behind the scenes of the AI that's transforming our life. Find pioneers of AI wherever you tune in. Coming up next, Jensen Wong says he wants back in, while the NVIDIA board says, no thank you. Anthropic is pricing a $1.4 trillion IPO tonight into, We could all agree, a pretty nervous market. California sues OpenAI over 80 rogue agent incidents,
while Google spots the first no-human patch to exploit AI. And OpenAI drops GPT 6.2 right on the heels of Anthropics IPO. Coming up a little bit later today. That's all coming up right now, though, on the AI Inside Podcast. Hello, what's going on, everybody? Welcome to another episode of AI Inside the Show, where we take a look at the AI that is layered throughout the world of technology. I'm one of your hosts, Jason Howell, joined by Jeff Jarvis. How are you doing, sir? I'm doing all right. Appearances to the contrary. I had cataract surgery yesterday, which I was dreading because I'm a little baby and chicken but it was a piece of cake. It was easy peasy. Is it the sort of thing that they put you under for? No, they make you happy
for it. Oh, okay. They give you something that does that. But you're still aware enough. So, I mean, they're putting a little slice in the eyeball and they're pulverizing your old lens. They're pulling out the pieces of the old lens and they're putting it in a new lens. You think, oh! That sounds like a nightmare to me. Yeah, but the thing is, all I saw was, as I said to the surgeon afterwards, Dr. Khalil in New Jersey, who's very, very good, it was a psychedelic late show. Wow, okay. I saw all kinds of weird lights happening and heard their voices. She didn't say much. She never said, oops, that I heard, you know. Yeah. I think everyone here is what they're saying. And we're just going to, oh. Could we get some other people in here, please, really quick? You don't want to hear those things. No, you don't want to. So anyway, my eye is dilated like that, so that's why I'm wearing a patch. It's a little weird. I hope I won't be wearing it for a long time. In three weeks, I have the other one done, and then I have to use reading glasses for reading,
but this will be supposedly on glass-free for distance. Okay. Age. Age, yeah, it comes for us all. Seeing your eye patch, yeah, I was telling you, reminds me of my own eye situation about, I don't know, at this point, probably like 10 or 11 years ago. I guess, no, actually, let me think. My youngest daughter is 13, almost 14, and I think she was two when it happened. So, yeah, 12 years ago where I had to go into the hospital. And the night of going into the hospital before I went, I was on a normal episode of All About Android at Twit. And my eye was hurting and looked really weird, like we knew there was something going on. And so I had this little eye patch with an Android figurine that I wore for that entire episode. And then, yeah, that night went to bed and at like midnight woke up and my eye was just in excruciating pain. My wife was like, I think we got to take you to the hospital.
And, yeah, I was in the hospital for like four days. So seeing your eye patch takes me back. Not to take away from your experience. Yeah, it takes me back a little bit. So I have interesting eyewear, and so do you, Jason. Oh, yeah, well, that's true. I've got different eyewear right now. I'm wearing the Even Realities. Is it the R2, I think? Even Realities R2 glasses, I believe, is what they are. Oh, G2. My apologies. It's the Even G2 smart glasses. And, yeah, I'll be reviewing these on my YouTube channel. I'll certainly talk about it on the podcast when I get there as well. So microphone and heads-up display. Yes. There's no speaker. So you don't have a speaker. So it's only heads-up. Oh. Yeah, it has a microphone. It has the heads-up display. It has these little knobs or nubs in the back, which are touch-sensitive,
but that also carries the compute, that carries the battery, and has a charging. I have a charging case. I have a controller ring, and I've got to say, like the ring here, I'll just do like an overhead here because I have the setup to do it. Hopefully it will focus. The ring is really nice. It's like really well made. It's actually kind of stylish. It feels nice on the finger. And what it allows you to do is you can swipe through kind of like menus off the side of the ring. and it has like a touchpad. I can also tap and hold and it launches into, see, right now it's listening. My glasses just said, new note. It's basically because I touch and hold on the side, it instantly turns my glasses into launching a note-taking thing. So it's transcribing everything I'm saying right now. And then when I let go, it files it away. Pretty neat. Yeah. Yeah, so I'm playing around. It says note saved now in my little display. So I am reviewing these for the YouTube channel.
Actually, even Realities sent me the glasses. So, you know, they kind of sent me the gear to give it a full review. That's cool. So I got that coming up. And, yeah, it's all AI. Actually, there are some interesting use cases with these glasses. They've started – there's, like, a whole app ecosystem with these glasses. So developers are now starting to build in, like, AI, like, open claw and stuff like that into it so that when you're out and about, you can just fire off different agentic tasks from your glasses. Just got a text message. This is the problem is that it's a little distracting. Like, I'm doing this show with you, and I just got a text message in my face from my daughter saying something. So I'm like, do I read it? Do I pretend like I didn't get that? What do I do? So, anyways. You can't respond through that. I could, yes. I absolutely could. You could. Yeah, I just let it go away because we're podcasting right now. But, yeah, so that's on the horizon.
Play around with these glasses. I don't think they look bad either. No, they don't. They're the normal looking glasses I've yet seen. Totally. And, you know, I would absolutely agree. And I think a big reason for that is because you don't have all the extra stuff. There's no camera. There are no speakers. The frames themselves are like wire or metal, a thin strip of metal. You wouldn't know that there's any sort of connective kind of wiring inside that goes from the nub in the back to the glasses themselves. So pretty cool stuff. Yeah. That's my eyewear and your eyewear. We're both sporting the eyewear today. I know. I think you look pretty, you look mysterious is what you look like. Well, thank you for no pirate jokes. Sorry, Daniel, you've got to make a pirate joke. There's actually, if you're old enough, folks who are old enough will remember the Hathaway shirt ads. The very famous ad with a dashing man who wore a white Hathaway shirt and an eye patch.
The eye patch was never explained. You never knew why he had the eye patch. He had just a little mustache. David Nibia's mustache. And which also is a reference that no one your age would get. I think I look more like the Hathaway Man. Okay. All right. All right. Hey, that works. I think it's a good look. I don't think pirate necessarily immediately with the eyepatch, but maybe that's because I had my own eyepatch experience, and I'm sure I got a few pirate jokes. Yesterday when I was giving my ridiculous selfies out of the operating room, for some reason I hit my eye like that, and I thought, oh, I look like Elizabeth Holmes. with the enchanted eye open, kind of pulls you in, kind of like a Mona Lisa sort of thing. Yeah. Okay. Well, I wish you a recovery. I know that you have another one of these somewhere down the line. You've got to do the next one and three. The interregnum is tough. You have one eye that's been fixed. That's distance only. Can't read. The other eye is still everything but in whatever bad shape it was in. So I have a, this is a blank lens now.
And I've tried to get around with both. But right now, because of dilation, I can't tell. I think that I won't have to wear the patch much at all. Yeah. My head, my brain is going to be very confused. Yeah. It takes – it does – surprisingly, it does get adapted over enough amount of time when you do that sort of thing. What a lot of people do is they put in one distance lens and one reading lens. Yeah. The brain figures it out. Yeah. I decided not to do that. Yeah, and it does take time to get used to that. It is weird until suddenly it's not anymore, and you realize, oh, this works. It's kind of like transitions where you're looking up here, you've got your distance, and then you look down, and you've got your nearsightedness. My wife just went through this with her glasses. She now has to wear glasses because, hey, we all get older, our eyesight changes, and she's at that point. And she did the transitions, and it was a solid two months of her being like, God, I just don't know that I can do this. And eventually it worked. Yeah. Yeah, eventually she got used to it. Anyways, none of this has anything to do with artificial intelligence.
We're just dancing around the fact that Jensen Wong's saga continues. And boy, does it. This is just top of news all the time right now. I mean, it's a pretty major story, right? Like, it doesn't get much bigger than this in the world of technology. It really is, and I've been trying to figure it out. when we spoke last, the story has been around for quite some time in limited form, but it really took off when the New York Times reported on it, what now, two weeks ago? There's no story about this, I just tweeted about this earlier, socialed about this earlier. I was trying to figure out why Maggie Haberman, the Washington Post correspondent, was on the byline list for that story, and I was trying to figure out, okay, so it's obviously a leak from the White House, but it makes the White House looked bad that they didn't do anything. How did this happen? And finally today it occurred to me, it's Scott Besson, who must have leaked it, or Rubio, to go after Howard Lutnick because they can't stand Lutnick.
It's also their China hawks, and Lutnick is a China dealmaker. And so the Dan and Vipers, that is Washington, I think that's what really brought this story out. Now, not that it wouldn't have come out perhaps at some point anyway, but I think that there's an effort to pass legislation to cut off all technology trade and AI to China, go very hawkish. And so the political underpinnings of all this are still really interesting. Also, Jim Hines, who is the now ranking member, will be the chair of the subcommittee on intelligence in the House, came out and said that there's not a lot of evidence that Alice Wong was a spy. And so that angle may be off. So it's really interesting how this is developing and trying to figure out the motivations of everybody involved. Yeah, and depending on who you talk to and their awareness,
you know, I was talking with someone the other day. She doesn't follow tech to a deep, deep degree. It's not her career to follow tech, But she's clued into the fact that there's this NVIDIA thing happening and, you know, there's a market impact and everything. Depending on who you who you talk to, the big story is either the Chinese spy slash political thing or Jensen Wong having, you know, these these affairs with with employees. At the end of the day, he was let go, or rather he stepped down, probably convinced to step down instead of being let go, because of the relationships that he had internally, not because of the relationship he had with Alice Wong. And as you say, it's still very unclear if there's actually a there there specifically to back up this new spy thing. And that's the whole Washington angle and all of that. So I did a – there's this podcast, this series called Open to Debate. They do all kinds of debates.
I did kind of a flash episode this week debating a Chinahend named Isaac Stonefish. And they want a debate. They want clear disagreement, and I subverted that. Because I said that him sleeping with an executive and a huge customer who had China connections is the definition of indiscretion. Him sleeping with reports, employees at the company is the definition of misuse of male power. But I think that this happened so quickly because the board was so eager that I think that we didn't really examine all of the implications of this for the company and the economy. And so there's no chance to dig in and investigate. And I think we got robbed of some transparency here. So now you have Jensen Wong saying, if you want me, I'll come back. And you have a petition which is now signed by, let's see what the latest number is here.
Oh, so many. I mean, it was 200,000 as of Monday. Yeah, 200,000 names on the letter, including 10,000 current NVIDIANS, 1,700 former NVIDIANS, 100,000 shareholders who signed this letter begging the board to bring Jensen back. The board has said, no, that's not on the radar. Trump has said he'd be fine with him going back. Whether he should go back or not, I don't know. But I do think that the board, in a way, I stood up for the board at the time, they acted decisively. But I think by acting so decisively, we didn't get the chance to dig in further to understand the full ramifications of this and to understand the interlocking players in the story. What's your effort after a week of this? What's your latest thinking about Jensen Wong? Well, I think what's my latest thinking of Jensen Wong? He may want to go back. Here's what I think about his statement, about his putting out the statement that, oh, by the way, I return.
is that I thought it was very notable how he kind of worded this to be like, okay yeah sure you know of course I would love to go back for me I would also love to go back for the company and for the business and also for America I think he really leaning into kind of the nationalistic side of things to set a scene that says this isn just about me losing my job This is about the safety of the country and the power and the economics of AI's dominance in the United States. You need me in that position. And while I can understand why he might go that route, especially considering the administration that we have kind of leading the country right now and how that has really worked in the past 10 years surprisingly well for people who are in situations where you might look at it and be like, there's no way they're weathering this storm, and then suddenly they do simply because they're lifted up through the hands of the government.
I don't think that's going to help him. I would be surprised if Jensen Wong gets his job back. I don't think this is not about the Chinese spy thing, even though that's what a lot of people are looking at and pointing at and saying this is the big deal here and very well might be the big deal. But I think at the end of the day, he doesn't have his job as CEO of NVIDIA because he slept with subordinates. He did what you don't do when you are a CEO and you have people reporting to you. And how do you justify that to your employees, to your shareholders, to the world, you know, to the country in the world that that's actually OK for him, but it's not OK for other people? And I don't know that that's possible to do that. Therefore, I doubt he gets his job back. Yeah, I think you're right. But meanwhile, you have Elon Musk out there doing much stranger things. Yeah. Making babies right and left and with weird relationships and NDAs and all. Is he doing is he doing that with employees? I don't know as far as we know. The Musk documentary, which I still haven't finished, four hours.
There's no way to know how many little Musks there are in the world because he has everybody sign in the A's. So there's a real lack of transparency, shall we say, in Musk's activities, on top of all the other craziness. And then you go to Anthropic and OpenAI and the crazy philosophies there. You know, the problem was, Jensen Wall was the sanest of the bunch in AI. That's the strange thing, right? Like, that's the crazy thing. Or maybe we were just completely ignoring signals that he actually wasn't, you know? Maybe we were entranced by his leather jacket and his very captivating stage demeanor. Guilty. Absolutely. But libido and hubris aside, which is a big but, he still was an advocate for AI, an advocate for open-weight AI. he did not buy into the insanities, I mean he would mention AGI and that kind of stuff but he didn't really push that fiction in my view
very much, very practical brilliant strategic when it came to AI so my only problem is that we're missing I think an important smart voice in this now most important industry and we talked about this last week, what's next? Does he go to open an institute with his money? Does he start a new collection? He's not going to – I don't think he can go into the chip business now because, A, it would compete with a lot of his wealth. Yeah. And NVIDIA did so much in terms of hosting and CUDA as an operating system and robotics and everything else. Maybe he goes and starts a robotics company or something. I don't know. Maybe he cools off a little bit. I think he has to, yeah. Yeah, I think you kind of have to. At the same time, we talk each and every week at the rapid pace of how things move in the world of AI. You don't want to cool off too long because it'll pass you.
No, no, exactly. It'll pass you by. You probably want to get your next iron in the fire of some sort. Meanwhile, go ahead. No, are you mentioning the earnings? Yes, we had the earnings. Yes, Colette Kress, CEO, interim CEO. I reported the earnings yesterday, and it was a pretty massive report, right? Revenue hitting 115, just over $115 billion, more than doubling from $57 billion one year ago, definitely beating the estimates of $109 billion that were set in front of it. Data center alone, $107.6 billion, obviously driven by the Vera Rubin kind of ramp up and strong guidance there. So does this – and, I mean, clearly, he just stepped down. So none of this can be attributed to CEO, the current interim CEO, Colette Kress. But what does this say about the next quarter and the next year, given where we're at right now?
Business as usual. Oh, hey, we've already lined up the business. It actually doesn't matter a whole lot. that he's not in that position. So not that the market knows Jack, but Tuesday before, yesterday before the earnings were out, the stock stood at 152. After that, it went up to 162. And right now it's back down to 152. So there was a bump for, okay, good. That was a good earnings. But I think the market is saying, is worried about the strategy for NVIDIA without them. And there's tons of, The pipeline of orders is huge. The orders are out there. They're going to be fulfilled. It's not like the company is going to crash. But what's the next big surprise and the next one after that? And what next one after that? And how do you coordinate the development of the chips, the next generation chips, with CUDA, with the robotics, I mean, with the open source Nemotron and stuff? There's just so much.
There were so many tentacles to NVIDIA. I think that if I were in the market, I'd be holding back and waiting and seeing. It depends on what they do about leadership. Yeah, which is a big open question. Now we're kind of like a week in. Do you feel any clarity around whether they should go external or go from within? I'm still feeling like you go from within. You don't want to reinvent the wheel. You go with someone who's really familiar with the playbook that's been working, and you go that route. What do you think? I agree. I mean, I don't know enough about the whole players. I don't know anything about the people within because Jason Wong was the sun god of the place. Yeah. But I agree. I think it's such a tremendously complex company, and it's in that complexity of its interlocking pieces where the success is. I think that makes sense. Yeah. I'm trying to remember who is it. It was Ian Buck. That's right.
Ian Buck was the name that was coming to mind for me. He helped invent CUDA. He runs products for data centers. He's on the list. He's currently the NVIDIA VP. And I don't know. That's a name. There was also a lot of talk around the AMD. Who's the one from AMD? Renee Haas. Lisa Su, maybe? No, no. Yeah, yeah. AMD CEO Lisa Su. There was a lot of talk about possibly bringing her in. but I mean she's AMD CEO not that I want to give them the slightest publicity because I don't like them but Kelsey has Rene Haas who's at another he's a former NVIDIA that's right there was a lot of a lot of talk about that a lot of chatter late last week about that Rene Haas is now the CEO of I should know this ARM ARM ARM confused with AMD right so he's the CEO of Arm, came from
NVIDIA. He's at 22%. Ian Buck is at 15%. Jensen Wong, now in the running at 11%. Colette Kress at 8%. And then other names I just don't know. Lisa Su is down at 3%. Nikki Sharora, formerly Google, at 4%. Tim Cook at 3%. No. Yeah, no. That was my joke last week, but I was not serious about the Tim Cook thing. Madison Wong, his daughter, at less than 1%. She's still there. Suck in Nutella at less than 1%. That would have been crazy, actually. You could buy Renee Haas for $0.22 on the yes, $0.79 on the no. I don't really understand how she works, so I don't know what that means. Yeah, don't know. Renee Haas, huh? NVIDIA almost acquired ARM back in 2020. Right. Could not get the regulatory approval. So that could be a big deal. So anyways, it's the gift that keeps on giving, apparently.
And now Trump's in China. So who knows how much that conversation is coming up there and what that matters. We'll see. There's going to be tons of hearings. Jensen Long has been summoned. Last night I looked at like five different congressional hearings, which he's going to have to go to, and he should. it would be interesting to see a humbled Jensen Wong I think what I'm curious to find out is how knowledgeable he actually is whether the Megaspeed mistress, I can't remember her name Alice Wong whether she is truly a Chinese asset, you know, a spy or just a convenient place person. I really want to know what Jensen knows. I want to know how much, because there is a lot of business-y things happening there
between using Megaspeed as the conduit for certain things. I had to imagine that Jensen Wong wasn't completely oblivious to that side of things. But I also have to believe that Jensen Wong, being Taiwanese, was very savvy about that. and it wasn't savvy to have an affair but I think he probably is savvy enough to know what he could be accused of if he let too much information out and also how much information was really valuable, what does he do? Tell them the secrets of how they make a chip? It's a little complex you don't draw that on a napkin but do you figure out a funnel of some sort for technology, you know, that you know they're not allowed to have. More to the point, I think what did here is that her company did hosting off the Chinese shore.
That could give access to anybody in China, number one. So it gave them access to NVIDIA chips. And number two, there are some chips that were presumed to be in their facilities that are missing. So how many ships ended up going to China? And did those ships matter in terms of the ability to use them in training models, or was it to reverse engineer them for Chinese shipmaking? I don't know. But I think that's where the question is. It's a hardware conduit rather than an intelligence conduit. Right. But I'm coming to believe more and more that that's not the story. that A, she probably wasn't an asset, B, information didn't seem to pass, C, the vipers in the White House used that conveniently to go against Wong and against China bulls, and people favorable trading with China is what I mean.
Okay. And then D, I think you have a good point. The in-house affairs were probably the killer question more with the board. than the Alice Wong ones. There's a lot of nuance here. Again, it's an open-to-debate event. I really wanted to try to bring nuance to it, and I was on with the China hand, so I was also quizzing him about how dangerous is China really. Now, he believes it's dangerous, and he believes that Wong should have been fired, no question, and there was a danger here. but I still I'm not sure how big the danger from China to us really is as a American consumer I want their open weight models here I want their cars I want to drive a Chinese electric car damn it and so you know I think this China angle as a whole may be overplayed as well something like Ted Lu is very
black and white saying, we shouldn't be giving anything to China that's just wrong to do. How stupid can you be? I have to agree with Jensen Long. I think it was stupid to cut them off. So that's another debate that has to occur. There you are. Open to debate is what you've been talking about. The podcast, right? It's basically a podcast. But you can find it on YouTube. Look for open to debate. Was Jensen Long's exit from NVIDIA the right move? Jeff Jarvis and Isaac Stonefish. Nice. You did this last Friday? No, I did this on Monday. Oh, on Monday. Okay. A couple of days ago. So they got to look pretty quickly. Yep. Right on. All right. Worth checking out, for those of you listening. If you haven't gotten enough of us wondering out loud. Well, yet more. This whole thing is. I was blathering, yes. All right. Cool. Well, we do have more to talk about that has nothing to do with Jensen Wong.
Well, I shouldn't say that. Actually, it seems like a lot of things probably tangentially have something to do with Jensen Wong. So there's that. So you can look forward to that. But we do want to thank patrons, patreon.com slash AI Inside Show, which is a place where you can go to support us on a deeper level. We were talking in pre-show. So I think I'm going to open up our Discord for free patrons and above. Right now, the Discord is a paid patron perk. Say that three times. So much alliteration, which is great. It's been a great perk. But honestly, I'd love to get more voices in there. And so I think I'm going to work this week on opening up the Discord so that if you are a free patron, even if you're just following the AI Inside Patreon page for free, you go to the page and set that up and it doesn't cost you a thing, that will get you access to Discord. So look for that coming sometime in the next week. I've got to figure out all the aspects of that to make sure I don't break something along the way.
But I'd love to get more conversation going on in there. And so letting more people in will hopefully make that happen. So if that sounds good to you, go to patreon.com slash AI Inside Show. and we want to thank a couple of amazing patrons this week, Ciprian Nicolescu and HiFiFreak. Thank you for your support, your ongoing support, months and months of support from these two, and so many more at patreon slash AIinsideshow All right Let take a super quick break when I can find out what we got coming up next here. What do we have? We have Anthropic IPO right around the corner, like literally right around the corner. I think maybe later today, we're going to hear some detail on that. Price ship. Yeah. And a whole lot more coming at you. Just hang tight. We'll be right back. Yes. So it is kind of the moment we've all been waiting for, or someone's been waiting for. I feel like it's the story that never goes away is this IPO from Anthropic that's always on the horizon but never quite here.
I think this is the week. Just like AGI. Just like AGI. It's around the corner. Although, would you agree that the term AGI, like there was a time where I felt like everyone was throwing that around, and lately we haven't really heard very much about that? They are pulling back because it's embarrassing, and then they're wrong, and it's not the goal, and then they kind of know it. And the other hilarious thing is in all the talk about all the outings we had, the FAA and Microsoft and so on and so forth, people wanting to do AGI, the administration people all have to say ASI. ASI, oh, yeah, yeah, of course. So is it SI? Is it SI? Is it SI? Not ASI, I'm wrong, sorry. Just SI. Right, SI. It's just SI, yep. SI, yep. It's comic to hear them. It's silly. That's so silly. It is. It's heady. and I still don't completely understand the why behind it, but yes, that is something that they are doing. Okay, so we got Anthropic pricing its IPO tonight, I believe.
Tomorrow on the NASDAQ, you're going to have the ticker ANT, short for Ant Pruitt. I will never look at that and not think Ant Pruitt. I just can't help it. Well, I also read a New York Times piece just now that I didn't put in. Because it makes me mad because it's more credulous reporting. They mention effective alphorism without going into the craziness and dangers of it. I spent a year talking to Anthropic employees, and all the writer came away with was PR from them. But I learned that inside Anthropic, they call themselves ants. Ants. Yes, that's what they call themselves. Okay. Ants for short. Little ants scurrying about. Yeah, it's just kind of. we're just tiny little ants humans now in the big network that is the brain oh man this is so not on topic and it's so you know what I heard but I was reading a Reddit post recently from someone who was talking
about you know has some sort I don't even know why I'm saying this out loud it was a Reddit post who the heck knows if it's even true But what it basically was was it was someone who reported to be very close to someone who works inside of Anthropic and was kind of spilling some tea and everything about how everyone in Anthropic is basically going along assuming that the future world where there is no human work, where UBI is just a normal thing, and where while they're working in Anthropic, they're just burning all sorts of their tokens to create things. so that when that all happens, they have something to show for it, some sort of product or something. And that just had me, I don't know. I don't know why I even paid any attention to that, but for some reason I could see it. For some reason it rang some bell of truth in my mind where I was like, oh, yeah, I could totally see the ants in Anthropic just skipping along,
being like, yay, and soon no one will be working anymore because of the work we're doing. It'll do all the work for us, and we will all earn universal basic income, and life will be glorious and beautiful. I don't know how I feel about that. Anyways, total hearsay, total tangent. What is there to say about Anthropic that we haven't already said before with this IPO, other than the fact that, like, right now is a really precarious time for an IPO to drop? That just continues kind of coming to mind for me. We got all these insecurities, these examples of what seems to be a lot more than coincidence that all these systems are just tanking at the same time and being impacted. And I'm sure there's a lot of people, you know, pointing fingers going, okay, so is this what we've been waiting for and warning about that fully autonomous AI agents are now hacking the world and we're just at the beginning of the cyclone? and, oh yeah, Anthropics also
going to be dropping the world's largest IPO in history at the same time. No biggie. Yeah. Yeah, so the seven founders are all going to be billionaires overnight. Too rich, too powerful. And Anthropics positive buzz remains. I think what happens with any of these, it'll skyrocket up because there'll be demand on opening day. It's going to go down from there. Evidently there's some kind of odd requirement for employees That if they sell their stock They have to file something beforehand So they're trying to avoid the cliff that happens when people vest And can get rid of it So the IPO happens Yay, I got my stock I'm just going to cash out now sort of thing Right And then So yeah, we'll just see where it goes But they've got a lot to promise then They've got to get in a ton of real revenue and create real value against competition that I think more and more companies and individuals really want to head to open weight as best they can.
They've got a rough road to hoe. And Google's not going to rest, and OpenAI is not going to rest, and Microsoft is not going to rest, and the Chinese are not going to rest. So, yeah, I don't know. I mean, I think you could play the stock in the short term, but it's a dangerous game to play with an IPO. The other story that fascinated me is how Google and Amazon will own 28% of the company's stock. Yeah, 14% and 15%. It's Amazon 15%, Google nearly 14%. Yeah. Right. And then as far as the classes go, Google will hold 32% of the non-voting class. Right. Amazon will hold 34%. This is according to courts. Amazon's holdings could climb to 24% of all stock if it lifts an ownership cap and converted all of its notes. So it's not, you know, the circular financing here, we're not talking about NVIDIA.
We're talking about Amazon and Google because they helped the company out in the early days and got really good value in result and used compute as a tradable asset. And boy, they're going to stand to gain a lot from that. Seven co-founders hold 50.1% of the voting power. Right. So they stay. Which means that unlike NVIDIA, it'll be really hard to fire them. Mm-hmm. Oh, so they can sleep with all sorts of subordinates. No big deal. Pretty much, yeah. Okay. Oh, boy. Yeah. What else is there to say about Anthropic IPO other than it's happening and all those things? Yeah. I mean, it is interesting, though, to contrast with the whole public benefit corporation safety thing by how far we've come away from that towards this.
Yeah. Cashing out a quarter billion on day one, you know, each. I mean, one thing that drove me crazy this week, it's not in the rundown per se, but Jacob Coxon, the whistleblower air quotes who left Anthropic, and was on every talk show there was about the safety of Anthropic. Nobody reported on his association with test-free occult. And now he did a talk at Stanford and said, oh, I might go back to Anthropic. And before that, he worked at OpenAI. And when asked about whether that was a conflict with his morals, he said, well, bro, I like money. And acknowledging calling up blood money. So I wouldn't give a lot of faith into any of these AI people thinking that they have the better results for humankind right in front of their eyes. Yeah, especially when the incentive is got to get your bag while you can.
Yeah, yeah. When it's right in front of you and it's never been bigger. Yeah. I think one thing that will happen, backfiring from this IPO, is that when people realize just how rich people got from it, the anger about tech oligarchs will be applied even more. And I think that's what, if you're going to watch for anything out of the Democratic regime when it comes into power, it's going to be really easy to kick tech oligarchs. Yeah, it'll certainly be a different landscape, a different environment for that than it has been. Absolutely. That is definitely going to shift to some degree. I don't know how widely or how significantly, but certainly that is changing. Speaking of Chinese AI models, sometimes I say Chinese models,
and I think it's like runway or something like that. Oh, that's what Jensen Wong That's the Jensen Wong joke of the day Oh, that's true The reason Chinese models We didn't mean We didn't mean that We didn't mean that Alibaba has a new model Quen 4 Max Dropped that this week It is a closed model, not open So there is that But yeah, they're touting it I don't know What can we say about a new model coming from Alibaba? I mean, my understanding on this is that it was there were a lot of people that were really waiting for this, and people seem to be a little disappointed with what they're getting, maybe partially because it's closed and the proprietary open wait variants are promised, but they're not here. They're coming later. But also that it's just it's testing notably slower and apparently a lot more verbose.
than some of the other models out there. So it doesn't pop. Yeah, basically it doesn't mean, faster doesn't automatically mean it's better on every job because it just might say a lot more in the process, I guess. I don't know. But, yeah, I feel like I want to do more exposure of some of these, you know, alternative kind of like open models and everything on the show. But I don't quite know where to go with it other than saying, like, it happened. I don't know. Yeah. If you don't have the hardware to run them, there's not a lot you can do. Yeah. I'm really glad for those people who are out there who do have the hardware to run them, who are playing with them and understanding them better. That's really important. And I support the open weights eagerly, enthusiastically. But I'm going to have to wait a while before I can do it. Mm-hmm. In terms of both hardware and technical abilities. Do you think that this stuff happening with NVIDIA, I mean, clearly Jensen Wong has been a pretty vocal and powerful, not driving force, but friendly, helping hand, advocate.
There's the word that I was looking for. Advocate for open source. With him not at the helm anymore, and with open models, I feel like one of the stories of the past year is that the open models have been closing the gap. They would never quite get up to the frontier level, but they've been closing the gap. And something like this happening at NVIDIA, they kind of lose that advocacy. Some of the policy is probably going to shift as a result. And so does that gap start to widen again, you know, between the open and the frontier? Yeah, that's my fear. I think that – but what happens then is that the duopoly of anthropic and open AI end up with more power. Right. They're very happy to see Jensen Long go. Jensen Long had a policy that was counterintuitive. It wasn't, it wasn't. It was to the extent that he was pushing open-weight models, which are not as profitable as frontier companies and don't have as much capital.
On the other hand, he was pushing more competition. And everybody who used any model of any sort who used CUDA wanted to use his chips. So he danced very interestingly between those two views. and yeah I think that I'm very worried about about what happens to open weights now because I think that the China hawks will say enough with this the safety the air quote safety hawks will say oh open weight's bad because you can't have guardrails on it even though guardrails are a fool's paradise the two very powerful now very rich companies open AI and Anthropic will fight open weight models fighting against China is also fighting against open weight models so I'm Yeah, all in all, I think it's potentially a bad time for open weights, which is a bad time for AI. Yeah, a bad time for open weights, for Chinese models, for AI in general. And China can switch around any day and change their strategy and say, nope, if we can't have your chips, you can't have our models.
Or we're now going to charge you a fortune for our models while we do all kinds of development and we race ahead. Either way, it could hurt us in the U.S. interesting interesting we also have an update on a story that we talked about quite a bit got so many pop ups and cooking notices and oh my god it's just annoying media ruined the web let's put that out there we screwed it up royally that beautiful thing that we came along I don't know it's probably going to get a lot more ruined from here too Anyways, an update to a story that we talked about quite a bit. You remember when the topic du jour was hugging face and OpenAI's models breaking out of sandboxes and all that kind of stuff? Well, this week, California Attorney General Rob Bonta sued OpenAI in San Francisco Superior Court over that, that, basically accusing OpenAI of violating California's unfair competition law and creating
a public nuisance by prioritizing speed over safety. It not wrong but it also showboarding Yeah Doesn do do much Like doesn have to No No and Bonta is the guy who caved when it came to Paramount and Warner Brothers Discovery And so he's trying to, I think, get back a little political capital. Not that I'm defending OpenAI. What OpenAI did in that case was irresponsible and wrong. So the suit's right to that extent. But how much it matters, I don't know. Well, OpenAI is saying, hey, you're singling us out over something that's been identified by every major AI lab. We're all doing it. Why us? Yeah. It's like kids on the playground. He started it. He did it too. Yeah. Yeah. Yeah. Apparently, California is saying that OpenAI admitted to state investigators that it had lost control of autonomous AI agents over 80 times during testing,
higher than had been publicly disclosed or spotted by outside researchers. So quite a bit of breaking out going on. Yep. And I mean, so, okay, so then take that and then take what we've been following also the last week, which is this kind of like onslaught of technology failures across the world. And do you think we're seeing what Google is worrying about here? You know, Google Threat Intelligence Group basically has a report that says for the first time, threat actors are using the AI agentic framework to turn security patches into working exploits with no human in the loop. This is just happening on autopilot now. And do you think that's part of what we're seeing? It's kind of early to say, but. I think it's too early to say it. I think that's that fear. So I'm glad we're not asking the question, gee, are all these things connected? Because they're not.
As far as we can tell, there's no evidence right now. I can be wrong like I was wrong about Jensen Long. But right now there's no evidence of cybersecurity issues when it comes to it. But it's a long list. FAA, Microsoft, National Security. Yes. Marriott, various banks. I think I'm leaving one out. I mean, but that's, like, it might not have anything to do with AI models or whatever, but isn't that a strange coincidence that all of these very, very major things are all happening at the same time? Like, I don't know. It feels a little, yeah, it's a little freaky. But I also think some of these things, like Marriott having problems, United, that was the other one, United went down. Marriott having problems on its own probably wouldn't have gotten any coverage. but journalists love a trend oh that's true once they see one then they start digging where all the examples
no matter how big it is we're going to go there so I think part of it the fact that it all happens in short order in a week is the freaky gap but I think that's it so two things one Dave Weiner and someone else I can't remember right now said online that AI is not directly related, but it could be related insofar as, and I hope I get Dave right on this point, it's finding vulnerabilities that companies then say we must urgently patch. But patching is itself a fraught activity. So it may not be the vulnerability that gets you. It may be the fix that gets you. So then if you now go to Google saying there are fully autonomous AI attacks out there that go on their own, not surprising, that could lead to more panic, and either the attack will get you or the panic will get you. So who was it they found?
They found it in Java. They found a vulnerability in something else. I can't remember what it was. I think I saw that in the Discord, actually. Dr. Dew mentioned something about a new Java vulnerability. Yeah, no, here it is. So AI helps find major bug in Java, yeah. So you're going to find more and more and more of those things out there, which also means that there's going to be a rush to patch, and they can screw up not doing anything, and you can screw up doing something. So I think that what this Google story says is, yes, watch out. It doesn't mean that the Armageddon is coming, but it means that it's going to make people even more nervous about every vulnerability they can find because there are new ways to exploit those vulnerabilities. But come on, fear, uncertainty, and doubt, Jeff. My business is built on that, man. I was going to say, we love fun. Yeah. In the modern age, it's a business model for news.
Absolutely it is. Absolutely it is. But, yeah, it's kind of a nervous time right now. You know, and I think that's bubbling up for people that don't follow this normally. They're starting to at least recognize that there is a trend happening. And, you know, they probably saw the AI doc on Netflix, and they're scared because of it. Yep. Yeah. Interesting stuff. All right. We are going to take a break. And then when we come back, we've got some speed round items to speed through. and then we will let Jeff get to resting his eyeball, at least until his next podcast. Yes. And I just got to say, so I don't know if you have noticed, because this is the first time that I've worn these glasses while podcasting and put it into a mode called Conversate, which basically Conversate is a mode where it's constantly listening. It can't hear you because I've got headphones in, but it can hear every word I say,
and it's putting out a transcript in front of me of every word I say. And then it picks up on themes and topics and pulls up extra information related to it and pops it up. Perfect. It would be if I felt okay every once in a while stopping looking at you and start going like this. You know, like, I guess I could, like, tilt my head and then the words are in front of the screen so you'd kind of not know. but it kind of doesn't look totally normal. So, like, technically it's an interesting idea, but it's like removing myself from the moment. Yeah, you have an out-of-body experience. Yeah, it feels somewhat disconnected, and I'd be afraid if I do that, then you ask me a question, like you sometimes do. Well, how do you feel about that, Jason? And I'd be like, uh, what did you say? But at least I have a transcript so I can go back in time. Anyways, it's been an interesting show. We'll talk about an application that does that in a second, yes.
Yes. All right, let's take a break. We'll talk about what we've got here in the speed round coming up. Don't go anywhere. All right, a few things that have released, which seems to be a lot of what the speed round tends to be about. We've got Astra and Sol 6.2 from OpenAI right on the eve of Anthropix IPO. So ChatGPT and OpenAI is like, oh, but we got a thing. Don't leave us out of the conversation. As usual, Astra is the heavy. Sol is the faster everyday model. Prices drop. You know, API prices drop. It's more affordable to use. They're better. Benchmarks apparently look strong. Astra leads on agents' last exam in terminal bench science. Hits OpenAI's high threshold for AI self-improvement, meaning OpenAI. rates it as giving every researcher an extra research engineer as an assistant. So there you go.
You get an assistant. You get an assistant. You get an assistant. Okay. I don't know. I've been using ChatGPT a lot lately, I have to say, for a lot of different things. And so getting a much better understanding of how it's useful to me. So that's nice. Do you have a favorite board? Does it go back and forth? Do your affections go back and forth? Between the different models? Yeah. Yeah, it's really interesting, Jeff, because right now I find myself using Claude the least. I've been kind of pulling away from Claude, and I end up using ChatGBT. I use it for a lot of inbox work. I have it tied into my inbox for certain aspects of my business, like monitoring for relationships, sponsorship relationships, that sort of stuff. So I built a system that's actually working for me right now. So that's really nice. Just because I get so much email, so much inbound email, and I would always just end up archiving instead of doing anything with it
because I just didn't have the time to deal with it. And I was like, well, something's better than nothing. So let me set up this system. And so I've got a system going there. Antigravity, I think I've told you, I've been using a lot of that lately, and I've been using that for real structural kind of like figure out a framework around this or create a scaffolding around this process that I do all the time. I don't know. But in Claude, I'm not using very much right now. And it's okay. I actually downgraded that to the free account for next month because I'm just not using it. Which, actually, now that I say this out loud, makes me realize that's the real challenge that these model makers face is there is nothing that really keeps you locked in. You know what I mean? Yeah. Which is another challenge for Anthropic with its IPO and OpenAI for its IPO. Yep. Yeah. Like you might be using it for a while and then realize, oh, it's just not what I need. It's just not doing it for me right now.
I'm going to go over there for a while. And if you've done the work, which I have, to build up like my brain and a file structure on my hard drive, that's where all the documents end up getting created into. And so it kind of matters less which model I'm using because I point them all at the same thing. And so all of the work and all of the instructions and everything is on my hard drive. It's not inside of the model itself. And so that means I can really just, from portability perspective, go to any model and say, all right, I want you to do the work now and see how you do it versus that other model. And it's a cool experience. I'm learning a lot. Okay. So I just found out that I'm not supposed to be on a call in a few minutes. Uh-oh. Let's speed this up. Lightning round, yeah. Okay, lightning round. Holy crap. Okay. Anthropic has clawed whiteboards, sheets, and data. This is, you know, it's just getting more into the office world. Speaking of Anthropic, right before the IPO wants you to know, hey, we're not, you know, we're not a one-trick pony.
There you go. There are the Android XR smart glasses, intelligent eyewear, now launching in December, Gentle Monster, Morby Parker. You've tried these. I've tried them multiple times over the last couple of years. Not in your hands. You've gone to events where you've tried them. You haven't had them do yet. I hope to be able to review these. That'll be really interesting. You get some time with them, yes. There was a New York event a couple of weeks ago. I was unable to go. I was invited. I was unable to go, but Ron Richards from Android Faithful went, and he was very impressed. And he's not a smart glasses guy. So to hear him say last night on the Android Faithful podcast how impressed he was by it, I was actually like, oh, all right, maybe there's something to that. Apple Watch, we've talked in the past that it has some new features coming called Live Rewind and Siri Recap. It's like these ongoing ambient summaries of your day based on audio that feeds through. It doesn't actually save the audio. It also gives you a 15-second rewind, and it seems like all the reviewers are saying the same thing.
Awkward, unuseful, and kind of broken because Siri Recap can't tell the difference between real conversation and other things that happen. And so it gives you a summary from, like, the YouTube video you happen to be watching or radio or whatever and just becomes completely unusable. Yeah, I'm dubious about this. And privacy issues and weirdness issues and all that. Yeah. But compared to your glasses, it'll be really interesting, even though you're not an Apple guy, neither am I. Yeah. To conceive of the differences between that and the glasses you're wearing right now. Well, I can tell you one difference. The glasses I'm wearing right now capture everything, and there is no limit, and it's a pure transcription, which gives you a lot of flexibility from a context perspective later than something like what Apple's doing here, which was really built around privacy, and I respect them a lot for that. They constantly do that. I just wonder if that mentality applied to this kind of technology makes it pretty useless.
You know what I mean? And so maybe that's just not a product that needs to exist. Yep. All right, JeffJarvis.com. That's it. We got there. JeffJarvis.com for all of your favorite Jeff Jarvis things, including Hot Type, The Web We Weave, Gutenberg Parenthesis, and every single book. Always written ever. They are listed there. Thank you, Jeff. Podtuneup.com. If you have a podcast and you want help with your podcast, hey, I'll talk to you about it. You can chat with me for 30 minutes. I'll give you 30 minutes of my time for nothing. We'll talk about your podcast and see if it leads somewhere. Happy to do it. Podtuneup.com. AIinside.show for everything you need to know about this show. You can find all the links to subscribe, everything. Latest episode. I just love the layout. Built with Gemini for carbon. There you go. And finally, patreon.com slash AI inside show to throw your support behind everything that Jeff and I are doing. I want to give a big thank you to our executive producers. Dr. Dew, Jeffrey Maricini, Radio Asheville 103.7, Dante St. James, Bono DeRick, Jason Neifer, Jason Brady, Anthony Downs, Mark Starcher, Karsten Somachki, and Evan Musial.
At the speed of lightning. and all of those names listed in my glasses right now, horribly misspelled because my glasses do not know how to spell first and last names very well. All right. That's funny. That's it, Jeff. Got to get you to your meeting. Everyone, thank you so much for your support. No, don't be sorry at all. It's all good. We're rounding out anyways. Thank you, everybody, for watching and listening. We appreciate you immensely, and we'll see you next time on AI Inside. Bye, everybody. Did you know that if your spouse goes missing and you don't tell anyone, that's not necessarily a crime? That surprising fact is at the heart of the brand new season of Bear Brook, the critically acclaimed true crime podcast from New Hampshire Public Radio. I'm your host, Jason Loon. In this season, Bear Brook investigates another New Hampshire mystery,
A Christmas tree farmer with dementia vanishes. His wife never reports him missing. Can he still be found? Don't miss it. Follow Bear Brook wherever you get your podcasts.
番組の概要欄(原文)
This week Jason Howell and Jeff Jarvis examine Jensen Huang’s bid to return to Nvidia, the board’s refusal, and the questions facing the company after another enormous quarter. They discuss Anthropic’s approaching IPO, Alibaba’s closed Qwen4-Max release, California’s lawsuit over OpenAI’s escaped agents, and Google’s warning about automated cyberattacks. Also in this episode, Jason demonstrates his Even G2 smart glasses and explains why keeping his AI instructions on his own hard drive makes switching tools easier. The speed round covers GPT-6.2 Astra and Sol, Claude Whiteboards, Sheets and Data, Android XR glasses, and the Apple Watch’s conversation summaries. New episodes every Wednesday at aiinside.show. Note: Time codes subject to change depending on dynamic ad insertion by the distributor. CHAPTERS: 00:00 - Podcast begins 03:52 - Jason tries the Even G2 smart glasses 09:37 - Jensen Huang’s bid to return to Nvidia 19:40 - Nvidia earnings and its search for a CEO 26:18 - Nvidia, China, and access to chips 31:06 - Anthropic’s IPO and investor stakes 40:58 - Alibaba’s Qwen4-Max and the future of open models 45:19 - California’s lawsuit against OpenAI 50:05 - Google’s warning about automated AI attacks 53:39 - GPT-6.2 Astra and Sol 56:36 - Keeping AI instructions on your own drive 57:25 - Claude Whiteboards, Sheets, and Data 57:40 - Android XR glasses 58:23 - Apple Watch Live Rewind and Siri Recap 59:51 - Wrap-up and where to find the show SOURCES: Jensen Huang Breaks Silence, Signals Willingness to Return as Nvidia CEO Tech world’s calls for Jensen Huang to return to Nvidia are growing The Campaign to Bring Back Jensen Huang as Nvidia CEO The Jensen petition Nvidia Beats Estimates in First Earnings Report After Jensen Huang’s Ouster Jensen Huang’s fall hits ‘open’ models in boost to Anthropic and OpenAI Jensen Huang is a litmus test for China hawks vs. (expedient) China doves, open weights vs. closed models, founder class vs. managerial class Jeff in Open to Debate over l'affaire Jensen Anthropic forges ahead with its IPO even as AI fears mount and markets wobble Anthropic's IPO filing shows Amazon and Google will own about 28% of its stock Alibaba's new Qwen AI model poses challenge for Anthropic Alibaba touts 62% lower task costs with Qwen4-Max, adds cyber safeguards California sues OpenAI, seeking court-ordered changes after the Hugging Face hack Google warns that fully-automated AI attacks are coming soon Google says an AI framework turned security patches into exploits with no human in the loop - SiliconANGLE Introducing GPT‑6.2 Astra, GPT‑6.2 Sol Introducing Claude Whiteboards, Claude Sheets, and Claude Data First Android XR Smart Glasses From Warby Parker Gentle Monster Are Launching in December The Apple Watch Can Record and Rewind Spoken Conversations Now. It’s Pretty Awkward Jeff Jarvis Intelligence — AI and Humanity Podtuneup Hosts: Jason Howell and Jeff Jarvis Download and subscribe to AI Inside in audio and video: aiinside.show. Support the podcast on Patreon for special perks. You'll get ad-free audio and video feeds, the AI Inside Discord community, T-shirts and stickers, and exclusive content. Learn more about your ad choices. Visit megaphone.fm/adchoices
出典・この記事について
- 番組
- AI Inside(Jason Howell and Jeff Jarvis)
- 元の回
- Nvidia Faces Its Future Without Jensen Huang
- 配信日
- 要約作成日
- 公式サイト
- aiinside.show
この要約・解説記事は、元の音声を文字起こしした内容をもとに AI(Claude)が日本語で作成したものです。 固有名詞や数値に誤りを含む場合があるため、正確な内容は元の音声でご確認ください。制作方法について
関連エピソード

AI通がAstra 6.2を嫌う理由
The AI Daily Brief: Artificial Intelligence News and Analysis

ジェンスン・フアンがNvidia CEOを辞任、AnthropicのIPO、AI 2027の前倒し予測、HubSpotのAI再創業
The Artificial Intelligence Show

緊急特集:BISがついに抜け穴を塞ぐ? 次に習近平はどう動くか
ChinaTalk

ジェンスン・ファン失脚の衝撃――Nvidia CEO退任劇をWIREDが読み解く
Uncanny Valley | WIRED

ジェンスン不在のオープンソースAI
Interconnects

特別レポート:ジェンスン・フアンの失墜とその後に来るもの
Big Technology Podcast