Price:

AI & TECH

Anthropic chief claims AI has solved coding

Thursday, June 25, 2026 · from 4 podcasts, 5 episodes
This Week in StartupsLenny's PodcastPodcasting 2.0Peter McCormack Show
  • Anthropic’s code-gen head says AI agents now write 100% of his work.
  • Solo founders build complex software in days, erasing the need for junior developer teams.
  • Veteran engineers warn this speed sacrifices the creative friction that produces quality work.

Boris Cherny hasn't edited a line of code by hand since November. As head of Anthropic’s Claude Code, he argues the act of programming is "virtually solved."

Speaking on Lenny's Podcast, Cherny detailed a new reality where he ships 10 to 30 pull requests daily by managing a team of AI agents. This isn't a projection. Anthropic reports a 200% jump in productivity per engineer, and Claude Code already authors 4% of all public GitHub commits. The job itself is changing.

"The title 'software engineer' will fade. I think it will get replaced by 'builder'."

- Boris Cherny, Lenny's Podcast

The economic pressure for this shift is immense. On This Week in Startups, discussion highlighted how the app Interval reached a million downloads with just three developers - a task that would have required a dozen engineers and millions in funding a decade ago. Peter McCormack, on his podcast, gave a sharper example: he personally built a complex management system for his football club in 11 days. An agency had quoted the project at £1 million and a 15-person team.

This hyper-productivity lets a single founder operate with the leverage of a mid-sized company. Fernando Nikolić told The Peter McCormack Show he runs his data business with a 94% profit margin and zero employees, relying instead on an "arsenal of agents."

But this efficiency has a cost. On the Podcasting 2.0 show, developer Dave Jones described the change as a kind of grief. He went from writing all of his project's code to just 10% in two months. Jones argues the struggle - the friction of building - is what ensures quality and meaning. By removing it, he warns, you just get slop.

Nikolić argues the only defensible businesses left are those with physical moats. If your work is based on information, like knowing a programming language, AI is closing that gap. But if you have to build hypersonic jets like Hermeus or physical game pieces like Board, that work is safe from being automated away in an afternoon.

The line between the person who defines a product and the person who builds it is dissolving. For the junior developer, whose job was to translate well-defined specs into code, there may not be a role left.

#Coding#Startups#Big Tech#Labor
Boris ChernyPeter McCormackAnthropicClaudeClaude CodePodcasting 2.0

Source Intelligence

- Deep dive into what was said in the episodes

This Week in Startups
This Week in Startups

Why the Future of Video Games is Moving Back to the Dinner Table • Jun 24

  • Bren Putnam's company, Board, created the first face-to-face gaming console, a 24-inch touchscreen that recognizes physical game pieces without internal electronics or sensors.
  • Board's proprietary technology uses a unique conductive pattern in game pieces and a custom software stack with embedded AI to detect, track, and interpret real-time physical interactions on the screen.
  • Developing Board took approximately two years, expedited by Bren Putnam's prior experience with Mirror and her CTO Ryan's background in machine learning and computer vision from self-driving cars.
  • Board manufactured its console with a tier-one contract manufacturer in Thailand, leveraging their experience with display products to prototype more quickly compared to Bren Putnam's previous hardware venture, Mirror, which was produced in Mexico.
  • Bren Putnam advises founders to validate the product experience with early, 'hacky' prototypes for investors rather than over-engineering, emphasizing that fundraising is effective storytelling about the customer experience.
  • The Board console sells for $399, including seven initial games. Individual games with custom pieces cost around $35, with a creator tools subscription launching later this year to drive long-term value.
  • Board launched direct-to-consumer for the holiday season last year, selling out 10,000 launch units within weeks, indicating strong market demand for its new category of family-centric gaming.
  • AJ Piplica's company, Hermeus, is developing Darkhorse, an uncrewed hypersonic reusable aircraft designed to provide asymmetric military advantage by operating faster than the SR-71 with greater impunity.
  • Hermeus's Chimera engine is a turbine-based combined cycle system that transitions from a turbojet (Mach 3, using a pre-cooler to manage 800°F air intake) to a ramjet for hypersonic speeds by using shockwaves for compression.
  • The Quarterhorse program iteratively de-risks hypersonic flight, building roughly one new aircraft per year. Its Mark 2.1 achieved Mach 1.21 in its first supersonic flight, 364 days after Mark 1 flew.
  • Hermeus recently raised a $350 million Series C, demonstrating venture capital's ability to fund ambitious aerospace projects, with a strategy to secure billions more in revenue from customer-funded R&D.
  • Hermeus monetizes its technology de-risking through a Defense Innovation Unit contract, mirroring SpaceX's model of customer-funded R&D to accelerate development and capital efficiency.
  • AJ Piplica contrasts Hermeus's rapid development with traditional defense primes, noting that F-22 and F-35 fighter jets took 20-25 years and $30-60 billion each for development, highlighting a structural advantage for agile startups.
#Startups#AI Infrastructure#Robotics#Macro#Energy

The hottest running app has nothing to do with speed | E2303 • Jun 22

  • Louie Phillips says Interval is a gamified running app where users run around the block to claim territory on a live global map.
  • Louie Phillips states Interval intentionally excludes speed metrics, focusing instead on volume-based competition to allow anyone, including a slow-moving neighbor, to capture territory.
  • Louie Phillips describes future features like arenas for speed-based competition with daily leaderboards and volume-based leaderboards for regular territory capture.
  • Louie Phillips says Interval has grown to about a million downloads and about 100,000 followers on Instagram through organic social media content without paid media.
  • Louie Phillips says Interval's paid ad funnel on Meta provides predictable growth, with a cost per trial start of about $12 and an average customer lifetime of about 17 months.
  • Jason Calacanis notes Blackstone purchased Hamilton Island for 1.2 billion Australian dollars, roughly $804 million US.
  • Alice Zhang says Verge Labs uses a multimodal transformer-based world model, representing each patient as a 512-dimensional vector, to fuse data and infer missing information like brain activity from blood alone.
  • Alice Zhang details Verge Labs' revenue models: custom partnerships, direct insights licensing, and platform-as-a-service formats.
  • Alice Zhang cites a partnership with Eli Lilly where 83% of AI-derived targets validated in wet lab experiments, far exceeding Lilly's 20% expectation.
  • Alice Zhang argues drug development costs $5 billion on average primarily due to nine out of ten failures at the expensive last stage, implying AI-driven efficiency could dramatically lower prices.
  • Alice Zhang claims Verge Labs' moat is its decade-long infrastructure for collecting brain tissue data, a difficult task others avoid, not just model scaling.
  • Jason Calacanis describes Agree's average time from contract sent to signature as under 7 hours and median time from invoice sent to paid as 36 hours.
Also from this episode: (2)

Science (2)

  • Alice Zhang says Verge Labs built one of the field's largest brain datasets directly from patients, comprising over 12,000 human brains from 6,000 patients.
  • Alice Zhang states Verge Labs learned a critical lesson from its own drug development experience: predicting patient response is more valuable than just discovering a drug.
#Models#Health#Startups#Macro#Biology
Lenny's Podcast
Lenny's Podcast

What happens after coding is solved? | Fiona Fung (Manager of the Claude Code and Cowork Teams) • Jun 21

  • Boris Cherny states that 100% of his code is written by Claude Code, with no manual edits since November 2024. He ships 10-30 pull requests daily and often runs five agents simultaneously.
  • Cherny argues coding is virtually solved for his work, and he predicts the title 'software engineer' will fade, replaced by 'builder'. He believes everyone will soon be a product manager who codes.
  • A SemiAnalysis report found Claude Code authors 4% of all GitHub commits, a figure Cherny notes is higher for private repos. The report predicts Claude Code will author one-fifth of all commits by the end of 2025.
  • Cherny says productivity per engineer at Anthropic has increased 200% since introducing Claude Code, measured by pull request volume. He contrasts this with his time at Meta, where annual productivity gains were only a few percentage points.
  • Claude Code's growth is accelerating, with daily active users doubling in the past month. Cherny built the initial prototype, called Quad CLI, as a terminal tool because the model improved too quickly for other form factors.
  • Cherny advocates underfunding projects initially to force teams to 'Claudify' and automate work with AI. He advises leaders to give engineers unlimited tokens for experimentation, then optimize costs only after an idea proves successful.
  • He sees the printing press as the best historical analog for AI's impact, enabling a transition from a specialized skill to a universal capability. He imagines a future where anyone can build software.
  • For using Claude Code, Cherny recommends always using the most capable model, starting tasks in 'plan mode', and experimenting with different interfaces beyond the terminal.
  • The Claude Co-work agent was built in 10 days using Claude Code. It emerged from observing latent demand, as users were hacking the coding tool for non-technical tasks like analyzing genomes or recovering photos.
  • Cherny describes a three-layer safety approach at Anthropic: mechanistic interpretability to study model neurons, laboratory evals, and studying model behavior in the wild through early product releases.
  • He observes that newer engineers often use Claude Code in more advanced, 'AGI-forward' ways than veterans, who can get stuck in old mental models. He cites an example where a junior engineer used Claude to debug a memory leak faster than he could manually.
Also from this episode: (1)

AI Infrastructure (1)

  • Cherny advises builders to design for the AI model six months from now, not its current capabilities. He warns against over-engineering workflows, arguing the more general model will always outperform a specialized, scaffolded one.
#Agents#Coding#Startups#Enterprise
Podcasting 2.0
Podcasting 2.0

Adam Curry

Episode 264: Podcast Plebicide • Jun 19

  • Dave mourns the abrupt shift from writing 100% of his code to writing only 10%, which happened within two months due to AI coding agents.
  • Adam argues that AI reduces build friction to near zero, which floods the world with trivial projects rather than meaningful contributions.
  • Adam describes how AI-generated art and music on No Agenda initially displaced human artists, but skilled creators later learned to use the tools effectively.
  • Dave explains Godcaster sends only one play event per user per hour, deduplicated via a ULID, to avoid issues with caching and unreliable connection data.
Also from this episode: (7)

Social Media (2)

  • Dave theorizes that isolated, private communication platforms like Slack distort attention, making single issues feel like the entire world and fueling conflict.
  • Dave argues federated platforms like Mastodon better mimic real-life conversation by allowing semi-public discussions where others can overhear and join.

Media (3)

  • Adam and Dave credit the Podcasting 2.0 project's success to its combination of Mastodon, GitHub, weekly podcast meetings, and live chat.
  • OP3 data shows Podcasting 2.0 had about 5,247 unique listeners in May, placing it in the so-called 'indie middle class' of 5k-25k monthly downloads.
  • Adam rejects the term 'indie podcaster', arguing all podcasters are independent by definition, and dismisses obsession with download metrics.

AI Infrastructure (2)

  • Dave clarifies James's proposal: full support for the transcript tag means supporting only the VTT format; anything else does not count.
  • Podping's trust system uses a plebiscite model where trusted nodes vote to add or remove other nodes from the trusted list, ensuring swarm integrity.
#Coding#Agents#Psychology#Social Media#Macro
The Peter McCormack Show
The Peter McCormack Show

Peter McCormack

#185 - Fernando Nikolić - AI, One-Person Companies & The New Economic Elite • Jun 19

  • Fernando Nikolić runs a one-person company using AI agents, achieving a 94% profit margin with an infrastructure cost of roughly $200-$250 per month on Google AI Ultra.
  • Peter McCormack rebuilt a comprehensive club operations system for his football team alone in two weeks. His previous agency estimated the same project would require 15 people, 18 months, and a £1 million budget.
  • McCormack built an automated media company infrastructure for his podcast, complete with simulated departments, weekly AI agent meetings, and a grading system for hundreds of weekly recommendations, in 11 days.
  • McCormack argues front-end UI and many SaaS tools are being commoditized to zero value, as custom software can be 'vibe-coded' with AI in days, rendering tools like Squarespace, Canva, and Grammarly obsolete for proficient users.
  • Both hosts observe AI is terrible at creative, subjective tasks like writing jokes, developing brand strategy, or generating original marketing content, which Nikolić calls 'slop' that is polluting the web.
  • Nikolić warns of a 'permanent underclass' if AI's exponential productivity gains are unevenly distributed to a wealthy elite, questioning whether those left behind will even be able to afford the AI utility bills.
  • McCormack built a system tracking 628 UK MPs, ingesting 10,915 tweets to grade their truthfulness with 81% accuracy, costing roughly £3,000 to process, as a prototype for AI-driven political accountability.
  • McCormack sees a parallel between the current 'chaos gap' of AI disruption and past upheavals like file-sharing's destruction of the music industry, driven by the collapse of information asymmetries.
  • Nikolić's company, Perception, focuses on 'narrative engineering' by aggregating and sanitizing expensive, tedious data from the digital asset space, which he argues remains a moat versus easily automated software.
  • McCormack estimates he'll spend at least $10,000 on AI inference tokens this month, primarily on Claude, but considers it 'peanuts' compared to the multi-million pound development costs it replaces.
Also from this episode: (2)

Protocol (1)

  • Nikolić argues Bitcoin has failed to achieve hyperbitcoinization and is not for everyone, requiring a radical 'Truman Show' moment of economic realization that most people are emotionally unwilling to undergo.

Society (1)

  • Both hosts note a craving for analog, high-trust experiences - like physical media, vinyl, and localized interaction - as a reaction to the overwhelming, sanitized digital landscape and AI-homogenized content.
#Agents#Coding#Startups#Macro#Media

Related Stories