Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Anthropic CPO, Mike Krieger: "over 70% of Anthropic pull requests are now generated by AI" but we're still figuring out what that means for code review and long-term architecture.

slow_developer's profile picture

Haider.

48,762 subscribers

97,999 görüntüleme • 1 yıl önce •via X (Twitter)

Bilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

11 Yorum

Juniper profil fotoğrafı
Juniper1 yıl önce

They don’t ship a lot

MightyBot profil fotoğrafı
MightyBot1 yıl önce

🧠 Unified Search. Smarter Meetings. Effortless CRM. MightyBot is your AI agent platform for seamless workflows—record meetings, automate CRM updates, and find answers across apps in seconds. 🌟 Focus on what matters. We'll handle the grind.

Adrian Ziegler profil fotoğrafı
Adrian Ziegler1 yıl önce

70% AI-generated today, 100% AI-refactored tomorrow.

Tech Brief Ai profil fotoğrafı
Tech Brief Ai1 yıl önce

Over 70% of pull requests at Anthropic are now AI-generated—wild stat from Mike Krieger. We're in uncharted territory. Speed is up, but the big question now: how do we ensure long-term code quality and architecture hold up? AI is changing how we build—but not why we review. (We cover shifts like this daily—check the newsletter in bio.)

Gabriel Cezar profil fotoğrafı
Gabriel Cezar1 yıl önce

PR with what exactly? A div? A CSS color? An internal survey?

Necati Demir, Phd profil fotoğrafı
Necati Demir, Phd1 yıl önce

And, the rest 30% of the PRs produce 70% of the total value ;)

John P Alioto profil fotoğrafı
John P Alioto1 yıl önce

explains a lot actually

David Rosa profil fotoğrafı
David Rosa1 yıl önce

The rest 30% is fixing that 70%

Product Guru’s profil fotoğrafı
Product Guru’s1 yıl önce

That’s why they don’t update..

Drawk Kwast profil fotoğrafı
Drawk Kwast1 yıl önce

Does anyone reading this (including the person who made this post) understand the difference between a “pull” and a “push” request???

Andrew profil fotoğrafı
Andrew1 yıl önce

interesting...

Benzer Videolar

Anthropic CPO Mike Krieger says that Claude is now effectively writing itself Engineers regularly ship 2–3,000-line pull requests generated entirely by Claude Dario predicted a year ago that 90% of code would be written by AI, and people thought it was crazy "today it's effectively 100%"
0:37

Anthropic CPO Mike Krieger says that Claude is now effectively writing itself Engineers regularly ship 2–3,000-line pull requests generated entirely by Claude Dario predicted a year ago that 90% of code would be written by AI, and people thought it was crazy "today it's effectively 100%"

Haider.

569,128 görüntüleme • 6 ay önce

.Anthropic CPO and Instagram co-founder Mike Krieger (Mike Krieger) on what comes next Don't miss this one. We discuss: 🔸 How Claude Code now writes 90-95% (!!) of Claude Code 🔸 What new bottlenecks appear once so much of your code is written by AI 🔸 How Anthropic plans to compete with OpenAI long-term 🔸 How to use Claude as your product strategy partner (with specific prompting techniques) 🔸 Why Mike shut down Artifact, and what other founders can learn from it 🔸 Where AI startups should focus to avoid getting squashed by OpenAI, Anthropic, and Google 🔸 Why MCP might reshape how all software works 🔸 The counterintuitive metrics that matter for AI 🔸 Much more Listen now 👇 • YouTube: • Spotify: • Apple: Thank you to our wonderful sponsors for supporting the podcast: 🏆 Stripe — Helping companies of all sizes grow revenue: 🏆 Productboard — Make products that matter: 🏆 OneSchema — Import CSV data 10x faster:
1:15

.Anthropic CPO and Instagram co-founder Mike Krieger (Mike Krieger) on what comes next Don't miss this one. We discuss: 🔸 How Claude Code now writes 90-95% (!!) of Claude Code 🔸 What new bottlenecks appear once so much of your code is written by AI 🔸 How Anthropic plans to compete with OpenAI long-term 🔸 How to use Claude as your product strategy partner (with specific prompting techniques) 🔸 Why Mike shut down Artifact, and what other founders can learn from it 🔸 Where AI startups should focus to avoid getting squashed by OpenAI, Anthropic, and Google 🔸 Why MCP might reshape how all software works 🔸 The counterintuitive metrics that matter for AI 🔸 Much more Listen now 👇 • YouTube: • Spotify: • Apple: Thank you to our wonderful sponsors for supporting the podcast: 🏆 Stripe — Helping companies of all sizes grow revenue: 🏆 Productboard — Make products that matter: 🏆 OneSchema — Import CSV data 10x faster:

Lenny Rachitsky

293,184 görüntüleme • 1 yıl önce

Anthropic CPO, Mike Krieger: Dario Amodei predicted the coding benchmark (SWE-bench) would reach 90% by the end of the year I’ve started taking AI timelines more seriously after seeing the progress. "mid-2025 now feels much closer than 2027"
1:15

Anthropic CPO, Mike Krieger: Dario Amodei predicted the coding benchmark (SWE-bench) would reach 90% by the end of the year I’ve started taking AI timelines more seriously after seeing the progress. "mid-2025 now feels much closer than 2027"

Haider.

46,067 görüntüleme • 1 yıl önce

Anthropic CPO, Mike Krieger: Within 1-3 years, Claude will act as an autonomous "coworker" Instead of waiting for instructions, it will stay in your business loop, monitor data, propose changes, and even write code with human giving final approval
0:56

Anthropic CPO, Mike Krieger: Within 1-3 years, Claude will act as an autonomous "coworker" Instead of waiting for instructions, it will stay in your business loop, monitor data, propose changes, and even write code with human giving final approval

Haider.

28,382 görüntüleme • 10 ay önce

BREAKING 🚨: Upcoming API Playground preview on Anthropic Workbench and a new analytics page for Claude Code. What % of your code is AI-generated? "coming soon" 👀
0:37

BREAKING 🚨: Upcoming API Playground preview on Anthropic Workbench and a new analytics page for Claude Code. What % of your code is AI-generated? "coming soon" 👀

TestingCatalog News 🗞

14,242 görüntüleme • 1 yıl önce

Anthropic code review this, clanker review that ... why don't you shut up and review+annotate your own code.... (yes im a loser who still manually reviews code) Originally inspired by a bunch of feature requests and then seeing Dillon Mulroy tweet a similar cool ux. plannotator for reviewing plans (primary focus) and code, fully oss. OpenCode, Mario Zechner 's and Claude Code and other clankers
0:38

Anthropic code review this, clanker review that ... why don't you shut up and review+annotate your own code.... (yes im a loser who still manually reviews code) Originally inspired by a bunch of feature requests and then seeing Dillon Mulroy tweet a similar cool ux. plannotator for reviewing plans (primary focus) and code, fully oss. OpenCode, Mario Zechner 's and Claude Code and other clankers

Michael Ramos

69,589 görüntüleme • 5 ay önce

“Real-world impact has come much faster with AI than it did with mobile.” -Mike Krieger AI is already changing budgets, workflows, and P&Ls. I sat with Mike Krieger, CPO at Anthropic. Here’s what he told me: • How AI is driving revenue today • Sonnet 4.5: Opus-level brains at 1/5 the cost • Agents that can run 30 hours on tough coding tasks • How Claude Code lands in the enterprise Watch our full conversation below 👇
58:29

“Real-world impact has come much faster with AI than it did with mobile.” -Mike Krieger AI is already changing budgets, workflows, and P&Ls. I sat with Mike Krieger, CPO at Anthropic. Here’s what he told me: • How AI is driving revenue today • Sonnet 4.5: Opus-level brains at 1/5 the cost • Agents that can run 30 hours on tough coding tasks • How Claude Code lands in the enterprise Watch our full conversation below 👇

Zain Kahn

739,802 görüntüleme • 10 ay önce

Today, we're launching @PolarBrowser for Windows. The AI browser that beats Anthropic and OpenAI on every single major web benchmark. Now it's on the other 70% of desktops. Try for free now:
0:22

Today, we're launching @PolarBrowser for Windows. The AI browser that beats Anthropic and OpenAI on every single major web benchmark. Now it's on the other 70% of desktops. Try for free now:

Kevin Jiang

393,153 görüntüleme • 13 gün önce

"I don't prompt Claude anymore. I have loops running that prompt Claude and figuring out what to do. My job is to write loops. And this is transition we're going to see for the rest of the year." - Boris Cherny, head of Claude Code at Anthropic.
0:40

"I don't prompt Claude anymore. I have loops running that prompt Claude and figuring out what to do. My job is to write loops. And this is transition we're going to see for the rest of the year." - Boris Cherny, head of Claude Code at Anthropic.

Rohan Paul

2,302,968 görüntüleme • 2 ay önce

Insane: 90-95% of Claude Code is now written by Claude Code. I asked Mike Krieger (Anthropic's CPO) what changes when so much of your code is written by AI? His answer: The bottleneck moves from coding/design to: 1. Upstream: aligning the team around what to build. 2. Downstream: the PR merge queue, reviewing and merging all of the new code into production. Anthropic is where other product teams will be in a year or two, so it's important we pay attention. "I would expect that a year from now, the way that we are conceiving of building and shipping software changes a lot. Because it's gonna be very painful to do it the current way."
0:36

Insane: 90-95% of Claude Code is now written by Claude Code. I asked Mike Krieger (Anthropic's CPO) what changes when so much of your code is written by AI? His answer: The bottleneck moves from coding/design to: 1. Upstream: aligning the team around what to build. 2. Downstream: the PR merge queue, reviewing and merging all of the new code into production. Anthropic is where other product teams will be in a year or two, so it's important we pay attention. "I would expect that a year from now, the way that we are conceiving of building and shipping software changes a lot. Because it's gonna be very painful to do it the current way."

Lenny Rachitsky

369,920 görüntüleme • 1 yıl önce

LeetCode is dead. Developers don't write code line-by-line anymore. They orchestrate AI agents working in parallel, review AI-generated code, and make architectural decisions. That's the job now. But most interview processes haven't caught up. They still test algorithm memorization instead of AI fluency, code review, and judgment. We're building assessments for next-gen hiring that mirror how developers actually work. Here's how we think about it:
2:43

LeetCode is dead. Developers don't write code line-by-line anymore. They orchestrate AI agents working in parallel, review AI-generated code, and make architectural decisions. That's the job now. But most interview processes haven't caught up. They still test algorithm memorization instead of AI fluency, code review, and judgment. We're building assessments for next-gen hiring that mirror how developers actually work. Here's how we think about it:

HackerRank

343,870 görüntüleme • 4 ay önce

Anthropic CEO: “In the next 3 to 6 months, AI will write 90% of the code, and within 12 months, nearly all code may be generated by AI.” the job isn’t coding anymore, it’s telling machines what to build.
0:32

Anthropic CEO: “In the next 3 to 6 months, AI will write 90% of the code, and within 12 months, nearly all code may be generated by AI.” the job isn’t coding anymore, it’s telling machines what to build.

CG

892,220 görüntüleme • 4 ay önce

📁 Boris Cherny, engineer at Anthropic, says 100% of his code is written by Claude. He ships 10 to 30 pull requests a day and has not edited a single line by hand in months. Claude writes it. Claude reviews it. Humans are still in the loop, but they are no longer the ones typing.
1:08

📁 Boris Cherny, engineer at Anthropic, says 100% of his code is written by Claude. He ships 10 to 30 pull requests a day and has not edited a single line by hand in months. Claude writes it. Claude reviews it. Humans are still in the loop, but they are no longer the ones typing.

Jon Hernandez

22,274 görüntüleme • 5 ay önce

Mike Krieger (Mike Krieger) is the CPO of Anthropic ($10B+ raised) and the co-founder of Instagram, which he sold to Meta for $1B. Here's the full video of my recent conversation with him. Mike has one of the AI industry's most interesting jobs. He shared with me how he and his team craft product strategy for the company that's leading the charge on AI in the enterprise. Specifically, we discussed how frontier model innovations both drive product and vice versa (how product ideas inform AI research). We also talk about the long term defensibility of models (inspired by the emergence of DeepSeek), and how Mike believes that not only will individual models have specific strengths over others (such as in areas like coding, science, etc), but that a model's "vibes" will also be a major factor for driving customers' choice. Mike also shared his view on how AI will reinvent media and the business model of advertising on the internet, drawn heavily from his work building one of the most successful ad products ever built (Instagram) and his work on Artifact, an AI news product he also co-founded. Lastly, Mike dove deep into what it's like building for the Enterprise for the first time in his career, and how lessons from Instagram and Meta inform not only product development, but how Anthropic thinks about scaling its team in this period of hypergrowth. Chapters: 00:00 Introduction 00:54 Mike Krieger's Journey to Anthropic 03:17 Building Product Strategy at Anthropic 07:43 Rapid Iteration and Safety 10:58 Differentiating AI Models and User Experience 17:57 Impact of AI on Consumer Products and Business Models 24:39 Enterprise vs. Consumer Product Strategy 29:19 AI in Personal Life Management 30:15 Open Source and Claude Integrations 33:09 AI-Assisted Product Development 37:13 Scaling Teams and Processes at Anthropic 42:17 Reflections on AI and Future Prospects
54:20

Mike Krieger (Mike Krieger) is the CPO of Anthropic ($10B+ raised) and the co-founder of Instagram, which he sold to Meta for $1B. Here's the full video of my recent conversation with him. Mike has one of the AI industry's most interesting jobs. He shared with me how he and his team craft product strategy for the company that's leading the charge on AI in the enterprise. Specifically, we discussed how frontier model innovations both drive product and vice versa (how product ideas inform AI research). We also talk about the long term defensibility of models (inspired by the emergence of DeepSeek), and how Mike believes that not only will individual models have specific strengths over others (such as in areas like coding, science, etc), but that a model's "vibes" will also be a major factor for driving customers' choice. Mike also shared his view on how AI will reinvent media and the business model of advertising on the internet, drawn heavily from his work building one of the most successful ad products ever built (Instagram) and his work on Artifact, an AI news product he also co-founded. Lastly, Mike dove deep into what it's like building for the Enterprise for the first time in his career, and how lessons from Instagram and Meta inform not only product development, but how Anthropic thinks about scaling its team in this period of hypergrowth. Chapters: 00:00 Introduction 00:54 Mike Krieger's Journey to Anthropic 03:17 Building Product Strategy at Anthropic 07:43 Rapid Iteration and Safety 10:58 Differentiating AI Models and User Experience 17:57 Impact of AI on Consumer Products and Business Models 24:39 Enterprise vs. Consumer Product Strategy 29:19 AI in Personal Life Management 30:15 Open Source and Claude Integrations 33:09 AI-Assisted Product Development 37:13 Scaling Teams and Processes at Anthropic 42:17 Reflections on AI and Future Prospects

Michael Mignano

85,991 görüntüleme • 1 yıl önce

""70, 80, 90% of the code written in Anthropic is written by Claude. (...) "I said something like this 3 or 6 months ago, people think of it as falsified because they think of it as like we're going to fire 70 80 or 90% of the software engineers but what really happens is that the 10% were still writing, you know, humans become managers of AI systems."
0:53

""70, 80, 90% of the code written in Anthropic is written by Claude. (...) "I said something like this 3 or 6 months ago, people think of it as falsified because they think of it as like we're going to fire 70 80 or 90% of the software engineers but what really happens is that the 10% were still writing, you know, humans become managers of AI systems."

Chubby♨️

295,805 görüntüleme • 10 ay önce

Creator and head of Claude Code, Anthropic: "I don't prompt Claude anymore. I have loops running that prompt Claude and decide what to do. My job is to write loops." He hasn't hand-written a line of code since November. 100% is Claude Code now. Still ships 10 to 30 pull requests a day. He breaks down his actual workflow in the 1h28m podcast. Loops writing 100% of the code. Engineer leveling up, or automated out?
30:22

Creator and head of Claude Code, Anthropic: "I don't prompt Claude anymore. I have loops running that prompt Claude and decide what to do. My job is to write loops." He hasn't hand-written a line of code since November. 100% is Claude Code now. Still ships 10 to 30 pull requests a day. He breaks down his actual workflow in the 1h28m podcast. Loops writing 100% of the code. Engineer leveling up, or automated out?

shmidt

521,927 görüntüleme • 1 ay önce

Dario Amodei says Claude now writes 70-90% of code at Anthropic Humans haven't been replaced; they've shifted roles. Engineers are becoming managers of AI systems "the future feels sci-fi… but when it arrives, it's oddly ordinary"
0:49

Dario Amodei says Claude now writes 70-90% of code at Anthropic Humans haven't been replaced; they've shifted roles. Engineers are becoming managers of AI systems "the future feels sci-fi… but when it arrives, it's oddly ordinary"

Haider.

96,068 görüntüleme • 10 ay önce

Anthropic CEO Dario Amodei just revealed the hidden bottleneck that will kill most AI companies in the next 18 months (Save this). The insight comes from a principle in computer science called Amdahl's Law. Dario's argument is simple when something starts working really well inside an organization, you have to immediately ask what isn't working well around it. Amdahl's Law states that the maximum speedup of any system is capped by the fraction you haven't improved and that applies to companies just as brutally as it applies to processors. If you can suddenly write three or four times as many pull requests as before, you don't get three or four times the output but you rather get a pile of code no one can review, verify, or trust. The data makes this impossible to ignore. Teams with heavy AI coding adoption are merging 98% more pull requests but PR review time has ballooned 91%, deployment velocity is effectively flat and 96% of developers don't fully trust AI-generated code reaching production. AI generated code produces 1.7x more issues per pull request than human written code, 0.83 issues per PR versus 6.45. Veracode's 2026 State of Software Security report found that 82% of organizations now carry security debt, up 11% year over year, with critical security debt surging 36% in a single year driven directly by AI-generated code reaching production faster than security teams can handle. What Dario is describing is a systems problem, not a software problem and coding is roughly 20% of the software delivery cycle. Even at infinite coding speed, you're still bottlenecked by review, security, verification, testing, and deployment which make up the other 80%. The enterprises that win are the ones that identify which part of their system is the new constraint after AI accelerates the old one and fix that next. This is why Anthropic's Claude Code focuses on the full development loop, not just generation, and why the verification and security layer of the AI stack is where the next wave of enterprise value gets created. This is also why Anthropic as a company is positioned differently than most people realize. Anthropic's 2026 Agentic Coding Trends Report found that organizations using full-loop agentic coding workflows where AI handles not just generation but testing, review, and deployment validation reduced their software defect rates by 43% while increasing velocity by 2.8x. Claude Code now authors 4% of all GitHub commits and is on track to hit 20%+ by year-end, with the full-loop use case growing 3x faster than pure code generation. Dario has been building Anthropic around the exact insight he's describing publicly ,the constraint isn't writing code but rather everything that has to happen after.
0:48

Anthropic CEO Dario Amodei just revealed the hidden bottleneck that will kill most AI companies in the next 18 months (Save this). The insight comes from a principle in computer science called Amdahl's Law. Dario's argument is simple when something starts working really well inside an organization, you have to immediately ask what isn't working well around it. Amdahl's Law states that the maximum speedup of any system is capped by the fraction you haven't improved and that applies to companies just as brutally as it applies to processors. If you can suddenly write three or four times as many pull requests as before, you don't get three or four times the output but you rather get a pile of code no one can review, verify, or trust. The data makes this impossible to ignore. Teams with heavy AI coding adoption are merging 98% more pull requests but PR review time has ballooned 91%, deployment velocity is effectively flat and 96% of developers don't fully trust AI-generated code reaching production. AI generated code produces 1.7x more issues per pull request than human written code, 0.83 issues per PR versus 6.45. Veracode's 2026 State of Software Security report found that 82% of organizations now carry security debt, up 11% year over year, with critical security debt surging 36% in a single year driven directly by AI-generated code reaching production faster than security teams can handle. What Dario is describing is a systems problem, not a software problem and coding is roughly 20% of the software delivery cycle. Even at infinite coding speed, you're still bottlenecked by review, security, verification, testing, and deployment which make up the other 80%. The enterprises that win are the ones that identify which part of their system is the new constraint after AI accelerates the old one and fix that next. This is why Anthropic's Claude Code focuses on the full development loop, not just generation, and why the verification and security layer of the AI stack is where the next wave of enterprise value gets created. This is also why Anthropic as a company is positioned differently than most people realize. Anthropic's 2026 Agentic Coding Trends Report found that organizations using full-loop agentic coding workflows where AI handles not just generation but testing, review, and deployment validation reduced their software defect rates by 43% while increasing velocity by 2.8x. Claude Code now authors 4% of all GitHub commits and is on track to hit 20%+ by year-end, with the full-loop use case growing 3x faster than pure code generation. Dario has been building Anthropic around the exact insight he's describing publicly ,the constraint isn't writing code but rather everything that has to happen after.

Milk Road AI

52,190 görüntüleme • 3 ay önce

Despite myths that there's "incredible demand for AI," 70%+ of hyperscaler AI revenues come from OpenAI and Anthropic. In FY26, $24.1bn of Microsoft's $34.33bn in AI revenue came from OpenAI - meaning that GPU rentals and AI software are single-digit billion dollar businesses.
1:32

Despite myths that there's "incredible demand for AI," 70%+ of hyperscaler AI revenues come from OpenAI and Anthropic. In FY26, $24.1bn of Microsoft's $34.33bn in AI revenue came from OpenAI - meaning that GPU rentals and AI software are single-digit billion dollar businesses.

Ed Zitron

99,098 görüntüleme • 1 gün önce