Podcast Summary: The Last Invention is AI

Episode: New Apple CEO, OpenAI's New Image Model, Vercel AI Hack
Date: April 21, 2026
Host: The Last Invention is AI

Episode Overview

This episode dives into major developments at the intersection of AI, big tech, and business strategy, including Apple’s huge leadership change, a cutting-edge image model from OpenAI, shocking security lapses at Vercel caused by AI tool integration, the consequences of AI workforce automation at Snapchat, and the latest massive funding waves in the AI infrastructure wars. The host analyzes the implications, offers incisive commentary on how AI is restructuring industries, and shares candid opinions on the trends shaping the future.

Key Discussion Points & Insights

1. Apple CEO Transition: Tim Cook to John Ternas

[00:00, 37:10]
- Tim Cook is stepping down on September 1st after 15 years; John Ternas, SVP of Hardware Engineering, is taking the helm.
- The timing is pivotal – Apple’s AI offerings (“Apple Intelligence”) have lagged behind, with Siri underperforming compared to AI-first competitors.
- Apple's AI approach: “We’ll do it on device, private, and you can trust us,” but execution has been lacking.
- Insight: Appointing a hardware engineer signals Apple’s belief that “the next AI war is going to be won on devices and on silicon, on power efficiency, and on the integration between all of the chip models.”
  - Quote [38:05]:
    
    “If they are making the correct bet there, then I think the transition is going to look amazing in five years… If they're wrong and AI keeps living in the cloud, then Apple just picked basically a hardware-first leader at the exact wrong moment.”

2. Vercel AI Hack & The Risks of AI Tool Integrations

[01:29]
- Vercel (behind Next.js, hosting a swath of modern web infrastructure) experienced a major data breach—not from being directly hacked, but via a supply chain attack through Context AI, used by a Vercel employee.
- Attack vector:
  - Context AI employee infected with infostealer malware
  - Credentials exploited, leading to broader compromise of Vercel’s environment
  - Hackers (Shiny Hunters) are selling stolen API keys, source code, and database data.
- Broader point:
  
  “Every company right now is bolting AI tools onto their stack…every new tool is a new vendor, new credentials, a new way for attackers to get in. … We’re going to see a lot more AI-adjacent supply chain breaches. This is going to happen more before it gets better.” [03:28]
- Advice:
  
  “If you’re at a company, audit who has access to what through your AI tools.” [04:03]

3. FTC Settlement: Clarifai’s Unethical Training on OkCupid Photos

[05:10]
- Clarifai used 3 million OkCupid user photos (dating back to 2014) without user consent to train facial recognition AI.
- The Clarifai CEO even directly emailed OkCupid cofounder about “collecting data.”
- As a result of the FTC settlement, Clarifai must delete both the images and any models trained on them.
- Major insight:
  
  “For years, AI companies have been operating kind of in this gray area…if it was on the Internet, we can use it. And the FTC just said, no, you actually can’t. If people don’t consent to this specific use, you can’t actually have it.” [06:40]
- Warning for startups: This settlement is “the one you should be studying” if you trained on scraped data.
  
  “You need consent for use of the data... I think we're going to see a lot more of these in the future.” [07:03]

4. Snapchat Layoffs & the Age of AI-Driven Workforce Automation

[07:31]
- Snapchat cuts ~1,000 employees (16% of its workforce) and closes 300+ open roles.
  - CEO Evan Spiegel’s memo: AI now generates over 65% of new code at Snap, driving $500 million in expected annual savings.
  - Stock surged nearly 8% after the news.
- Larger trend:
  
  “The question isn’t whether AI is going to reshape your job. It already did.” [09:03] “If you were someone that got cut, you’re going to want to use these tools, you’re going to want to learn how to really integrate them and make yourself more valuable.” [10:07]
- Power-users who integrate AI are the ones kept; non-adopters see their jobs consolidated.
- Candid takeaway:
  
  “That’s probably what you’re already doing if you’re listening to the podcast.” [11:35]

5. OpenAI Image 2.0: Major Jump in Multilingual, Text-Savvy Image Generation

[12:00]
- OpenAI launches Image 2.0, a significant upgrade from 1.5:
  - Clear, multilingual text generation (including Japanese, Korean, Hindi, Bengali)
  - Handles non-Latin scripts, poster and infographic creation, slide decks, and even consistent pages for comics
  - Can generate up to 8 consistent images per prompt
  - “First OpenAI image model with reasoning baked in”—it now searches the web and verifies content before generating final images.
- Host’s use case: Helps in regenerating podcast cover images in multiple languages.
  - Quote [14:52]:
    
    “This is just an awesome tool. I’m super stoked about it.”
- Available free for Go users; Plus and Pro users get higher tiers; API available immediately.

6. OpenAI vs. Anthropic: Debates Over Model Safety and Fear-Based Marketing

[16:09]
- Sam Altman (OpenAI CEO) publicly calls out Anthropic’s Mythos preview as “fear-based marketing.”
  - Anthropic claimed Mythos is so good at finding vulnerabilities, they can’t yet release it publicly.
- Host opinion:
  
  “There’s a consistent pattern at Anthropic of announcing a model alongside a report on how dangerous it is… Sam Altman is kind of right to call that out…” [17:10]
  - Recalls similar “doom” framing with GPT-1 by OpenAI itself
  - Cautions: Cherry-picking and exaggeration may be happening, but Altman is “not a neutral party.” Still, skepticism about hype is warranted.

7. OpenAI & Novo Nordisk: AI Rollout in Pharma

[18:27]
- Novo Nordisk (Ozempic producer) inks a multiyear partnership with OpenAI:
  - AI to be rolled out across drug discovery, trials, manufacturing, supply chain by 2026.
  - If AI drastically reduces drug discovery timelines, downstream impact on health could be profound.
  “I think that every major pharma company is also going to be signing deals that look basically exactly like this in the future.” [18:44]

8. Amazon’s Massive Investment in Anthropic and the AI Compute Arms Race

[22:08]
- Amazon commits $25B more to Anthropic, atop previous $8B (total $33B) — intricately linked to Anthropic’s $100B+ AWS cloud commitment over the next 10 years.
  - $5B upfront at Anthropic’s $380B valuation; $20B more linked to commercial milestones.
  - Gives Anthropic access to over 5 gigawatts of Amazon’s custom Trainium chips—scaling up to 1 gigawatt this year (“that’s the output of a large nuclear reactor”).
- Strategic insight:
  - Distribution: Claude available natively in AWS for Fortune 500s
  - “Amazon…is not picking the winner. They’re just going to be giving money to anyone that’s going to be giving it back to them and helping their platform grow.” [24:38]
- Infrastructure gold rush: Model training is now a capital moat and a core competitive axis.
  
  “Every major lab is now pre-committing tens of hundreds of billions of dollars to cloud providers and the compute alone is basically becoming a moat.” [25:08]
  - According to rumors, current OpenAI model training is using “billions and billions of dollars” in compute.

Notable Quotes & Memorable Moments

On Apple’s AI struggle:

“Apple has not…they made a lot of promises so it's, it's kind of wild to see Tim Cook having all of these kind of AI features announced and then just saying, all right, handing over the reins to the next guy to deliver on them.” [37:30]
On being an AI power-user:

“If you were someone that got cut, you're gonna want to use these tools, you're gonna want to learn how to really integrate them and to make yourself more valuable.” [10:07]
On the AI compute gold rush:

“Every major lab is now pre committing tens of hundreds of billions of dollars to cloud providers and the compute alone is basically becoming a moat.” [25:08]
On FTC and data privacy precedent:

“If people don’t consent to this specific use, then you can’t actually have it. And I think the models that are trained on it have to also come down, which is incredibly expensive.” [06:42]

Timestamps for Important Segments

Apple CEO Transition/AI Strategy: [00:00], [37:10]
Vercel AI Security Breach: [01:29] – [04:03]
FTC/Clarifai Data Misuse: [05:10] – [07:03]
Snapchat Layoffs & AI Automation: [07:31] – [11:35]
OpenAI Image 2.0 Launch: [12:00] – [15:06]
OpenAI vs. Anthropic/Fear-based Marketing: [16:09] – [17:47]
OpenAI & Novo Nordisk Deal: [18:27] – [19:14]
Amazon & Anthropic Cloud Arms Race: [22:08] – [25:08]

Conclusion

This episode provides a fast-paced, insightful tour of AI’s most pivotal business, ethical, and technical developments—highlighting both risks and opportunities as the AI revolution accelerates across tech, healthcare, and finance. The host’s candid takes, industry analysis, and predictions are a valuable resource for anyone tracking the interplay of AI, business, and society.

Transcript

A (0:00)

Apple CEO Tim Cook is officially stepping down on September 1st and John Tarnas is going to become the CEO. This is really interesting. There's a lot of AI things at play for Apple and you know, they really haven't delivered on their AI promise. We're going to be getting into that. Vercel, the company that I personally use for being my web host for all of my AI tools, has had a major security incident because of an AI tool that one of their employees was using. And the FTC just dropped a settlement with Clarify or over the company using 3 million OkCupid photos to train facial recognition. Snapchat laid off around a thousand people and said that is part of their AI initiative because AI is writing more than 65% of their new code. They dropped a major new image model over at OpenAI called Image 2.0. That is a huge step up. I've been playing with it today. We'll talk about that. And in addition, Anthropic's Mythos model has, has of course it's starting to get a bad rep and Sam Altman is going and calling them out for their fear based marketing. They also signed a huge enterprise deal at Novo Nordisk, the Ozempic company we'll be talking about. And I think one of the biggest stories is that Amazon is pouring over $25 billion into anthropic on top of the hundred billion dollars that the AWS team has already committed. Now, if you're still paying for ChatGPT, Claude Gemini, Grok, any of the audio models, Elevenlabs for audio, and any of the image models, I've got to tell you about AI Box. This is what I personally have built. It's what I'm recommending to my friends who ask me how to actually use AI without going broke on the subscriptions. You get access to over 80 different AI models in one place, all of the top models. So you can pick whichever one is best for the task that you are doing. The part that I think is super useful is our automation builder that we have just created. You describe what you want in plain English and it builds out a workflow for you. You don't need to know how to code, you don't need to, you don't need know how to, you know, wrestle with a, with a new platform. You describe what you want and it builds it for you. So it's 8,99amonth and I hope this is something that saves you a ton of money and is incredibly useful for getting your hands on all of the different AI models to use them. All right, let's talk about Vercel's security breach. This is the company behind Next js. They host a huge chunk of the modern web. I basically create all of my new AI apps on them and they have confirmed that they have got breached and customer data was so. But this is the interesting part for me in particular. They don't get hacked directly. The actual the attacker actually broke into Context AI, which is an AI tool that Vercel, one of their employees was using. And from there they got into that employee's Google workspace. Workspace, and from there into Vercel's environment. So there's kind of like this multi step approach that they took in order to get in. And the original compromise actually happened back in February. We're just hearing about it now. I feel like this is how a lot of these kind of kind of attacks happen. It, you know, the event happens, they don't really want to leak it and then it finally, you know, comes out. So a context AI employee actually got hit with the info stealer malware and the attacker basically just sat on the valid credentials until they could chain them into something bigger. Vercel says that the number of affected customers is limited. But a group claiming to be they're calling themselves Shiny Hunters is on hacker forums trying to sell API keys, source source code and database data that they say they pulled from this whole hack. So in my opinion, every company right now is bolting AI tools onto their stack and every new tool is a new vendor, it's new credentials, it's a new way for attackers to get in. I think if you're gonna see, you know, a lot of these new AI tools get built, we're also gonna see a lot more AI adjacent supply chain breaches. This is gonna happen more before it gets better. If you're at a company audit, who has access to what through your AI tools? I don't think this is something theoretical. I think everyone should be thinking about this. The FTC also just settled with the AI company called Clarify over the use of 3 million photos from OkCupid, that's the dating website. And they use these to train facial recognition AI. And the best part is that they actually have receipts. So the Clarify CEO literally emailed the OkCupid co founder saying, quote, we're collecting data now and just realized that OkCupid must have a huge amount of awesome data for this. The photos were then shared going back. All of the photos that they had were shared going back to 2014. Users didn't know what was going on. So as part of the settlement, Clarify has to delete the 3 million photos and any models trained on them. I think that this is, you know, something that is absolutely insane. For years, AI companies have been operating kind of in this gray area that, you know, they're like, well, if it was on the Internet, we can use it. And, and the FTC just said, no, you actually can't. If people don't consent to this specific use, then you can't actually have it. And I think the models that are trained on it have to also come down, which is incredibly expensive. So this is just not good for anyone in the industry. You need consent for use of the data. I think that this is a really interesting precedent, this story in particular, because if you're an AI company that trained on scraped data without clear consent, basically this is the settlement that you should be studying. I think we're gonna see a lot more of these in the future. Okay, Snapchat just announced that they're cutting roughly a thousand people. That's about 16% of their global workforce. They're closing more than 300 open roles. And even Spiegel sent the memo that's kind of announced all this. Their CEO and the reason that he gave is basically the thing that I think is critical. He said that AI is now generating more than 65% of new code at Snap, and the company expects to cut over $500 million in annualized expenses by the back half of this year. So they are cutting in a massive way. The Stock jumped almost 8% on the news. And I think that tells you exactly how Wall street is feeling about this trade off, in my opinion. I think this is something that a lot of us have been saying was coming. I think this is going to be repeated because here we have a CEO writing a memo to a public company. AI is writing a majority of the code. Smaller teams are shipping more, and they're also being directly, you know, they're attributing this directly to AI and the market is rewarding them. Wall street is rewarding them. So I think we're going to see this in a lot more companies. This is what I do think. It tells us if you're a knowledge worker, the question isn't whether AI is going to reshape your job. It already did. I think the question is whether you're using it to become really a power user on the team or whether your tasks just got consolidated into someone else's workflow. I know this sounds really ruthless, but I think this is a very honest take. And looking at this case where he cut about 16% of their global workforce. Well, who didn't he cut? In my opinion. In my opinion. I mean, there's still a huge chunk of the workforce that they're keeping, and those are the people that are becoming the power users. They're using AI. So I just think it's really critical now. And if you were someone that got cut, you're gonna want to use these tools, you're gonna want to learn how to really integrate them and to make yourself more valuable. So the next job that you find, they're gonna want you to basically be kind of like an AI expert in your industry. So anyways, I know this is kind of. It's kind of a ruthless, Ruthless way to put it, but I do think it's true. And so I just want to put it out there that we should all be taking AI to the. Pushing it to the limits, testing out everything new. If you're listening to the podcast, that's probably what you're already doing. Okay, OpenAI has a ton of news today. I'm going to tie it all together. The first thing is that they dropped ChatGPT Image 2.0. This is something I've been playing with previously. They had image 1.5. Basically, it's really good at doing something that the last model 1.5 was pretty bad at, which is rendering text. If you ask it for a poster, you know, you would always get this kind of like elfish looking. I don't know, the letters weren't that. Weren't that great. In fact, I recently had like a plumber or, sorry, an AC repairman come to my house and he sent me his business card via text. And I could tell it was generated with ChatGPT and there was like a typo in it because the letters just weren't that good. Anyways, 2.0, in their words is a quote. Step change. They always say step change. So I don't even know what that means. It's better. Okay, fantastic. Basically, it makes really clear multilingual text, so it's good at doing other languages as well. It can do non Latin scripts like Japanese, Korean, Hindi and Bengali. Personally, I'm excited about that because I actually built a feature into my software company, podcast studio.com, which isn't public. I haven't dropped anything yet. I'm personally using it and have a couple beta testers on it, so it's not something big. But one of the features that I added to it was essentially, it's basically to help you publish podcast episodes faster and easier. And one thing it does is you can select multiple languages. So like, for my own podcast that you're listening to here, I also publish this in French and like a bunch of other languages, but we can actually build you your own podcast in that language and help you get all of your distribution set up and stuff. And the cool thing is we actually regenerate your podcast cover image. So, like, if you have text with your podcast name and you want it, it will regenerate your whole cover image, just changing the text to whatever the new language is. Anyways, it's one of the things I built. I'm stoked about it because it was a bit of a struggle. When you translate it to a new language, the text sometimes would get kind of garbled and I'd always have to double check to make sure that it actually said what it was supposed to. So anyways, this is what it's good at. It's good infographics, slide decks, maps, pages of comic books. It can produce up to eight consistent images from one prompt. And I also think that it's the first OpenAI image model with reasoning baked into it so it can search the web and verify that it's generating something, something correctly before it finishes the images. That's a massive upgrade just in and of itself. It rolled out today. It's for free and go. Users plus and Pro get higher quality tiers basically. And it's already in the API. So this is just an awesome tool. I'm super stoked about it. Sam Altman also went on a podcast recently and he took a very direct shot at Open or at Anthropic's Claude Mythos preview. He called it, quote, fear based marketing. So if you missed that, Anthropic basically said that Mythos is so good at finding software vulnerabilities that they, you know, they can't release it publicly. Oh no, it's going to take over the world. And Sam Altman basically made the argument that their framing was just basically a way to look powerful and keep AI exclusively to themselves or their, you know, rich companies. I think that probably both of these things are true. I think Mythos does seem like a really capable model, but I also do think that there's a consistent pattern at Anthropic of announcing a model alongside a, you know, report on how dangerous it is or how crazy it is. And I think Sam Altman is kind of right to call that out. It definitely is annoying. You can See articles like dating back to the first version of ChatGPT, like GPT1 when Dario was over at OpenAI and they were saying like, it's so dangerous to release this thing anyways. I don't think Sam Altman's definitely like necessarily a neutral party on that, but I do think it's good to call people out when they're saying that their model is going to destroy the world when it probably isn't. Some people have been looking at the data and saying they cherry picked a bunch of it and, and it's not as dangerous as they basically led on to say. All right, Novo Nordisk, the Ozempic company, recently signed a massive partnership with OpenAI to roll out AI across their entire business. Drug discovery, clinical trials, manufacturing, supply chain, everything. Full deployment by the end of 2026. If OpenAI can actually help Novo compress drug discovery timelines, then the main, then the downstream impact on human health I think is gonna be really big. This is an area I'm excited about, is AI and healthcare and, and I think that every major pharma company is also going to be signing deals that look basically exactly like this in the future. Okay, Apple just announced they have a CEO transition. This is basically the next big player. We had Steve Jobs, Tim Cook, and we have the next big player finally coming in. Tim Cook has been the CEO of Apple for like 15 years. And they're going to be pulling in John Terenas, who is currently SVP of hardware engineering. He's going to become the CEO effective September 1st. Cook is going to move to be the executive chairman. Teranis has been at Apple since 2021, so he's an engineer. He has quietly overseen basically every major hardware product that Apple has shipped in the last decade. And this is the first big CEO change since Cook took over from Steve Jobs in 2011. So the reason why I want to talk about it on the, on this AI podcast is because the timing I think is really important. I have been very critical of Apple, saying that they are far behind for, I don't know, as long as I can remember on AI. And Apple Intelligence was late. It was definitely underwhelming. Siri after 15 years is still Siri. Meanwhile, OpenAI, Anthropic, Google, Meta, they're all shipping these really aggressive AI products. And Apple's main public AI pitch has basically been, you know, we'll do it on a device and private and you can trust us. But they've never really delivered on this. Apple has not. They made a lot of promises so it's, it's kind of wild to see Tim Cook having all of these kind of AI features announced and then just saying, all right, handing over the reins to the next guy to deliver on them. I think handing the company to a hardware engineer at right now is basically a big choice. I think that Apple believes the next AI war is going to be one on devices and on silicon, on power efficiency, on the integration between all of the chip models. And I think that's what Teranis entire career is hopefully has prepared him for. If they are making a correct bet there, then I think the transition is going to look amazing in five years. If they're wrong on this and you know, AI keeps living in the cloud, then Apple just picked basically a hardware first leader at the exact wrong moment. So I think we'll be watching this one closely. The last thing I want to bring up is that Amazon is going to put $25 billion into anthropic in a massive cloud deal. They announced that they're investing all of this on top of the 8 billion that they've already put in since 2023. And while the $25 billion sounds incredible, it's sort of one of those circular money deals because Anthropic is also committing over $100 billion to AWS spend over the next 10 years. So here's how the deal basically is going to be rolled out. Amazon is putting in $5 billion right now, and this is at Anthropic's new $380 billion valuation. The other 20 billion is tied to commercial milestones. So it's going to be basically performance linked. And part of this deal is that Anthropic gets access to over 5 gigaw on Amazon's custom Trainium chips. And they're planning to bring nearly 1 gigawatt of Trainium 2 and Trainium 3 capacity online by the end of this year. So 1 gigawatt is roughly the output of a large nuclear reactor. And all of these companies are building AI data centers at, you know, power plant scale. So this is, you know, this is a really big part. Anthropic is taking money from someone who's going to be able to help them out and Amazon is investing money into someone who's going to be giving it back to them. So I think there's two pieces. First, the full Claude platform is going to be available directly inside of aws. If you're already an AWS customer and a huge chunk of the Fortune 500 companies are, you can now use CLAUDE through your existing account billing and security controls. I think that's a really big distribution unlock. The second thing is that Amazon just two months ago also agreed to put up to $50 billion into OpenAI. So Amazon is now the cloud backbone for both frontier models. And Amazon has Azure and OpenAI. Google has Google Cloud and Gemini. Amazon now has AWS and anthropic plus AWS and OpenAI. So they basically are not picking the winner. They're just going to be giving money to anyone that's going to be giving it back to them and helping their their platform grow. I think this one is basically going to cement AI infrastructure and basically it's showing that this is an actual gold rush to not the models, not the apps, but the compute underneath. Every major lab is now pre committing tens of hundreds of billions of dollars to cloud providers and the compute alone is basically becoming a moat. Right? And I even saw a rumor that with the latest model, OpenAI is training that the amount of compute they're using is billions and billions of dollars. I mean, this stuff is absolutely wild. All right, that's the show for today. If you got something out of it, it would mean the world to me. If you drop a comment on Apple podcasts or hit the stars over, it's the about tab on Spotify where you can hit some stars or drop a comment over on Apple. It helps us show out a ton. And also if you want to get access to over 80 AI models and an automation builder that runs on plain English for 8.99amonth, go check out AI box. AI. There's a link in the description. I'll catch you guys all in the next episode.