Why Everyone Is Obsessed With Claude Code

Today on the AI Daily Brief why Everyone is Obsessed with Claude Code Right now the AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright friends, quick announcements before we dive in. First of all, thank you to today's sponsors, kpmg, Superintelligent and zencoder. To get an ad free version of the show, go to patreon.com aidaily brief or you can subscribe on Apple Podcasts. And to learn more about sponsoring the show, send us a note at sponsorsideaily Brief AI now this was one of those days where it was more about the discussion going on in the space than about the news in the space. And this whole Claude Code Opus 4. 5 inflection point has been really brewing for a number of weeks now. Because of that, I ended up doing the whole show about that. We will be back with our normal headlines in domain format tomorrow. And the last thing before we dive in, after having so many of you sign up to participate in this AIDB New Year, I have gone ahead and launched a free community for people who are participating in that and just for discussion of AI building in general now. Additionally, this is imagined as primarily about the AI DB New Year's project, but we'll see if that changes. I'm certainly thinking about it as a community that might go on longer, which is why I've called it AI Operators. If you sign up@aidbnewyear.com, there's a link to the community on the program page and you can also find a link from the main website page. AIDAILY Brief AI welcome back to the AI Daily Brief. Ever since Anthropic released Opus 4.5, there has been a sense that we've fundamentally shifted in terms of what AI coding is capable of. In point of fact, it's been actually a combination of model advances as well as the tools and platforms through which we access them, like Claude Code, that have contributed to this sense of something fundamental having shifted. But boy, is there a sense that something fundamental has shifted. On January 4, Simon Willison posted It genuinely feels to me like GPT5.2 and Opus 4.5 in November represent an inflection point, one of those moments where the models get incrementally better in a way that tips across an invisible capability line where suddenly a whole bunch of much harder coding problems open up. OpenAI's Greg Brockman agreed. Reposting and sharing. It does feel like models have just cleared a threshold of utility in software engineering. AI builder and YouTuber Theo writes, feels like we're actually in one of those everything is about to change moments, and the din is loud enough that publications like Axios are picking up on it and writing stories, even though there's no particular news that the conversation is based on. Now, to be fair, some called this early from the moment that Opus 4.5 launched, Dan Shipper and Every have been very much on the tip that this represented something different. Back in December, he wrote a piece called Opus 4.5 collapsed six months of development work into one week, and he started with this pronounced declaration. Humans, he wrote, have always had two main intuitions about what we'll find when we travel to the end of the earth. 1 An edge where the known world falls off into nothingness, chaos, or monsters. 2. A new vista where unexplored, lush, and perhaps perilous territory extends toward a new horizon. The first is terrifying, a place to be avoided. The second represents possibility and an entirely new world. These days, most new AI model releases are incremental. Sometimes, though, a new model brings us right up to the edge of the known and allows us to take a peek at what lies beyond. Is it nothingness, Dragons, or a new Horizon? Anthropics? Opus 4.5 is one of those models, and I've been peering over the edge for about a week now. Here's what's over the horizon. We are in a new era of autonomous coding. You can build astonishingly complex apps without looking at a single line of code. Two Agent native apps are now possible. You can use Opus 4.5 as a general purpose agent to power your app's features. This turns new features into an exercise in prompt writing rather than coding. All of this led Dan to declare the birth of an infinite vibe coding machine. And yet for many, it seems to have taken a holiday of hacking for all of this to really click. MidJourney's David Holz wrote, I've done more personal coding projects over Christmas break than I have in the last 10 years. It's crazy. I can sense the limitations, but I know nothing is going to be the same anymore. That post had over a million views. Even Elon Musk responded to it, saying, we've entered the singularity. Flo, the founder of Lindy, agreed, writing, this is definitely how I felt these past two weeks, getting AGI shock every day. Chat with the Lindy team on Friday. AI coding is solved. At this point, if you think you know what's going to happen next, you're very confused, AI News curator Andrew Curran wrote. People who work in code have felt a sea change, but the jump comes from frontier models inside their coding shells, not from the shells themselves. Gains are currently concentrated in one area code, but the driving force is intelligence, and the more time you spent on the Internet, the more you saw people sharing specific projects. Pietro Scarano wrote, claude code isn't just for coding. I fed it my raw DNA data from an ancestry test and used it to find health related genes I should keep an eye on. The file is massive, but its ability to search what matters makes it possible. DeepMind's Xiaoma wrote, I'm this close to throwing in the towel and giving coding agents my entire financial history. Been using beancount for personal accounting for years. Every January, download statements parsed with custom scripts. I've been manually updating forever. Started having AI upgrade the scripts 2 years ago. Clunky workflow, but it worked. Every year the AI gets a little better. Maybe this is finally the year to just give it all. Lenny Ryczycki of Lenny's podcast shared his vibe Code Project of the Week, a soundboard of his toddler's cutest words. He took a $20 microphone off of Amazon, used a script to slice up each word, Claude code to clean up the audio quality and lovable to build it. Google's Logan Kilpatrick wrote, I spent the last week building what could easily be a $100 million venture backed business. It's truly wild how much leverage AI gives you. 10X's Alex Lieberman wrote, I've never taken a coding class, I've never shipped a piece of software pre AI, yet I'm now creating vector embeddings of 812,918 iMessages from the last 5 years of my life. The seismic shift happening before our eyes is truly once in a lifetime. And what's important to note with all of this is that these are not new people coming into AI for the first time and getting a glimpse of what's possible. These are singularly the most active, frankly proficient AI users in the world, and they are increasingly shocked at what's possible now. Ex Google engineer Rohan Anil wrote, I feel if I had agent decoding and particularly opus, I would have saved myself the first six years of my work compressed into a few months. Maybe the most notable tweet of the whole period came from John Adogan, a principal engineer at Google, who wrote, I'm not joking and this isn't funny. We've been trying to build distributed agent orchestrators at Google since last year. There are various options not Everyone is aligned. I gave Claude Code a description of the problem it generated. What we built last year in an hour. It's not perfect and I'm iterating on it, but this is where we are right now. If you are skeptical of coding agents, try it on a domain that you are already an expert of. Build something complex from scratch where you can be the judge of the artifacts. In other words, as many pointed out, even Google engineers are blown away by Claude code. So whatever is going on, it's clearly widespread. In fact, when Alex Lieberman again posted on January 2 that he wanted to start a community dedicated to Claude code, more than 7,000 people signed up. So many, in fact, that he's now hiring a program and community manager to run that community. Now, one of the key themes is that Claude Code is actually misnamed. Investor Nikunsh Kothari wrote Claud Code should be thought of as Claude Computer because that's what we're getting a glimpse of. Alien intelligence that has human tools, browser file system, terminal commands and others with generalized ability to do any task. Terminal UX is all that's stopping folks from utilizing this all the time. But the GUI will come either through Anthropic directly or someone packaging the agent SDK neatly. And indeed, one of the things that you see is that the people who are really getting the most out of Claude code are using it for everything. Alex Vanovic wrote started using Claude code for managing my team admin, program management, meetings, agendas, etc. Basically, Claude is my chief of staff now. Marketing and crypto entrepreneur Amanda Cassatt wrote, I'm now coding air quotes all day with Claude code. If you haven't tried it, you don't understand. It literally takes over your computer and does everything for you. There was something wrong with the sound on my Mac and I used Claude code to fix it rather than open settings. She also tweeted, claude code is leaving everyone in the dust right now. If you haven't tried it, you need to. Now you just speak in English like it's the starship enterprise. It takes over your computer and does everything for you. It's not like asking a chat model to help you code, where it puts code in the weird little box. It's not a copilot checking over your code or teaching you how to code, unless you want that. No copy and pasting between windows. No learning to code, no learning curve at all. Chicago booth professor Alex Emas wrote to those who are curious about the hoopla, Claude code is not just for code. It's an AI agent that can essentially act like a combination of personal assistant and colleague that has access to your machine. Need your files organized? Need to create a better calendar app and organize your life? Need a sounding board for a presentation that's coming up? Need to organize your notes that are all over the place for that presentation. Claude code will do that. Just write it out in plain text and make it happen. Alex also agreed that the word code in the name will slow adoption. Boris Czerny, who built Claude code, also shared how he used it. And it gives you a sense of just how comprehensively the power users are putting this tool into practice. Boris writes, I run five clauds in parallel in my terminal. I number my tabs 1 to 5 and use system notifications to know when a Claude needs input. I also run 5 to 10 clauds on Claude AI code in parallel with my local clauds. As I code in my terminal, I will often hand off local sessions to the web or manually kick off sessions in Chrome, and sometimes I will teleport back and forth. I also start a few sessions from my phone from the Claude iOS app every morning and throughout the day and check in on them later, and so on and so forth. You can see that this guy's got 15 different instances of CLAUDE running at any given time now. One of the things that's clear is that the new capabilities just feel different. 37signals Jason Fried writes, AI workflows are technically impressive, but there's a deeper reason people are really amped about AI agents. This isn't just new tech, it's new psychology. Until now, very few people have known what it feels like to delegate to total competency. If you manage great people or lead great teams, you know how it feels to put someone in charge who will get it done, get it done right, and get it done without drama. That kind of delegation is pure joy. Delegating to competency lets you forget about it completely. That's real leverage. And now anyone can experience that. Everyone can feel it, and it feels effing great. That's a big reason why the excitement is real and fully justified. Cursor's head of design, Rao Liu, called this a software renaissance. The printing press didn't make monks faster at copying manuscripts, he wrote. It made copying obsolete. It created new worlds. Anyone with an idea can now make software. But now the tools are no longer the true bottleneck. We are. Even some folks who might have previously been skeptical are coming on board. Theo again wrote, if you think AI coding is a fad, I get it. I felt the same way about a year ago. If you tried these tools out during the Copilot autocomplete era, I understand entirely why you wrote them off. Things have changed a lot since then. These tools are here to stay. I don't fault anyone for feeling this way. The rate of progress seems impossible, and keeping up on it all is tedious and obnoxious. Lots of people are selling snake oil right now. You might have accidentally bought some yourself. I know I have. That doesn't make the trend irrelevant. Every few months there's an absurd leap in capability. Six months ago I was blown away with GPT5. It could do decent designs and navigate a giant code base. Awesome. Now OPUS can complete large feature work by itself. Actually insane. To be blunt, all of the best devs I know are using AI heavily in their work. Once you see what these tools can do, it's hard to talk about it genuinely without sounding like a paid shilling. The gap between devs evangelizing AI and devs skeptical of AI is getting massive. People who insist these tools are useless, I get it, they felt useless just a few months ago, but your takes are dated. AI is fundamentally changing how we work. David Henemeyer Hansen shared a very similar sentiment in a recent blog post called Promoting AI Agents. He wrote, I never really cared much for the in editor experience of having AI autocomplete your code as you were writing it. That was the original format pioneered by GitHub's copilot and cursor, but it left me code. But with these autonomous agents, the experience is very different. It's more like working on a team and less like working with an overzealous pair programmer who can't stop stealing the keyboard to complete the code. You were in the middle of writing with a team of agents. They're doing their work autonomously and I just review the final outcome, offer guidance when asked, and marvel at how this is possible at all. Yes, I'm ready to give the current crop of AI agents a promotion. They're no longer just here to help me learn, answer my questions, or check my work. They're fully capable of producing production grade contributions to real life code bases. I get that some programmers are eager to tune it all out. The hype drones on relentlessly. The most fantastical claims are still far off from being substantiated, and there's real uncertainty where this will all leave the profession in the future. But that's still not reason enough to miss out on this incredible moment in human and computing history Foreign friends if you've been enjoying what we've been discussing on the show, you'll want to check out another podcast that I have had the privilege to host, which is called you can with AI from kpmg. Season one was designed to be a set of real stories from real leaders making AI work in their organizations, and now season two is coming and we're back with even bigger conversations. This show is entirely focused on what it's like to actually drive AI change inside your enterprise and and as case studies, expert panels, and a lot more practical goodness that I hope will be extremely valuable for you as the listener. Search you can with AI on Apple, Spotify or YouTube and subscribe today. Today's episode is brought to you by my company, superintelligent in 2026 one of the key themes in enterprise AI, if not the key theme, is going to be how good is the infrastructure into which you are putting AI and agents Superintelligence Agent readiness audits are specifically designed to help you figure out one where and how AI and agents can maximize business impact for you, and 2 what you need to do to set up your organization to be best able to leverage those new gains. If you want to truly take advantage of how AI and agents can not only enhance productivity, but actually fundamentally change outcomes in measurable ways in your business this year, go to BeSuper AI. If you're using AI to code, ask yourself, are you building software or are you just playing prompt roulette? We know that unstructured prompting works at first, but eventually it leads to AI slop and technical debt. Enter zenflow. Zenflow takes you from vibe coding to AI first engineering. It's the first AI orchestration layer that brings discipline to the chaos. It transforms freeform prompting into spec driven workflows and multi agent verification where agents actually cross check each other to prevent drift. You can even command a fleet of parallel agents to implement features and fix bugs simultaneously. We've seen teams accelerate delivery 2x to 10x, stop gambling with prompts, start orchestrating your AI, turn raw speed into reliable production. Grade output at Zenflow Free. Now. Honestly, at this point the excitement has gotten so loud that there are some who are trying to show where there are still challenges. OpenAI uber researcher Noam Brown wrote in one memorable debugging session with Claude Code. I asked it as a sanity check what the expected value would be of an always fold strategy when the poker player has $100 in the pot. It told me that according to its algorithm, the expected value was around $93. When I pointed out how strange that was, hoping it would realize on its own that there's a bug, it reassured me that $93 was close to $100, so it was probably fine. Codex was not much better on this and ran into its own set of interestingly distinct bugs and algorithmic mistakes that I had to carefully work through. J.F. puget, who does machine learning at Nvidia, did a Claude code test and came to the conclusion I am more and more convinced that its value is where you know the best. The rest is vibe coding, which can be fine for personal projects, but definitely not fine for product development. Now, to be honest, I think this is in and of itself a temporary state and a temporary opinion. I think a lot of the things that people think currently are barriers on pressing send on production software right now will just be blown over in the future with builders content to just try to fix the problems when they come up. Which is not to say that I get every use case that's being shared right now. Entrepreneur David Siegel shared an admittedly extremely cool project called Claude Canvas where he basically gave Claude code the capability of creating a generative UI based on the particular task at hand. Now this was undeniably super cool and technically awesome. However, the three examples that he gave of scheduling a call with a cto, booking a flight, and writing emails just did not resonate with me at all. Specifically, the user experience of doing those things in Claude code, as opposed to the highly optimized UXs that have been created for them, just doesn't make sense. From where I'm sitting, calendar scheduling is not some massive pain. You just go into your Google calendar, pull up your teammate's schedule with their email, and pick a time that works for you. Both for flights by no stretch would I say that flight discovery UX is perfected, but seeing Claude code reapproximate the seat layout of an airplane, rather than just using the version of that on, for example, the Delta app, just seems really not time saving to me in any meaningful way. What's interesting though is that it's very clear that David is not using Claude code in a one off to one off way. He's using it very comprehensively as a personal agent across a huge span of his work and interactions with the world of the Internet. And so what strikes me is that while it's possible that he's just one of those folks who's really excited about a new tool and is exploring all the different use cases, even though some of them won't make sense for that tool in the long run. It strikes me as perhaps more likely that David is in the midst of a transition to a post UX world. A world, in other words, where user experience matters less because the user, which is us, a human, isn't actually having the experience. Instead, we are simply delegating to an agent which doesn't need the Delta app or the Google Calendar web app. And so in some ways what he's built is a generative GUI crutch for this transitional period. That's just a solve for him to be able to give the agent, which doesn't need a UX, enough information to complete the task successfully. Still, my broader point is that there is a lot of exploration happening and not every use case that people get incredibly excited about is going to be one that you have to run out and go try for yourself right away. So what are some predictions for where this all leads? Well, one of them is in fact this post UI World Broadloom CEO Todd Saunders writes, Every single day I become more convinced that the next winners in vertical software won't have a ui. They will be API first, agent first products that integrate directly into a company. Slack teams, email or browser sales team doesn't want another dashboard, they want deals automatically qualified in their CRM. Your accountants don't need another portal, although they do love portals. They want invoices reconciled in the tools they already live in the UI list. Future is coming. My working thesis still tbd, is that we are moving from software you visit to software that visits you. Another prediction, which is really more of a sensibility or a feeling which you can see in some of these posts, is that although this is amazing, there is something being lost too as the job of the software engineer changes. In another viral post, Dukka writes, I'm not sure if other developers feel like this, but I feel kind of depressed like everyone else. I've been using Claude code for a while. It's not a recent thing and it's incredible. I've never found coding more fun. The stuff you can do and the speed you can do it at now is absolutely insane. And I'm using it to ship a lot and solve customer problems faster. So all around it's a win. But at the same time the skill I spent ten thousands of hours getting good at programming, the thing I spent most of my life getting good at is becoming a full commodity extremely quickly. As much fun as it is, and as much as I like using the tools, there's something disheartening about the thing you spent most of your life getting good at now being mostly useless, writes Pathak. Everyone is going to feel this soon. Another line of predictions as expressed by Lewis everyone who tried Claude Code over the holidays and has been thinking deeply about this for seven days. All white collar jobs are toast. It's over. Although, to be fair to Louis, he was being a little tongue in cheek and making fun of the Twitter hype hosts that love big pronouncements like all white collar jobs are toast. That's certainly part of the sensibility that is floating around out there right now, maybe more discreetly. Outside of all jobs being toast, there's certainly a sense that the way work happens and the way that business happens is going to look quite different. Damian Player writes Things I'm bearish on in 2026 because of Claude code 1. Massive teams Opus 4.5 can now do the work of five people for $200 a month. Headcount is a liability now. 2. Manual workflows if a human touches the same task twice, it should be automated. No exceptions. 3. Agency services not tied to ROI. No one's paying $3k a month for email when AI can write, test and optimize the entire funnel. The companies that figure this out first will run circles around everyone else. The ones that don't will die off. Another big theme in the predictions, which I completely agree with. We'll call mainstreaming. Riley Brown writes, you thought 2025 was the vibe coding boom. Three months from now all platforms, not just X will be talking about it. Lovable and replit have no moat. The space is wide open. And I think Riley is dead on when he notes that this is still so contained to the people who were already either a coding or b deeply involved in AI in some way. And it is not going to take long for the wide world of normies to realize that they have been given godlike creation powers unlike anything they've ever experienced before. Hamal Hussain gave an example. He tweeted Wife who doesn't code created this website for her educational podcast with Claude Code in 30 minutes. Claude Code guided her by deploying on Vercel and getting a domain downloaded and analyzed all her content for website copy, made a responsive version and resolved all links to each episode in RSS Feedback. And if you have kids, by the way, go check out Wonderwise Science. But it's even easier than the setup he just described. This is frankly 5 minutes in repliter lovable with the point being that although Claude Code might be out in the lead, there is an entire ecosystem which is going to drive a huge amount of this activity online alongside the mainstreaming. One of the areas where we're going to see the expansion of Vive coding this year is of course across the enterprise. This is one of my main predictions in my 50 predictions for 2026 that by coding was going to go everywhere into every department. Carvalotti wrote a hilarious post where he says, I showed Claude code to my company last month and every department now believes they have AGI attached to. It was a hilarious meme of a soldier handing a chimpanzee in AK47 with the title marketing people discovering Opus4.5 and Claude Code. Now, in addition to this just changing what people can do inside the organization, there is likely a broader organizational change coming in terms of operational expectations. Paul Graham reposted that tweet from the Google engineer who was talking about how Claude code recreated their entire work of the last year in an hour, with Graham writing, this illustrates an aspect of AI that I hadn't thought about till now. It cuts through bureaucracy. If a big organization is paralyzed by indecision, AI doesn't care. It will happily generate a version one and that becomes the starting point because there is no other version 1. Indeed, all of this infiltration into the enterprise is going to create new skill set needs. Ethan Malik writes, when you see how people use Claude Code, Codex, etc. It becomes clear that managing agents is really a management problem. Can you specify goals? Can you provide context? Can you divide up tasks? Can you give feedback? These are teachable skills. Also, UIs need to support management now. He points out that this is going to get an assist likely from ui when he writes, a big question for the coming year is whether the major AI labs can rethink the experience to be more suited towards delegation on other tasks. I think the answer will be if they don't, there will be other interfaces that do that now, beyond the enterprise world, there are also a lot of thoughts about what new possibilities open up with application development. Balaji Srinivasan writes about a golden age of local and decentralized apps. He writes, the reason we will have this golden age is that Claude code allows you to quickly clone any moderately complex cloud based app into a decent local one that runs on only your files. Apps of moderate complexity without strong global network effects are suddenly easy to clone. The clone won't be perfect right away, but it'll be pretty good and if the cloning dev sticks with it, it'll get better. This gets to that idea of personal software which is a big theme for our predictions next year. Likewise, the apps are going to look different. Dan Shipper talks a little bit more about what the idea of agent native apps really means. He writes that with traditional software architecture, you write code that defines what happens and the computer executes your instructions. With agent native architecture, you define outcomes in natural language. The agent figures out how to achieve them using tools. In an agent native worlds, features are prompts, not code. All of this, he says, enables emergent capability in your app. It can do things you didn't plan for. This, he says, allows you to discover latent demand from your users that inform your roadmap. Dan points out, by the way, that this is fairly similar to how Boris Cherny actually builds features into clawed code. And contrary to the prediction that Claude code and these new models are just going to destroy all jobs, there are many who are extremely excited about the entrepreneurial possibilities that these tools and models are going to unlock. Greg Eisenberg writes, I regret to inform you the amount of opportunity to build software startups and mobile apps with Claude Code, etc. Is mind bending at the moment. In another tweet, he writes, The $1.3 trillion app store doesn't die, it's going to explode. Claude code makes it so that anyone can ship an app, which means the App store fills up 10,000 times faster than before. Most are AI junk, but a few are surprisingly good and personal instead of universal. Apple will get stricter and add more filters. How else are they going to deal with the flood of apps and making it easier for people to find them? At the same time, the types of apps will change over the next 24 months. People will start building very specific apps for very specific situations. An example he gives is an app that tells me if my snoring last night was bad enough to wake my partner instead of a generic sleep tracker. Hearkening to my prediction that these new app entrepreneurs will have a different economic model, he writes, most of these apps never try to get huge in the traditional VC sense. Ultimately, software feels more indie, more personal. This era is way better for people who want to build apps, people who want more personal apps, but maybe worse off for Apple who have to deal with the slob. Crisping up his predictions, Greg finishes Claude code will probably make 50,000 people millionaires, if not way more. Be on Twitter, I think captures some of my feeling when he writes, it's cliche af, but you literally have the entire world at your fingertips right now. You could hostile takeover boomer businesses acquire billions of real estate. Clone a SaaS and take the entire market share. Run up a bag with affiliate marketing. The list goes on and on. You should sit down at your computer every morning with pure excitement as you could be 10 button pushes away from retirement. Pick anything, but for the love of God, pick something. McKay Wrigley echoes the sentiment. Claude Code in Opus 4.5, he writes, injected the immaculate hacker vibes back into AI that we haven't had since GPT4. Everything is new and fun and weird again. You can feel it. Another oom of new ideas and latent economic value is waiting to be unlocked. And building has never been this fun. So what's the conclusion? It is not totally clear exactly what set of changes the combination of Opus 4. 5 and Claude Code, or 5.2 Codex, or whatever set of latter day models and tools you want to point to have done to tip over into this new era. But the tip has happened. The people who are the most proficient users of AI and code all feel it, and the exhilaration is palpable. Now if you've been too intimidated to try these tools yet, first of all, I would say there's no shame at all in using something like Repl Dot or Lovable instead of the more intimidating Claude code to start. It's still going to give you a feel for just how much has changed in what you can do. But also Claud code is itself getting more accessible. Claud codepm Cat Wu just in fact tweeted earlier this week if you want to use Claude code but don't like the terminal interface, you can now use local Claude code from Claude Desktop. Download Claude Desktop, open the sidebar and click Code Toggle. Select the folder you want Claud Code to have access to and submit your prompt. I do think we are starting 2026 in a unique moment and I am very excited to see what you guys all go out and build. For now though, that is going to do it for today's AI Daily brief. Appreciate you listening as always and until next time, peace.

0:00