Vertex ($VERX) podcast mentions
Timestamped Vertex ($VERX) mentions across our podcast corpus.
I do not. I do not. I had the great good fortune the fortune of having a board chair early in my tenure, Joshua Boger, who was the founder of Vertex Pharmaceuticals. And like the founders of companies, I'm told more generally, they don't like to have a lot of oversight. And Joshua said to me, my job is to support you or fire you.
Michael Roth · Freakonomics Radio · How to Run an Institution in an Anti-Institution Era | Better in Person · 2026-09-02Bloomberg News reported that the group had previously offered more than $50 billion for the company. The offer is a fraction of the roughly $360 billion valuation PayPal commanded back in 2021. Advent, Stripe, and Block first approached PayPal together back in April, but Block exited the group before Stripe and Advent submitted their offer. PayPal, which was founded in the late 1990s, has faced mounting competition from Apple Pay and Google Pay in recent years. Now for a look at a few other articles that are trending. The US is reportedly close to striking a massive deal for Venezuelan oil fields. This next one is the topic in today's edition of the Wall Street Breakfast newsletter: What will Warsh say at Jackson Hole if he doesn't offer guidance? And Michael Burry buys Nvidia calls as a hedge and expands shorts on Oracle and Palantir. Gotta tell you, that article has about 130 comments thus far. I have one item for you on our Catalyst Watch for the day: shareholders with Crynetics Pharmaceuticals will vote on the takeover offer from Vertex. On Wall Street, stock index futures are in mixed territory. Crude oil is down 0.5% at $83 a barrel. Brent crude is down 0.5% at $88. The FTSE 100 is up 0.2%, and the DAX is up 0.6%. And I have a couple of items for you on our economic calendar. At 10:00 AM, consumer sentiment. Also at 10:00 AM, something we've been talking about all week. Federal Reserve Chairman Kevin Warsh is scheduled to speak at the Jackson Hole Symposium. That's it for today's Wall Street Breakfast. Thanks for listening. To take full advantage of Seeking Alpha with coverage on significant stocks and ETFs, become a premium subscriber. Check out seekingalpha.com/subscriptions. I'm your host, Julie Morgan. Go out and make it a great day.
Julie Morgan · Wall Street Breakfast · Marvell beats, street shrugs · 2026-08-28Well, it's that. And also, let's just, let's just say this right now. I'm looking at the Barclays numbers in front of me. OpenAI dies tomorrow. That is, let's look at it, 28% of Amazon's AI revenue next year. Gone. Goodbye. Just, just evaporated. And indeed, 27% for the next 2 years as well. If Anthropic and OpenAI, if one of them dies, what, that's like 20%. Let's see, OpenAI is, uh, well, it looks like about $12.5 billion of Google's 2027 AI revenue, just like gone, evaporated. And otherwise, their other AI revenues are pretty flat. It looks like Vertex, Vertex kind of increases, but again, that's through selling Anthropic and OpenAI's models. It's— I don't think people realize that there's a great chart coming in this article as well. How much of Azure's revenue as well is just OpenAI spend? And it's, let's see, let's take, let's take a look at this chart of Azure revenue as well. 62%, let's see, yeah, it's about 72%, 68 to 72% in fiscal year 2026 and as much as 74% in 2027. If these companies die, they lose most of their AI revenue, evaporated. And that's not including the amount of money that comes through Azure AI renting out their models.
Ed Zitron · Better Offline · The AI Demand Bubble with Ed Elson · 2026-08-05Yeah, well, the hidden gem style, I've been doing this with Tom for a long time and benefited from so much of his thinking, but that hidden part Travis, and that can mean a lot of different things. It could mean under— it's basically underappreciated or undervalued assets from what the market sees. Asset strength and brands, exceptional leadership, financial power, intangible assets, and growth opportunities. Those are— and, and, and it's a mixture of both. But I think about that combination. What is the market really not seeing? And I think one interesting spot is something like a retailer that really doesn't quite get the attention that other retailers might in the US. And that's a company like Aritzia, which had earnings this week. They were outstanding. Think about the growth profile. It's a Canadian women's retailer that does this very accessible medium to high fashion retail, and it's just doing things in a different way with that leadership team. But it was Canadian-based, and now they're moving more aggressively into the US, very profitably into the US. And so that kind of story When you think about getting in, into an early growth story like that, that's the kind of hidden gems that I really love to find and love to invest in. I look about our history of the same thing with a company like Buffalo Wild Wings, which was very early on in trying to grow out chicken wings, buffalo wings in a way that really wasn't done at scale, or a company like Chipotle and Starbucks many, many years ago. So those Growth opportunities, and then you have a company that's taking advantage of those. So I think still a company like Aritzia and retailer, I see some in the biopharma space, whether it's a company like Vertex or Krystal Biotech, symbol K-R-Y-S, that is doing some really cool things when it comes to treating butterfly skin disease. And they have a very proprietary process and platform, and that's what I'm trying to find and that are not really quite appreciated by the market that is constantly talking about large-cap tech stocks.
Andy Cross · Motley Fool Hidden Gems Investing · What History Teaches Us About Today’s Market · 2026-07-10Well, you know, they clearly are going to be very aggressive on the pricing. I mean, they talked about that today. Right. And so my sense is like what you're going to start seeing is that people don't need to operate at the frontier and you have a lot of use cases that work just fine with and basically are just good enough, right, with some legacy model. And so it just comes down to then, okay, well, that's commodity. And so is it priced like a commodity? Yeah. Like, like if you, if you think about like, and, and also like, I think we're gonna see a lot more, um, uh, like, uh, people relenting from needing to be on the, on the frontier when they've built stuff using a model that then gets upgraded. And the whole idea there is like, well, am I gonna upgrade this tool because I have to adapt it to the new model, right? Like if I'm, if I, uh, if I, you know, cuz essentially like the model name, like if you, if you use like Vertex AI, right? You, you just gotta model name as a variable. Like you're sending this, you know, system prompt to, to, you know, Google, but like, it's just a variable. You could swap that out in 30 seconds. But the, the fact of the matter is you're sampling from a new distribution if you do that, and it's gonna change the output. It's gonna qualitatively change the output, and it might change it in quantitative ways that, like with retention and engagement. And so the thing is like, okay, well now we're talking about this big cycle. It's a new product development cycle because I have to adapt this product that I built to this new model and the output that it provides, right? And maybe it was working perfectly, it was working exactly as I expected it to before. Mm-hmm. Now I've got to invest a bunch of hours, much of engineering time in adapting it to this new model. So even if it's just a swap of a variable name, it's still a whole lot of testing, QA, determining like how that impacts long-term retention, going to A/B test. So my sense is like you're going to see a lot more people just saying, no, this works fine, this is perfect. And getting more, getting like more robust output. Let's say the token price is exactly the same. Getting more robust output wouldn't benefit me. Why am I going to invest the resources into adapting to the new model?
Eric Seufert · TBPN · Meta Releases Muse 1.1, GPT-5.6 Sol Reactions, New Robot Hand Alert | Eric Seufert, Bernt Børnich, Josh Lindgren, Jeffrey Morgan, Thibault Sottiaux, Sean Frank · 2026-07-09Yeah, it's funny. Um, so prior to Google, I, uh, was both doing my undergrad, master's, kind of concurrently in computer science and stats focused on, um, resilient, like machine learning algorithms specifically for like privacy preserving use cases. Then I went over to Google and I worked on across a number of different teams, including Google Research, where I was working on, if you remember, the early days of agents actually looked like the Google Assistant as well, which was like this embedded assistant in all these different devices. We were working on, um, more deep learning focused use cases within the assistant. So that was my first exposure to, uh, agents more than a decade ago at this point. And then the second, uh, uh, area that I worked was actually on Google Cloud's AI platform. So, um, I remember we hadn't even named it Vertex AI when I was there, which is the name of it now. Uh, but we were really forming kind of the initial basis for the fact that more people were starting to do machine learning and AI workloads on the cloud. And how do we build a platform that people could actually use across all that? So that's what I was working on until the end of 2020, the pandemic, when I met, uh, my co-founders who had built these frameworks at Uber to make it easier to be able to use machine learning internally and build these applications. These frameworks got popular, so they open-sourced them. As things happen to be open-sourced, they got popular externally. And so we all decided to team together and start a company around it, which was Predabase in early 2021.
Devvret Rishi · Eye On A.I. · The Biggest AI Security Problem Isn't the Model. It's This. | Devvret Rishi · 2026-07-07We're a layer above it. My point of view is that there's lots of good orchestration tools that exist out there. There's lots of good platforms to build agents today. You know, each of the hyperscalers has their own, um, agent platforms, including orchestration. So Vertex AI has, um, you know, their own agent, uh, ADK. You have Bedrock inside of AWS with Agent Core. You have a suite of tools inside of Azure. Salesforce has Agentforce. ServiceNow is an agent platform. Plenty of good ways now to be able to build agents. What we've noticed is actually hard is how do you place some level of controls around making sure that the agents, regardless of where they're built, actually have some level of consistent runtime guardrails on what they can and can't do. And this actually, I think, is a very tricky problem because agents look less like conventional software and inside of an IT organization, a little bit more like what you or I might be able to do. And so like my agent, as an example, which, uh, my agent, the one I use most heavily is out of personal uses like Claude Code and Claude Cowork, my agent has access to Salesforce and my agent has access to email because I use it to write emails and I use it to summarize opportunities in Salesforce. But, uh, what I don't want it to do is take something that's sensitive from Salesforce and put it out and send it out in an email autonomously, right? And so that's just one thing that happens with Claude, uh, Code and Cowork that I need to secure. But if you were using a different agent harness, if you were using OpenAI's Codex, it'd be the same concern. And so my view is that there's lots of different good systems. They're building excellent ways to integrate into the different areas, good harnesses, good orchestration. Registration frameworks, ways to be able to run in the cloud, like Anthropic launched Managed Agents yesterday. But how do you have a consistent way to be able to monitor, observe, and actually runtime guardrail any of the things that those agents are doing? That's where we come in. So if you think about like the core layer, it's like your models and your MCP tools or your other API tools. The layer above is going to be like your agent platform builders plus orchestrators. And then what we think about is like having the security kind of umbrella across this with the agent cloud.
Devvret Rishi · Eye On A.I. · The Biggest AI Security Problem Isn't the Model. It's This. | Devvret Rishi · 2026-07-07That's one of the settings where I think it's the most compelling. Um, and frankly, I think that this would be true for most enterprises, sort of similar to how you saw in the cloud era, like you had organizations adopting multi-cloud. I think you're seeing even more of that in the, um, AI era because these things are coming out so quickly. I think where Rubrik Agent Cloud actually really shines is you have multiple different agent platforms. Platforms that you're using. At Rubrik, we ourselves have multiple different agent platforms. We in fact have like a couple different coding agents that we, uh, have, including Cloud Code, but we also build agents on top of Vertex AI. You know, we have agents that are, um, built in a number of different— and also other third-party tools that I'm not even mentioning. And so the question is like, how do you secure across all of this? This is where I think, you know, our agent cloud provides a tremendous amount of value really at day zero, because oftentimes what people want to know even at the beginning is what are the agents that are even deployed in all these different ecosystems. Systems. So we can plug in, provide the automatically discovered and scanned inventory, and then allow you to start to get the system of control across all of those.
Devvret Rishi · Eye On A.I. · The Biggest AI Security Problem Isn't the Model. It's This. | Devvret Rishi · 2026-07-07Constellation's Ray Wang. Samsung's quarterly profits surged 19-fold and we're seeing chipmakers fall in early trading. SpaceX also falling. Well, now down about a third of a percent. The Elon Musk-led company is joining the NASDAQ 100 index today. China's Deepseek is developing its own chip to help power artificial intelligence systems. This is according to Reuters, which says the chip is designed for inference, the stage in AI when models are trained and begin generating responses to user queries rather than for building new models themselves. And DL News, shares of Kinetix Pharmaceuticals are soaring Vertex Pharmaceuticals has agreed to buy the endocrinology-focused biotech company for $10 billion. The purchase price of $85 per share is a 102% premium over where Krenetic shares closed yesterday. And the American dream of a World Cup is over, the U.S. losing to Belgium 4-1. The game was marred by controversy over FIFA's decision to allow the U.S.'s Fuller and Belligan to play. Again, tech futures leading the way lower, with the NASDAQ futures down 1.1%, S&P futures down 2/10 of 0.4%. Dow futures up a quarter percent. 10-year Treasury yield 4.49%. And IMEX crude oil is higher, up 9/10 of a percent at $69.17 a barrel. Brent at $72.71. And that's news when you want it with Bloomberg News Now. I'm Karen Moscow, and this is Bloomberg.
Karen Moscow · Bloomberg News Now · Trump Arrives in Ankara, Qatari LNG Ship Struck in Strait of Hormuz, More · 2026-07-07Yeah, so every quarter we're deploying vastly more compute than the prior quarter and there's more data centers built than the prior quarter. This year there's going to be 20 gigawatts, even accounting for the delays. And next year there's going to be more than 30 gigawatts accounting for the delays. Of course delays happen on everything. Anything hardware can have a delay. That's just the reality of life. Are we going to have a compute crunch for the rest of our lives? It depends on what happens with models. The TAM for Mythos 5, Fable 5 is not just 2x that of Opus. The model is so much better and it can do so many more tasks that the TAM for it is way larger than that. And yet compute in the world did not double in the last 6 months from Opus or maybe 7 or 8 months since Opus 4.5 launched to now. Huge 4.6, 4.7, 4.8 were improvements, But Fable and Mythos were a huge step function improvement. The world's compute did not double or quadruple or whatever in that same time frame. But the demand for useful tasks that can be done by AI, the number of useful tasks and the value of them that can be done by AI has. So now the question is what happens? Well, obviously Anthropic in Q2 is profitable. They're net income profitable excluding stock-based compensation. I think by Q3 they may even be profitable including stock-based compensation. That's how profitable they're getting. And their margins on an OPUS token, at least OPUS-48 token, is north of 80% for the API price. They've got a lot of deals where their total corporate gross margins get clawed down a little bit because of how they do Bedrock deals and Vertex deals and things like that. But ultimately, their per-token margin is so high Well, then if you don't have the capability to pay, ultimately every GPU they buy at an above-market rate, they also bought GPUs at above-market rate from SpaceX, which is below the rate of Google, but that's because they signed earlier. It's something that other companies, maybe a venture-backed company or a company that's not really got positive margins, can't necessarily do. What is the cost-benefit ratio?
Dylan Patel · Training Data · Why Hardware-Software Co-Design Is AI's Real 100x: Dylan Patel of SemiAnalysis · 2026-06-30Uptime page? I don't know what you call that. Email me if you know what they call that, or I could just look after this. Going to keep going though. Nevertheless, people want name brands. People like paying for a company that ostensibly works on this and keeps it updated. How true is that? Who knows? I don't think people are buying LLMs because of particularly sensible reasons anyway. But what might change things would be Microsoft, Amazon, or Google offering GLM 5.2 via their Foundry, Bedrock, or Vertex AI platforms. And even then, it's hard to tell if customers are excited about AI or just paying OpenAI and Anthropic to use buzzy models. Really isn't clear. We should know by now. All that being said, now does feel like the perfect time for a shift towards open source models, if only because everyone is crapping themselves about costs. Everyone is really freaking out. And I think that this is going to be the time when we start separating the wheat from the chaff, the people that actually like this stuff versus the people who just feel good because they're doing AI and they don't know what else to do with their time and they like forcing people to use it. I'll get to who actually does in a minute. Support for today's episode comes from Square, the business platform that helps sellers become neighborhood favorites. Whether you're gearing up for a busy season or just trying to keep up with everyday demand, Square keeps your business running smoothly, from payments and POS to online orders, inventory, staff, and more, all in one place so you can focus on your customers, not your to-do list. I like Square a lot as a customer, and from talking to various business owners who use it, everybody seems really happy with how easy and intuitive everything is. Business owners don't have to juggle a bunch of tools, and their hardware and software are designed to be easy to use, so day-to-day options feel a lot simpler. With 3 clear plans depending on the tools you need, sellers who use Square see 9% higher sales on average. Square helps you run your business more smoothly, bringing payments, operations, and insights together in one place so you're ready for whatever's next.
Ed Zitron · Better Offline · Monologue: Tokenpocalypse Now · 2026-06-26Yeah. So on desktops you got all These new graphics APIs, I don't know. On Windows I guess you have like the latest version on DirectX, which is probably different than the previous ones. On Mac you have Metal. Metal, yeah, they all. I think you've spoken in the backend then. So that's a new type of graphics API where in OpenGL you're basically firing commands. You're saying like, hey, set this texture now set this array of vertices and now draw this primitive using those vertices and then it executes those in order and you have to wait for it with WebGPU Vulkan, it's a bit different. You say like, okay, here is a texture, a vertex buffer and now cue the rendering command with those. So it's More asynchronous where WebGL you really like, you're waiting for the result, you're waiting for something to draw. With GPU you can batch a whole bunch of commands and then say, okay, send this to the GPU to render. I don't want to wait for the result. I'm going to do other stuff. So it's a lot faster that way. You're spending a lot less time waiting. But also much more complex API to use. But of course, on the web. Yeah, you have engines taking care of that. Yes, sure. But for the engines it's also a bit complicated. I know, because it's such a different way of rendering where you're not waiting for stuff, you're batching all these things. They usually have to quite change their pipeline to make use of this.
Eric Dabelbor · Software Engineering Daily · Web Native Game Development · 2026-06-04I do think it's fascinating. The returns to the frontier, all the economic returns to AI at the model layer, not all of them, but an overwhelming amount of them have been at the frontier, which is surprising to me and I think it's been surprising to a lot of people. This is one of the most important questions to be answered and you need to have a hypothesis on it as an investor. Are frontier tokens going to continue capturing the overwhelming majority of economic value created at the model layer? And it is surprising. I remember when Gemini 3.1 Pro came out, it was mind blowing to me. It was so good. Today it's intolerable. There's probably a little bit of a dynamic where companies prototype with frontiers, then when they put something into production. You're hearing a lot of people do use vertex or open source. But still it is a fact today that the overwhelming majority of these economic returns come from frontier tokens. And that's surprising. And whether or not it continues I think is a very interesting question. And I'm much more open minded to that having had the experience I've had with Gemini 3.1 and then Opus and then I do use Grok 4.3 a lot. It is on the Pareto frontier. The companies that are on the Pareto frontier are, and this is by the way a big change and a consequence of what we talked about last time. Google losing their per cost token leadership as a result of making very conservative design decisions with TPUV8 to try and take it away partially from Broadcom and Nvidia continuing to make aggressive choices. But Google dominated the Pareto frontier. The Pareto frontier being intelligence versus cost. And I think this is the most important thing to look at to analyze AI labs. Google dominated that nine months ago. Every point on the Pareto frontier, OpenAI, XAI and Anthropic were inside of them. Now the Pareto frontier is dominated by anthropic, OpenAI and then Grok 4.3 is on the Pareto frontier. It's clearly the best, lowest cost, 500 billion parameter model and then Gemini 3.1 is hanging onto the Pareto frontier. And if I were to bet or bet that they're subsidizing that out of pride, I would just say a violation of Richard Sutton's bitter lesson is for sure the biggest risk to this trade to all of AI. The closer someone is to AI, the more skeptical they are this will occur. One thing I think contributed to weakness in March was a much more stupid version of Deep seq, which was this thing called Turboquant. And Turboquant is some Google memory optimization that was written up in a paper a year ago. And then during the middle of an agreement while Google was negotiating with Micron, Samsung and Hynix to sign some LTA that would lock in really high prices for a long time, they released this. What people do is always more important than they say and they just kind of publicize it on X and it goes viral like, oh my God, DRAM is cooked. Here's this. DRAM optimization. I was unable to find a single AI engineer on planet Earth who believe that turboquant would have any impact on DRAM demand. But nonetheless, a violation of Richard Sutton's bitter lesson. More compute will always outperform human algorithmic ingenuity. More compute and data. Chinchilla optimal. Beyond chinchilla optimal. I guess what people increasingly do today, that's a real risk, man. The people who are building these models are skeptical of that risk. The reason I am a little less skeptical is I think we are very close to ASI. And who knows if the bitter lesson holds for 400 IQ models. Maybe we get a temporary period where these. If you get to asi, the first thing it wants is probably to be smarter and have more resources. How does it do that? It makes itself more efficient. I think that is an actual risk. The bitter lesson, literally, I believe, includes humans in it. So we're about to find out whether the bitter lesson we'll find out if it applies to 300 IQ, AIs, then 400, then 500 and 600. And at some point we may have like a temporary violation of the bitter lesson based upon AI and asi.
Speaker E · Invest Like the Best with Patrick O'Shaughnessy · Gavin Baker - Watts and Wafers · 2026-05-20Still a ball in the air kind of game. BCG had this theory, I think I talked about this once before, called the rule of three, where they've shown time and again that any stable, mature competitive market evolved to a 4 to 2 to 1 ratio of market share for basically 90% of the market. So there's a market leader that has four times the market share of the second place, that's two times the market share of the third place. This is the case in pretty much every mature kind of competitive market. So you can kind of think about AI probably evolving into a consumer market and an enterprise market. OpenAI, even if they're not at a billion, they're still at 900 million weekly users, which is well ahead of whatever Claude is at. I think Claude is like probably sub 100 million Sachs, you may know, and then Gemini is probably closer to them at 700 to a billion. Somewhere in that range, probably pretty neck and neck with OpenAI. So, you know, the consumer market looks like it's trending towards a ChatGPT Google fight for first place and second place and then probably anthropic in third place and maybe Elon emerges and takes off enabled by his compute capacity. And then the enterprise market is a little bit of a different story and that's its own market, which is kind of anthropic or probably Google in the lead. Actually, if you look at all the Vertex use, Google claims that 75% of GCP customers are active users of Vertex. So there's probably a pretty sizable market share that Google's captured on the enterprise side as well. This is also probably why Google stock has absolutely ripped over the last couple of months, is they're literally in first place or fighting for first place in enterprise and consumer. But I still think that there's a lot of opportunity to Chamath's point about the compute and energy capacity constraints in improving how we actually scale and deploy models in both the enterprise and the consumer setting. And it is such early days. And I just want to highlight this paper that came out from MIT from these two scientists and these guys published a paper on pruning techniques in neural networks. This paper showed that you could actually reduce the size of these networks by 90% and get the same accuracy out by pruning very large models down to smaller models. And then you can make a selection on which model to run for inference. And by doing this you can actually reduce inference costs by 10x. You can get 10x the output per energy unit that goes into the data center with no loss of accuracy. And so it's a really interesting, call it algorithmic technique that can be applied to the existing large models to actually make them much lower energy use. So if you think about it, you're firing up a very large model to answer a very simple question, you can actually prune away that model. Now, this is probably going to be the case in AI applications as it is in traditional Google search. There's a long tail of searches, but there's a few searches that account for a large percentage of search volume. It's like, what is the weather, what are the movies times, what's the stock price? There's a certain set of things that make up the bulk of consumer energy and there's probably a certain set of things that probably make up the bulk of coding output as well. And so if you can get that 80% of searches or chat interfaces or coding requests reduced down through pruning techniques to smaller models, and then you have a whole set of smaller models that can be called dynamically and you reduce inference cost by 90%, you can make much more use, call it 10 times the use on data center and energy capacity than we can today. So I would argue that we're still in the very early days of getting efficiency in terms of output and tokens, and we're just in the very kind of early stage of that, which also unlocks the opportunity for guys like Elon to reinvent how this is done and potentially compete pretty aggressively.
David Friedberg · All-In with Chamath, Jason, Sacks & Friedberg · OpenAI Misses Targets, Codex vs Claude, Elon vs Sam Trial, Big Hyperscaler Beats, Peptide Craze · 2026-05-01One of the sales folks on Quad Code, he realized he was making these repetitive decks over and over and over again. And so he actually has this web app that he built with the examples of the core Quad code decks that we know work well. So like a 101, 201 and mastering quad Code. And then he has a way to input specific customer context that pulls from salesforce that pulls from gong that pulls from other notes so that we can customize the decks for specific customers. And so pull out things like, okay, this customer is using like Bedrock or Call for Enterprise or Consul which affects what features are available to them. It will pull out things like, okay, this customer is concerned about like the code review stage of the SDLC and so we'll add a slide about our code review features there. It'll pull out things like, okay, this customer needs to be like HIPAA compliant or needs XYZ security controls. And so we'll make sure to add a slide or two in their deck about that. And then for example, if this is a customer that's on Vertex or Bedrock and doesn't want to use cloud for enterprise, then we'll just take out some of the slides that are cloud for enterprise only features. And so normally this is like manual work that could take 20, 30 minutes. And so people either like spend that time doing it or they'll just decide not to do it and use the general deck. With this it takes like a few seconds and you get a tailored deck.
Speaker A · Lenny's Podcast_ Product _ Career _ Growth · How Anthropic’s product team moves faster than anyone else | Cat Wu (Head of Product, Claude Code) · 2026-04-23