Anthropic confirms software engineering is NOT dead

Anthropic confirms software engineering is NOT dead 🔗 **Source:** [Anthropic confirms software engineering is NOT dead](https://www.youtube.com/watch?v=Te2I2muO-4c) > [!note]+ Document Notes > tags: [javascript, ai-tooling, acquisition, anthropic] > author: ThePrimeTime > source: YouTube > reference: https://www.youtube.com/watch?v=Te2I2muO-4c > date created: 2024-12-19 14:30 > > # 🏷️ Title: Anthropic confirms software engineering is NOT dead >  ## 🧰 Action Inbox - [H1] Analyze Anthropic's acquisition strategy — acquisition motivations/0:03:45 #business-strategy - [H2] Evaluate sustainability models for open-source projects — Bun vs Deno monetization/0:01:30 #open-source - [H3] Track AI tooling market consolidation — language runtime acquisitions/0:08:45 #market-analysis ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [Bun GitHub](https://github.com/oven-sh/bun) - [Anthropic Claude](https://www.anthropic.com) ## 📝 Summary - Anthropic acquired Bun, the JavaScript runtime, in a strategic move to enhance their Claude Code tooling - The acquisition contradicts claims that "software engineering is done" - Anthropic invested millions in engineering talent - Bun will remain open-source (MIT license) but gain resources for faster development - The move reflects AI companies' recognition that they need expert developers to build competitive tools - Investment actions speak louder than marketing claims about AI replacing programmers ## 📖 Main Content Anthropic's acquisition of Bun represents a significant strategic investment in JavaScript tooling infrastructure. Bun, created by Jared Sumner, has been a VC-backed company struggling to find sustainable revenue models, unlike competitors like Deno which monetized through Deno Deploy. The acquisition solves Bun's sustainability question while giving Anthropic control over a critical piece of infrastructure for their Claude Code CLI tool. The irony highlighted in the video is that despite widespread claims that AI will replace software engineers, Anthropic's actions demonstrate the opposite belief. Rather than using AI to fork and improve Bun themselves, they chose to acquire the team and expand it, investing millions in human expertise. This contradiction between public statements about AI capabilities and actual business decisions reveals the current limitations of AI in complex software engineering tasks. Bun will maintain its MIT license and open-source nature while benefiting from increased resources and hiring. The acquisition allows Anthropic to ensure their Claude Code tool has the best possible foundation for shipping self-contained executables across platforms. ## 🗒️ Notes #### ⏱️ [00:00:00](https://www.youtube.com/watch?v=Te2I2muO-4c?t=0) — [H1] Anthropic Acquires Bun JavaScript Runtime - Key idea: Anthropic purchased Bun to enhance their AI tooling capabilities - Evidence: "Bun has been acquired by Anthropic. Yes, I never thought I'd live to see the day a JavaScript engine being sold to an AI company" #### ⏱️ [00:01:15](https://www.youtube.com/watch?v=Te2I2muO-4c?t=75) — [H2] Bun's Sustainability Problem - Key idea: Bun lacked a clear monetization strategy despite VC backing - Evidence: "People would frequently ask, 'How is Bun sustainable?' If I bet my company's tech stack on Bun, will Bun still be around in a few years? We didn't have a great answer" #### ⏱️ [00:02:30](https://www.youtube.com/watch?v=Te2I2muO-4c?t=150) — [H3] Comparison with Deno's Business Model - Key idea: Deno had clear revenue through Deno Deploy, unlike Bun - Evidence: "Dino, kind of a competitor to that, has always had Dino Deploy and Dino Deploy, of course, they make money. You can see right here, transparent pricing, $20 for pro" #### ⏱️ [00:03:00](https://www.youtube.com/watch?v=Te2I2muO-4c?t=180) — [H2] What Won't Change Post-Acquisition - Key idea: Bun maintains open-source status and development focus - Evidence: "Bun is going to stay MIT and open source. Bun will continue to be extremely actively maintained" #### ⏱️ [00:03:45](https://www.youtube.com/watch?v=Te2I2muO-4c?t=225) — [H3] Anthropic's Strategic Motivation - Key idea: Control over Bun's roadmap ensures optimal CLI tool development - Evidence: "Bun allows Claude Code to be able to ship self-contained executables that can run on any platform easily" #### ⏱️ [00:04:30](https://www.youtube.com/watch?v=Te2I2muO-4c?t=270) — [H1] The "Software Engineering is Done" Contradiction - Key idea: Claims about AI replacing engineers contradict actual business investments - Evidence: "Software engineering is done. Now, I don't know what it is. I don't know why it is. It just for some odd reason, this irks me" #### ⏱️ [00:05:15](https://www.youtube.com/watch?v=Te2I2muO-4c?t=315) — [H2] Pattern of AI Doomsaying - Key idea: Every AI model release triggers premature claims about engineering obsolescence - Evidence: "This has been going on since Chad Jippity 35. When Chad Jippy 35 came out, software engineering is done. When 40 came out, software engineering is done" #### ⏱️ [00:06:00](https://www.youtube.com/watch?v=Te2I2muO-4c?t=360) — [H3] MIT License Reveals True Beliefs - Key idea: If AI could replace engineers, Anthropic would have forked Bun instead of buying it - Evidence: "What is the license of Bun? Well, Bun itself is MIT licensed. What does that mean? I can fork bun and have Prime Bun" #### ⏱️ [00:06:45](https://www.youtube.com/watch?v=Te2I2muO-4c?t=405) — [H2] Actions vs. Words Analysis - Key idea: Anthropic's investment contradicts claims about AI replacing engineers - Evidence: "Why didn't Claude just simply fork Bun and do all the changes itself? Well, I think Claude recognizes very well that they neither have the expertise nor the ability" #### ⏱️ [00:07:30](https://www.youtube.com/watch?v=Te2I2muO-4c?t=450) — [H3] Reality of Investment Decisions - Key idea: Companies invest in human expertise because they believe it's necessary - Evidence: "They are not giving in to the premise that software engineering is done in six months. In fact, they're doing the exact opposite" #### ⏱️ [00:08:15](https://www.youtube.com/watch?v=Te2I2muO-4c?t=495) — [H2] Impact on Developer Community - Key idea: Doomsday predictions cause unnecessary panic among developers - Evidence: "I genuinely get so many emails every week with people in an absolute panic because they think software engineering is gone in 3 months" #### ⏱️ [00:08:45](https://www.youtube.com/watch?v=Te2I2muO-4c?t=525) — [H3] Language Wars Theory - Key idea: Tech giants may be strategically positioning around programming languages - Evidence: "96.9% of it's written in Rust. And if you look at bun, almost 60% of it's written in Zigg. I think these billionaires are playing some language games" ## 📚 References - [Bun Official Website](https://bun.sh/) - [Anthropic Claude Code](https://www.anthropic.com/claude) - [Deno Deploy Pricing](https://deno.com/deploy/pricing) - [MIT License Details](https://opensource.org/licenses/MIT) ## 🗣️ Paraphrase - Anthropic bought Bun not because AI can replace engineers, but because they need expert developers to build competitive tooling - their investment proves they believe human expertise remains essential - The JavaScript runtime acquisition gives Anthropic control over critical infrastructure for their Claude Code CLI, enabling better cross-platform executable distribution - Despite repeated claims that each AI model advancement makes programming obsolete, companies continue investing heavily in engineering talent, revealing the gap between marketing hype and business reality ## 📰 Newsletter 🚀 **Big Tech Doubles Down on Developer Talent** Anthropic's acquisition of Bun sends a clear message: despite AI advances, expert developers remain invaluable. While headlines claim "software engineering is done," Anthropic invested millions to acquire and expand the Bun team. **Key Takeaway**: Watch what companies do, not what they say. Investment decisions reveal true beliefs about AI capabilities versus human expertise. **For Developers**: Your skills matter more than the hype suggests. Focus on building expertise in foundational technologies like runtimes, tooling, and infrastructure. Resources: [Learn WordPress](https://learn.wordpress.org/), [Developer Docs](https://developer.wordpress.org/), [WP-CLI](https://wp-cli.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create content addressing AI impact fears, showcase human+AI collaboration examples - Education: Develop workshops on AI-assisted development best practices - Developer experience: Evaluate tooling acquisitions for ecosystem impact - Developer marketing: Counter doomsday narratives with practical AI integration stories - Community management: Address developer anxiety about AI replacement in forums #### OSPO - Policy & Governance: Monitor open-source acquisitions by AI companies for licensing changes - License Compliance: Review MIT-licensed dependencies for acquisition risks - Security Management: Assess supply chain implications of AI company acquisitions - Software Inventory: Track runtime and tooling dependencies on acquired projects - Developer Support: Provide guidance on navigating AI tooling integration - Community Engagement: Support sustainable funding models for critical open-source infrastructure - Strategic Alignment: Align OSS strategy with AI tooling ecosystem evolution #### Community - Maintainers: Develop sustainable funding strategies beyond acquisition - Contributors: Create succession planning for key infrastructure projects - Sponsors: Increase funding for JavaScript runtime ecosystem sustainability - Ecosystem: Build redundancy in critical tooling infrastructure #### WordPress - Core: Evaluate Bun adoption for WordPress development tooling - Core Performance: Monitor JavaScript runtime performance implications - Community: Address developer concerns about AI impact on WordPress careers - Training: Develop AI-assisted WordPress development curriculum - Documentation: Update tooling guides to include modern JavaScript runtimes - CLI: Consider Bun integration for WP-CLI performance improvements - Meta: Assess .org infrastructure dependencies on acquired tools ## 🧪 Hypotheses - AI companies will continue acquiring developer tooling companies rather than replacing engineering teams - Open-source sustainability crises will drive more acquisitions by well-funded AI companies - Developer productivity tools will become strategic assets for AI companies competing on code generation quality - The gap between AI marketing claims and actual engineering capabilities will persist for several years ### 📊 Methodology - Track acquisition patterns: Monitor AI company purchases of developer tooling vs. claims about AI replacing engineers - Survey developer sentiment: Measure anxiety levels about AI replacement before and after major acquisitions - Performance comparison: Benchmark human vs. AI capabilities on complex software engineering tasks like runtime optimization ### ❓ Questions - Which other critical JavaScript ecosystem tools might become acquisition targets? - How will open-source sustainability models evolve in response to AI company interest? - What does Anthropic's roadmap reveal about the future intersection of AI and developer tooling? ### 🔮 Future Research - Analysis of open-source project sustainability in the AI era - Comparative study of AI company acquisition strategies in developer tooling - Long-term impact assessment of tech giant control over programming language ecosystems ### 💡 Implications - Practice: Developers should focus on building expertise in foundational technologies that AI companies are acquiring rather than replacing - Policy: Open-source communities need sustainable funding models that don't depend on acquisition as the primary exit strategy - Security: Concentration of critical tooling under AI companies creates new supply chain risks - Education: Developer education should emphasize AI collaboration skills rather than AI replacement fears - Funding: Investment in developer tooling infrastructure becomes strategic for AI competitiveness ### 🧠 Learning Styles Assessment Reading level estimate: 8th-9th grade #### 👀 Visual learners - Present: GitHub contribution graphs, acquisition timeline visualization - Improve: Flowchart showing acquisition rationale, comparison charts of runtime features #### 🔊 Aural learners - Present: Video format with clear narration and commentary - Improve: Podcast version with structured segments, audio summary of key points #### 🗣️ Verbal learners - Present: Clear analogies (MIT license explanation), technical term definitions - Improve: More analogies for business strategy concepts, glossary of acquisition terms #### 🧩 Kinesthetic learners - Present: Practical examples of forking vs. acquiring decisions - Improve: Hands-on tutorial for evaluating open-source sustainability, interactive timeline #### 🧠 Logical learners - Present: Clear cause-and-effect reasoning about acquisition motivations - Improve: Decision tree for acquisition vs. build-vs-buy analysis, financial impact modeling #### 🧑‍🤝‍🧑 Social learners - Present: Community discussion prompts, developer impact stories - Improve: Forum for discussing similar acquisitions, peer review of sustainability strategies #### 🧘‍♀️ Solitary learners - Present: Individual reflection on career implications - Improve: Self-assessment worksheet for AI impact, personal skill development planning #### 🌿 Naturalistic learners - Present: Ecosystem thinking about JavaScript runtime environment - Improve: Sustainability frameworks, lifecycle analysis of open-source projects ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading structure, descriptive timestamps - Improve: Alt text for visual elements, more descriptive link text than "right here" #### Keyboard navigation - Present: YouTube player supports keyboard navigation - Improve: Timestamp links should be clearly labeled for screen readers #### Contrast and readability - Present: Good text contrast in video captions - Improve: Ensure transcript formatting meets contrast requirements #### Text resizing and reflow - Present: YouTube interface scales appropriately - Improve: Provide transcript in responsive format #### Media and captions - Present: Auto-generated captions available - Improve: Professional captions with speaker identification, downloadable transcript #### Forms and interactions - Present: Standard YouTube interaction patterns - Improve: Custom interactive elements should follow accessibility patterns #### Standards reference - Video content should include audio descriptions for visual elements per WCAG 2.2 AA ### 🌍 DEIB Assessment #### Cultural breadth - Present: Focus on global tech industry implications - Improve: Include perspectives from international open-source maintainers, non-English language ecosystem impacts #### Inclusive language - Present: Generally uses accessible technical language - Improve: Define more technical terms, avoid dismissive language about concerns #### Neurodiversity support - Present: Clear structure with timestamps and sections - Improve: Provide content warnings for animated elements, consistent pacing #### Representation - Present: Focuses on technical merit and business analysis - Improve: Include diverse voices in open-source sustainability discussion #### Psychological safety - Present: Addresses legitimate developer concerns about career impact - Improve: More empathetic framing of developer anxiety, constructive rather than dismissive tone ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Sound logic connecting acquisition behavior to true beliefs about AI capabilities - Inductive strength --- ## Full Transcript [00:00] Bun has been acquired by Anthropic. Yes, [00:03] I never thought I'd live to see the day [00:05] a JavaScript engine being sold to an AI [00:08] company, but hey, here we are. Now, I [00:10] think this actually has a lot of [00:12] implications, and I think one of the [00:14] biggest implications, which I'm going to [00:16] like yapagedon about here shortly, has [00:18] to do with this tweet right here. I [00:20] believe this new model in Claude Code is [00:23] a glimpse of the future we're hurdling [00:24] towards. Maybe as soon as the first half [00:27] of next year, software engineering is [00:30] done. Yes, this acquisition of Bun makes [00:33] a huge statement about this exact tweet [00:36] from somebody who works on Claude Code. [00:38] But first, we got to take care of [00:40] business. Okay, we need to know why did [00:42] Anthropic buy Bun? Why was Bun looking [00:44] to be sold? And most importantly, why [00:47] you should just hit the subscribe [00:48] button? Because guys, I'm almost to a [00:51] million. When I cross a million, I never [00:52] have to ask this question again. So, [00:54] just press like, press subscribe, get it [00:56] done with so we could ALL JUST MOVE ON. [00:58] You see this black? This is silver. I [01:00] don't want silver. I want gold. All [01:01] right. So, first off, why did Bun sell? [01:04] Well, it's actually pretty obvious right [01:05] here. Jared is actually so shockingly [01:06] open about this. People would frequently [01:08] ask, "How is Bun sustainable?" If I bet [01:11] my company's tech stack on Bun, will Bun [01:13] still be around in a few years? We [01:15] didn't have a great answer to this [01:16] question until today. This is honestly, [01:18] this is shockingly honest. This is not [01:20] something that I think you often see on [01:22] Twitter. So, hey, I'd really like to say [01:24] great job Jared for just awesome [01:26] communication there. Bun practically [01:28] since its inception has been a VCbacked [01:31] company. They raised $7 million for [01:33] their seed round and well, they've been [01:35] in search of a way to make money and [01:37] they've never really had a strong [01:39] answer. Now, Dino, kind of a competitor [01:41] to that, has always had Dino Deploy and [01:44] Dino Deploy, of course, they make money. [01:46] You can see right here, transparent [01:48] pricing, $20 for pro or custom for [01:50] enterprise. And so everybody just [01:52] figured Bun would go down the same road [01:54] and well they really haven't. And so a [01:57] lot of people had to ask well you're a [01:59] VC back to open source project. How are [02:02] you going to make money? They didn't. [02:03] They just were bought. And just to kind [02:05] of make sure everybody understands [02:06] things that won't change. Bun is going [02:08] to stay MIT and open source. Bun will [02:11] continue to be extremely actively [02:14] maintained. And of course if you ever [02:16] looked at the creator of Bun's commit [02:18] graph, it does look very extreme. Which [02:20] by the way, do you see this little dot [02:21] right here? On the 28th, there was no [02:23] contributions on the 28th. Other people [02:26] asked what was happening right there. [02:27] And then it turns out the acquisition of [02:30] Bun 4hour walk probably happened on that [02:32] day. The same team still works on Bun. [02:34] Now, this is going to lead to me having [02:36] a crash out here shortly. Don't worry, [02:38] we'll we'll get to it. Bun is still [02:40] public on GitHub and Bun's roadmap will [02:42] continue to focus on high performance [02:44] JavaScript tooling, Noode.js [02:45] compatibility, and replacing Nood.js JS [02:47] is the default serverside runtime for [02:49] JavaScript. So, Bun's continuing to [02:50] claim it's going to operate pretty much [02:52] as usual. The only big difference [02:54] they're going to make is this right [02:56] here. They're going to be hiring more [02:58] engineers to work on Bun to deliver [03:00] features faster because well, they got [03:02] that backing from Anthropic now. But why [03:04] would Anthropic want to purchase Bun? [03:07] Well, I think a very simple answer is [03:09] this. Bun allows Claude Code to be able [03:12] to ship self-contained executables that [03:14] can run on any platform easily. So, I [03:16] know as much as it pains me to say this, [03:19] TypeScript is not a bad option to be [03:21] able to run these tools, except for of [03:23] course on like say RAM. RAM it just just [03:25] it just loves it. Just eats all of it. [03:27] And so, the motivation to me seems at [03:29] least fairly obvious. Anthropic needs to [03:32] be able to control the road map of Bun [03:35] so that they can make sure that the tool [03:37] they are building will be hopefully the [03:39] best of all the CLI tools. And that's [03:41] really because Open Code is an [03:44] incredible tool. It's growing like [03:46] absolutely wild right here and it allows [03:50] you to have any of your models you want. [03:52] You're not trapped on one. It is by far [03:55] a very compelling choice and it's also [03:57] built with bun. So, Anthropic obviously [03:59] wanted to make sure they have every [04:01] single edge they can possibly get in [04:03] developing a good CLI tool. But this is [04:06] where the crash out is going to happen. [04:08] Okay, cuz I did kind of allude to this [04:10] earlier. Software engineering is done. [04:13] Now, I don't know what it is. I don't [04:16] know why it is. It just for some odd [04:18] reason, this irks me. It's like I I feel [04:21] just this like great upsetness boiling [04:23] inside of me every single time I see [04:26] this. Now, this has been going on since [04:28] Chad Jippity 35. When Chad Jippy 35 came [04:31] out, software engineering is done. When [04:34] 40 came out, software engineering is [04:36] done. When Claude 35 came out, it was [04:38] done again. Grock, it was done. Chat, [04:41] Jippy 5, we're cooked. Everything's [04:43] over. 5.1, nobody really cared. Claude [04:46] 37, yeah, no one really liked that one. [04:48] Claude 4.5, oh yeah, again, it's so [04:52] over. Gemini 3, it's so over. But you [04:55] can tell that they don't believe what [04:57] they say. And here's the simple [04:59] observation why. What is the license of [05:01] Bun? Well, Bun itself is MIT licensed. [05:04] What does that mean? What does what does [05:06] the president of MIT have to do with [05:08] software? Well, an MIT license simply [05:12] means you can do whatever you want with [05:14] the code. I can fork bun and have Prime [05:17] Bun. That's it. It can become mine. And [05:20] there's really nothing that Jared or [05:22] anyone from the bun team can do. Now, [05:24] the reality is no one's going to want to [05:25] use Prime Bun because everybody's [05:27] working on Bun. Now, let's just give the [05:30] premise that software engineering is [05:32] done. Why would Anthropic spend money to [05:34] buy Bun? Why would Anthropics spend [05:37] money to aqua hire the entire bun team? [05:41] I thought software engineering was done. [05:43] I thought within the next 3 to 6 months, [05:45] again, by the way, that we were going to [05:47] have no more software engineering. Why [05:50] didn't Claude just simply fork Bun and [05:54] do all the changes itself? Well, I think [05:55] Claude recognizes very well that they [05:58] neither have the expertise nor the [06:00] ability to make Bun significantly better [06:03] by themselves. And so the best thing [06:05] they can do is hire the people who run [06:08] and make bun. They are not giving in to [06:10] the premise that software engineering is [06:11] done in six months. In fact, they're [06:13] doing the exact opposite. They are [06:15] hiring the experts. They are expanding [06:18] the team and saying, "Go as fast as you [06:21] can and make this product better and [06:23] better." Even though we've been told, [06:25] "Hey, this is all over." Here they are [06:28] not living up to the words they say, but [06:30] instead living up to reality. Because no [06:32] matter how much you want things to be [06:34] true, reality always asserts itself in [06:37] the end. And simply believing you can [06:39] make software engineers go away because [06:41] you got a statistical word calculator [06:44] does not mean it's true. Sure, there's [06:46] going to be a lot of improvements in our [06:47] life and software engineering is going [06:49] to continue to keep changing over the [06:51] next decade. And I would be a fool to [06:54] tell you where software engineer is [06:56] going to be in 10 years, let alone even [06:58] in 6 months. I have no idea. I'm not out [07:01] there attempting to be the Oracle of [07:04] Deli to tell you exactly what your [07:06] future's going to be. I can tell you one [07:08] thing. Anthropic believes it's good to [07:10] have hard skills. You know how I know? [07:12] Cuz they paid millions of dollars to get [07:15] those hard skills. Stop listening to the [07:17] news presses. Stop listening to Daario [07:20] saying, "Hey, guess what, boys? It's all [07:22] over. Everybody's cooked." And start [07:23] looking at where they spend their [07:25] dollars because that's the real story. [07:28] That's what they really believe. Of [07:30] course, Anthropic aren't the only one. [07:32] Gary, of course, which by the way, YC [07:34] has some sort of stake in Replet, [07:36] Emergent Labs, and Tascade is out here [07:39] saying that you can just over a weekend [07:41] replace some sort of SAS service, and [07:44] the SAS service is $30 per seat. I just [07:46] want you to think about that. We're all [07:48] software engineers. We've all done this. [07:50] You're paying a little bit of money to [07:51] some service, and then you don't like [07:53] it. So, what do you do? You spend [07:55] hundreds of hours replacing it, which is [07:57] a little ridiculous. your time valuing [07:59] yourself at like $1 per hour. And so [08:02] when I see these things, it's just like [08:04] this constant wave of, hey, guess what? [08:06] It doesn't matter. Everything you build [08:08] doesn't matter. What he's really saying [08:09] is every single business he's ever [08:11] invested in is also rubbish because you [08:15] over a weekend can simply just vibe code [08:18] whatever. Oh, you're using Replet. Why [08:20] would I use Replet? I'm just going to [08:22] vibe code Replet really quickly. Oh, why [08:25] would I even use cursor? I'm just going [08:26] to use cursor to make my own cursor and [08:28] make a bajillion dollars. Like this [08:30] premise is just fundamentally dumb. [08:33] Unless if you have a toddler's [08:35] understanding of software engineering. [08:38] Okay, I'm sorry for the crash. Okay, I'm [08:40] sorry. I'm sorry. It's just I feel [08:41] strongly about this because I I just I [08:44] genuinely get so many emails every week [08:47] with people in an absolute panic because [08:50] they think software engineering is gone [08:52] in 3 months and what have I done? I've [08:54] just spent the last year trying to get [08:55] good at something and now it's gone. And [08:57] I feel like the people spreading this [08:59] stuff, they're actively invested against [09:02] the words they even say. So yeah, I'm a [09:04] little bit triggered. But you know what? [09:06] I think I actually know the real reason [09:07] why Daario acquired Bun. Okay, this is [09:11] this is kind of like cutting edge stuff. [09:13] So you know, you're probably not going [09:14] to see a lot of this in the news. But if [09:15] you go over and you look at open codes [09:18] tool codeex, you'll see right here that [09:21] 96.9% [09:23] of it's written in Rust. And if you look [09:24] at bun, almost 60% of it's written in [09:28] Zigg. I think these billionaires are [09:29] playing some language games. Okay, [09:31] OpenAI clearly cornering the market on [09:33] Rust. Daario just bought a major [09:36] position in Zigg. Okay, these [09:37] billionaires, they're playing games we [09:39] don't even understand with the languages [09:41] we love. I even heard that Windsurf is [09:43] in talks with Phil C. The name is the [09:47] primogen. The two biggest problems in [09:49] computer science is cash invalidation [09:51] and pronouncing things. I personally [09:53] spent all day on teams arguing whether [09:55] it's SQL or Squeal. Well, I have great [09:57] news for you today. We have a company [09:59] that solves both of these problems. [10:01] Convex. Uh I'm pretty sure that's [10:03] Convex. [10:04] >> Teach out of my ad. Convex lets you ship [10:08] your backend without a backend team. Its [10:10] real-time database makes it easy to keep [10:12] all of your queries up to date. Convex [10:14] allows you to stop worrying about [10:16] complicated SQL and untyped messes. [10:19] Instead, you can embrace Convex's deep [10:21] TypeScript integration with type [10:22] completion and validation and database [10:24] queries that are just pure uncut [10:27] TypeScript functions. Try Convex today [10:29] in your language of choice, TypeScript, [10:31] Python, and even Rust. Visit convex.dev.  ## 📺 Video ![](https://www.youtube.com/watch?v=Te2I2muO-4c) --- # 📄 Full Transcript *Auto-generated from YouTube captions · timestamps roughly every 30 seconds · [watch on YouTube](https://www.youtube.com/watch?v=Te2I2muO-4c)* **[00:00]** Bun has been acquired by Anthropic. Yes, I never thought I'd live to see the day a JavaScript engine being sold to an AI company, but hey, here we are. Now, I think this actually has a lot of implications, and I think one of the biggest implications, which I'm going to like yap again about here shortly, has to do with this tweet right here. I believe this new model in Claude code is a glimpse of the future we're hurtling towards. Maybe as soon as the first half of next year, software engineering is **[00:30]** done. Yes, this acquisition of Bun makes a huge statement about this exact tweet from somebody who works on Claude code. But first, we got to take care of business, okay? We need to know, why did Anthropic buy Bun? Why was Bun looking to be sold? And most importantly, why you should just hit the subscribe button, because guys, I'm almost to a million. When I cross a million, I never have to ask this question again. So, just press like, press subscribe, get it done with, so we could ALL JUST MOVE ON. YOU see this plaque? This is silver. I don't want **[01:00]** silver. I want gold. All right, so first off, why did Bun sell? Well, it's actually pretty obvious right here. Jared is actually so shockingly open about this. People would frequently ask, how is Bun sustainable? If I bet my company's tech stack on Bun, will Bun still be around in a few years? We didn't have a great answer to this question until today. This is honestly this is shockingly honest. This is not something that I think you often see on Twitter. So, hey, I'd really like to say great job, Jared, for just awesome communication there. Bun, practically since its inception, has been a VC-backed company. They raised $7 **[01:32]** million for their seed round, and well, they've been in search of a way to make money, and they've never really had a strong answer. Now, Deno, kind of a competitor to that, has always had Deno Deploy. And Deno Deploy, of course, they make money. You can see right here, transparent pricing, $20 for pro, or custom for enterprise. And so, everybody just figured Bun would go down the same road, and well, they really haven't. And so, a lot of people had to ask, well, you're a VC-backed open source project. How are you going to make money? They **[02:03]** didn't. They just were bought. And just to kind of make sure everybody understands, things that won't change, Bun is going to stay MIT and open source. Bun will continue to be extremely actively maintained. And of course, if you ever looked at the creator of Bun's commit graph, it does look very extreme. Which by the way, do you see this little dot right here? On the 28th, there was no contributions on the 28th. Other people asked what was happening right there. And then it turns out, the acquisition of Bun, four-hour walk, probably happened on that day. The **[02:33]** same team still works on Bun. Now, this is going to lead to me having a crash out here shortly. Don't worry, we'll get we'll get to it. Bun is still public on GitHub, and Bun's road map will continue to focus on high-performance JavaScript tooling, Node.js compatibility, and replacing Node.js as the default server-side runtime for JavaScript. So, Bun's continuing to claim it's going to operate pretty much as usual. The only big difference they're going to make is this right here. They're going to be hiring more engineers to work on Bun to deliver features faster. Because well, they got that backing from Anthropic **[03:03]** now. But why would Anthropic want to purchase Bun? Well, I think a very simple answer is this. Bun allows Claude code to be able to ship self-contained executables that can run on any platform easily. So, I know, as much as it pains me to say this, TypeScript is not a bad option to be able to run these tools. Except for of course on like say RAM. RAM it just it just it just loves it. It just eats all of it. And so, the motivation to me seems at least fairly obvious. Anthropic needs to be able to control the road map of Bun, so that **[03:35]** they can make sure that the tool they are building will be hopefully the best of all the CLI tools. And that's really because open code is an incredible tool. It's growing like absolutely wild right here, and it allows you to have any of your models you want. You're not trapped on one. It is by far a very compelling choice, and it's also built with Bun. So, Anthropic obviously wanted to make sure they have every single edge they can possibly get in developing a good CLI tool. But this is where the crash **[04:07]** out is going to happen, okay? Cuz I did kind of allude to this earlier. Software engineering is done. Now, I don't know what it is. I don't know why it is. It just for some odd reason this irks me. It's like I I feel just this like great upsetness boiling inside of me every single time I see this. Now, this has been going on since ChatGPT 3.5. When ChatGPT 3.5 came out, software engineering is done. When Furro came out, software engineering is done. When Claude 3.5 came out, it was done again. **[04:39]** Grok, it was done. ChatGPT 5, we're cooked. Everything's over. 5.1, nobody really cared. Claude 3.7, yeah, no one really liked that one. Claude 4.5, oh yeah, again, it's so over. Gemini 3, it's so over. But you can tell that they don't believe what they say. And here's the simple observation why. What is the license of Bun? Well, Bun itself is MIT licensed. What does that mean? What does What does the president of MIT have to do with software? Well, an MIT license **[05:11]** simply means you can do whatever you want with the code. I can fork Bun and have Prime Bun. That's it. It can become mine. And there's really nothing that Jared or anyone from the Bun team can do. Now, the reality is no one's going to want to use Prime Bun, because everybody's working on Bun. Now, let's just give the premise that software engineering is done. Why would Anthropic spend money to buy Bun? Why would Anthropic spend money to acquire the entire Bun team? I thought software **[05:41]** engineering was done. I thought within the next 3 to 6 months, again, by the way, that we were going to have no more software engineering. Why didn't Claude just simply fork Bun and do all the changes itself? Well, I think Claude recognizes very well that they neither have the expertise nor the ability to make Bun significantly better by themselves. And so, the best thing they can do is hire the people who run and make Bun. They are not giving in to the premise that software engineering is done in 6 months. In fact, they're doing **[06:13]** the exact opposite. They are hiring the experts. They are expanding the team and saying, go as fast as you can and make this product better and better. Even though we've been told, hey, this is all over, here they are not living up to the words they say, but instead living up to reality. Because no matter how much you want things to be true, reality always asserts itself in the end. And simply believing you can make software engineers go away because you got a statistical word calculator does not **[06:44]** mean it's true. Sure, there's going to be a lot of improvements in our life, and software engineering is going to continue to keep changing over the next decade. And I would be a fool to tell you where software engineer is going to be in a 10 years, let alone even in 6 months. I have no idea. I'm not out there attempting to be the Oracle of Delphi to tell you exactly what your future is going to be. I can tell you one thing, Anthropic believes it's good to have hard skills. You know how I know? Cuz they paid millions of dollars to get those hard skills. Stop listening **[07:16]** to the news presses. Stop listening to Dario saying, hey, guess what, boys, it's all over. Everybody's cooked. And start looking at where they spend their dollars. Because that's the real story. That's what they really believe. Of course, Anthropic aren't the only one. Gary, of course, which by the way, YC has some sort of stake in Replit, Emergent Labs, and Taskade, is out here saying that you can just over a weekend replace some sort of SaaS service, and the SaaS service is $30 per seat. I just want you to think about that. We're all **[07:48]** software engineers. We've all done this. You're paying a little bit of money to some service, and then you don't like it. So, what do you do? You spend hundreds of hours replacing it, which is a little ridiculous. You're time valuing yourself at like $1 per hour. And so, when I see these things, it's just like this constant wave of, hey, guess what, it doesn't matter. Everything you build doesn't matter. What he's really saying is every single business he's ever invested in is also rubbish. Because you, over a weekend, can simply just vibe code whatever. Oh, you're using **[08:20]** Replit? Why would I use Replit? I'm just going to vibe code Replit really quickly. Oh, why would I even use Cursor? I'm just going to use Cursor to make my own Cursor and make a bajillion dollars. Like this premise is just fundamentally dumb. Unless if you have a toddler's understanding of software engineering. Okay, I'm sorry for the crash. Okay, I'm sorry. I'm sorry. It's just I feel strongly about this, because I I just I genuinely get so many emails every week with people in an absolute panic, because they think software engineering **[08:51]** is gone in 3 months, and what have I done? I've just spent the last year trying to get good at something, and now it's gone. And I feel like the people spreading this stuff, they're actively invested against the words they even say. So, yeah, I'm a little bit triggered. But you know what? I think I actually know the real reason why Dario acquired Bun, okay? This is This is kind of like cutting-edge stuff. So, you know, you're probably not going to see a lot of this in the news. But if you go over and you look at Open Code's CLI tool Codex, you'll see right here that 96.9% **[09:23]** of it's written in Rust. And if you look at Bun, almost 60% of it's written in Zig. I think these billionaires are playing some language games. Okay, OpenAI clearly cornering the market on Rust. Dario just bought a major position in Zig. Okay, these billionaires, they're playing games we don't even understand with the languages we love. with Phil C. The name is the primogen. The two biggest problems in computer science is cache invalidation and pronouncing things. I personally spent all day on Teams **[09:54]** arguing whether it's SQL or squeal. Well, I have great news for you. Today, we have a company that solves both of these problems, Convex. Uh Prime, pretty sure that's Convex. Pete, get out of my ad. Convex lets you ship your back end without a back end team. It's real-time database makes it easy to keep all of your queries up to date. Convex allows you to stop worrying about complicated SQL and untyped messes. Instead, you can embrace Convex's deep TypeScript integration with type completion and validation and database queries that are **[10:25]** just pure uncut TypeScript functions. Try Convex today in your language of choice, TypeScript, Python, and even Rust. Visit convex.dev.