Price:

AI & TECH

Coding agents commoditize junior devs as AI revenue pivots from subscriptions

Saturday, May 23, 2026 · from 4 podcasts, 6 episodes
All-InThis Week in AIAI Daily BriefThis Week in Startups
  • AI coding agents now match frontier model performance at a tenth of the cost, directly targeting entry-level developer and QA roles.
  • Anthropic's revenue doubled every six weeks to a $44B annualized run rate, marking a seismic shift from SaaS subscriptions to token-based agent consumption.
  • Recent graduates are booing AI commencement speakers, seeing the technology as a betrayal of their professional prospects.

A quiet revolution in software development is wiping out the economic rationale for junior developer and QA roles. Cursor's Composer 2.5, a coding-specific model, scores within the same performance tier as frontier giants like GPT and Claude, but at 50 cents per million input tokens - roughly one-tenth the cost. This commoditization is no longer hypothetical. Firms like Pulsia are running fully agentic operations, hitting $6 million in annualized revenue with a single founder and no employees.

Nathaniel Whittemore argues this marks the 'second moment' for AI, moving from chatbots to autonomous builders that execute work. The business model has flipped. Where software was once sold by the 'seat,' it's now consumed by the token. In the agentic era, a single user can generate hundreds or thousands of dollars in monthly token sales. This is obliterating legacy SaaS valuations and creating a new economic hierarchy.

'Unless you own the tokens, you are a cost center.'

- Jason Calacanis, This Week in Startups

Anthropic is the poster child for this new calculus. Its annual recurring revenue (ARR) surged from $9 billion to over $44 billion this year, a pace that dwarfs historical growth at AWS or Salesforce. The company is adding $96 million in ARR per day. This explosive growth is funded by a massive infrastructure buildout, with SpaceX renting its Colossus supercomputer clusters to Anthropic for $1.25 billion per month.

The human impact is curdling into a generational rift. On This Week in Startups, Jason Calacanis argued recent graduates don't just fear AI; they feel betrayed by the tech leaders who built it. This sentiment turned physical during graduation season, where speakers mentioning AI were booed. A viral New York Times essay by a Stanford senior claimed AI has dissolved the foundations of liberal arts education faster than the workforce.

'They spent years earning degrees only to see the very people who built the tools they use predicting the evaporation of their job market.'

- Jason Calacanis, This Week in Startups

Yet macro data complicates the doomer narrative. Labor market figures show software engineering job postings are up 18% from last year, hitting their highest levels since late 2023. Stripe Atlas incorporations grew 130% year-over-year in Q1, suggesting an explosion of entrepreneurship as the cost of intelligence collapses. The real story may be one of violent displacement, not net loss.

Enterprise adoption is maturing, but its focus is shifting. Survey data cited by Whittemore shows a sharp decline in users citing 'time savings' as AI's primary value, dropping from 20% to 13% in a month. Companies are now prioritizing increased throughput and new capabilities, like Generative Engine Optimization, a field projected to grow from under $1 billion to $34 billion by 2034. The era of using AI to save minutes is over; the era of using it to invent new business models has begun.

#Agents#Coding#Enterprise
Nathaniel WhittemoreAmazonAnthropicNew York TimesSpaceXClaudeATLASCursorGPT

Source Intelligence

- Deep dive into what was said in the episodes

All-In with Chamath, Jason, Sacks & Friedberg
All-In with Chamath, Jason, Sacks & Friedberg

SpaceX's $2T Case, Nvidia's Shock Selloff, America Turns on AI, Trump Pulls AI Order, Bond Crisis? • May 22

  • Andre Karpathy is 39 years old, coined 'vibe coding,' built Auto Research (82k GitHub stars), and leads a new recursive self-improvement team at Anthropic.
  • Jason criticizes CEOs like Matthew Prince and Mark Zuckerberg for dystopian messaging around AI-driven layoffs, creating fear that employees are training their replacements.
  • Gavin Baker estimates the LLM market could reach $200-400 billion ARR by year-end, excluding large tech companies' internal ROI from improved recommender and ad systems.
  • Baker notes Nvidia's GPU financing advantage: domain-specific accelerators allow older GPUs to have a 10-15 year useful life, enabling asset-backed loans at rates as low as 6-7%.
Also from this episode: (11)

AI & Tech (6)

  • Friedberg argues AI is entering a utility-focused phase, citing AI solving decades-old math problems and generating viable drug candidates entering clinical trials.
  • Friedberg asserts AI backlash stems from perceived power imbalance, foreign state intervention campaigns, and the technology's anti-humanist psychological impact.
  • Friedberg sees AI proliferation as inevitable, akin to the nuclear arms race; slowing US development risks creating an asymmetric power imbalance with China.
  • Friedberg suggests AI regulation should focus on KYC for frontier models and post-harm legal recourse, not preemptive government power which becomes a one-way ratchet.
  • SpaceX's AI compute business grew revenue to $3.2B (doubled YoY) but had $6.4B operating losses; Anthropic pays $1.25B/month ($15B/year) for Colossus 1 & 2 compute.
  • Chamath says SpaceX's value lies in its terrestrial data center build speed (Colossus 2 in 91 days), AI compute business scaling, and Elon Musk's 'one more thing' civilizational creativity premium.

Startups (1)

  • SpaceX filed for IPO aiming to raise $75B at a $1.75T valuation; Starlink generated $11.4B revenue (50% growth) and $4.4B operating income with 10M subscribers.

Space (1)

  • Friedberg argues space-based data centers and Starlink create a backup for civilizational progress, offering an internet alternative not controlled or destroyed by terrestrial governments.

Chips (2)

  • Nvidia reported Q1 revenue of $81.6B (85% YoY growth), $58B net income, $48B free cash flow at 75% gross margins, and authorized an $80B buyback.
  • Gavin Baker says the AI semiconductor market is cross-sectionally inefficient, with memory makers at 3-5x PE, Nvidia low, and power/cooling/optical names discounting different futures.

Macro (1)

  • Baker argues America is relatively advantaged by Strait of Hormuz closure due to energy self-sufficiency, AI leadership, and being the 'best house in a bad neighborhood' of global debt.
#AI Infrastructure#Markets#Big Tech#Startups
This Week in AI
This Week in AI

Grads boo AI, Reese Witherspoon gets dunked + Karpathy joins Anthropic | TWiAI E14 • May 20

  • Andre Karpathy’s move to Anthropic is more about communication than research, according to Jason Calacanis. He argues Dario Amodei’s grim predictions make him a poor AI spokesperson, while Karpathy’s credibility can alleviate industry pressure.
  • Anthropic’s API pricing penalizes third-party providers. They offer a 20x token savings plan only for customers using Anthropic’s first-party products, a subtle anti-competitive move aimed at locking users into their ecosystem.
  • Fundamental builds tabular models for enterprise structured data, a modality poorly handled by LLMs. They have a confidential compute partnership with AWS, allowing models to be deployed and encrypted within a customer’s own VPC.
  • Linear is shifting from project management to AI execution. The product now includes an agent that can research feedback, write proposals, examine codebases, and delegate tasks, with a native coding agent in development.
  • Kanjun Q sees AI enabling bespoke, personalized user interfaces. She built her own agent UI for email and task management, stating that design principles shift when creating for a single user versus a mass audience.
  • Jeremy Frankle defines poor AI etiquette as shifting the burden of reviewing AI-generated slop onto coworkers. He asserts all AI output is the user’s responsibility and must be reviewed before delegation.
  • Jeremy Frankle calls graduating students hypocrites for booing AI while using ChatGPT for essays. He argues this is the best time to graduate, as AI is a powerful tool for creative expression and starting companies.
Also from this episode: (8)

AI & Tech (4)

  • Imbue co-founder Kanjun Q bought a 10,000 H100 GPU cluster in 2022 as an investment to fund the company, which now generates substantial rental revenue. She avoided venture capital, taking investment from corporate arms and a non-profit.
  • Karri Saarinen argues AI-generated design is often soulless and can worsen product quality. Founders who delegate design to AI without understanding the problem produce aesthetically pleasing but non-functional outputs.
  • The best use of AI is as a reflective surface to ask better questions, not just a solution generator. Imbue open-sourced ‘Blueprint,’ an agent skill tuned to ask high-quality questions to gather user context.
  • AI industry leaders are forming distinct cultural cults. Jason Calacanis categorizes them: SpaceX for tech libertarian monks, Anthropic for the left-leaning and earnest, and OpenAI for cutthroat capitalists.

Startups (1)

  • Kanjun Q warns frontier AI labs will vertically integrate into profitable application layers. The defense for startups is building headless products with orchestration layers that can easily swap underlying models.

AI Infrastructure (1)

  • A cost-effective local AI cluster can be built by daisy-chaining multiple Apple Mac Studios with high RAM. ExoLabs provides software to address multiple units as a single cluster.

Society (1)

  • Graduates are booing AI commencement speeches due to real fear and disempowerment. They perceive a future where entry-level jobs are automated and wealth creation excludes them, reacting against condescending advice.

Media (1)

  • A New York Times editorial attacked Reese Witherspoon for encouraging AI adoption. Kanjun Q argues this conflates two separate issues: using a helpful tool versus critiquing systemic power concentration.
#Agents#Startups#Coding
The AI Daily Brief: Artificial Intelligence News and Analysis
The AI Daily Brief: Artificial Intelligence News and Analysis

Nathaniel Whittemore

9 Codex Tips From the Codex Team • May 19

  • Cursor's Composer 2.5 coding model matches frontier model performance at a tenth of the cost, scoring 69.3% on Terminal Bench 2.0 and 79.8% on SweBench multilingual.
  • Cursor prices Composer 2.5 at 50 cents per million input tokens and $250 per million output tokens, half the cost of Opus 4.7 or GPT-5.5.
  • Chamath Palihapitiya argues enterprises using OpenAI or Anthropic directly are letting competitors into their data, creating an opening for model-agnostic harness-first companies.
  • Jason Lou's first Codex tip advocates using durable, long-running threads for key work streams, relying on OpenAI's improved context compaction to maintain persistent memory.
  • Lou argues voice interaction with Codex unlocks richer context by allowing users to provide messy, uncertain backstory, letting the AI help clarify thoughts.
  • The steer feature in Codex lets users update prompts mid-execution, enabling parallel human-AI work instead of rigid turn-based prompting.
  • Lou built a structured Obsidian file vault for Codex memories, arguing work should leave behind inspectable artifacts, not just trapped chat history.
  • Codex's tool use - computer, browser, and connectors - transforms it from a chat interface into an evidence-gathering work system that needs full environmental access.
  • Lou uses heartbeats, or scheduled check-ins, to create autonomous feedback loops where Codex monitors tools like Slack and triggers actions without human intervention.
  • The side panel in Codex is where parallel processing happens, allowing users to inspect and annotate artifacts while the agent continues working.
Also from this episode: (3)

AI & Tech (3)

  • Cursor is training a new model from scratch using XAI's Colossus 2 cluster, which has a million H100 GPU equivalents.
  • Cloudflare's review finds Anthropic's secretive Mythos model can create multi-step exploit chains and generate functional exploit proofs, working like a senior security researcher.
  • A jury dismissed Elon Musk's lawsuit against OpenAI in two hours, ruling his breach of charitable trust claim was barred by a three-year statute of limitations.
#Models#Startups#Safety#Agents#Enterprise

Beating the AI Doom Cycle • May 18

  • Klein suggests AI is more likely to displace a limited number of workers rather than cause mass unemployment. He argues this localized displacement, similar to the 2 million jobs lost to Chinese competition, might be harder to manage than a mass event.
  • Anthony Pompliano notes a shift in his view, now observing increasing hires for software engineers and growing open roles. He states that new college grad hires rose 5.6% over the past 12 months, and unemployment for 20-24 year-olds with college degrees dropped from nearly 9% to almost 5%.
  • A Wall Street Journal and LinkedIn analysis reports AI created 640,000 jobs in the US between 2023 and 2025, including new white-collar positions like Head of AI. Pompliano suggests companies are aggressively hiring, leveraging AI to boost employee productivity.
  • The market narrative has shifted from "seats" to "tokens," as AI agents consume vastly more resources, making a single user generate hundreds or thousands of dollars monthly in token sales. This redefines the economic model for AI companies.
  • Anthropic's run rate revenue surpassed $30 billion by early April. SemiAnalysis reports its ARR surged from $9 billion to over $44 billion today, doubling every six weeks, with inference margins reportedly at 70%, up from 38% last year.
  • Ming Li highlights Anthropic's rapid growth, noting it's adding $96 million in ARR per day. He contrasts this with AWS taking 13 years to reach $35 billion and Salesforce over 20 years to pass $20 billion, indicating that old software valuation frameworks are obsolete.
  • Atlassian's stock surged almost 30% after its earnings report, with revenue growth accelerating to 32% (from 23% last quarter). Customers using their new AI search tool, Rovo, increased their own ARR at twice the rate of non-users.
  • Atlassian's Rovo leverages existing knowledge graphs in Jira and Confluence for context, significantly reducing token usage compared to token-hungry RAG search. This token efficiency, crucial in an agent intelligence era, helps maintain seat-based pricing models.
Also from this episode: (14)

AI & Tech (11)

  • Nathaniel Whittemore identifies an emerging "AI vibe shift," characterized by a growing discourse that moves beyond doomerism, appearing in both job market discussions and financial markets. This dual emergence suggests a potential narrative shift rather than a temporary blip.
  • Jasmine Sun's New York Times essay, "Silicon Valley is bracing for a permanent underclass," reflects a prevalent doomer narrative among AI builders in Silicon Valley who believe AI will eliminate jobs. Whittemore questions over-reliance on builders' perspectives due to potential biases and their limited understanding of broader economic impacts.
  • Ezra Klein argues that the AI job apocalypse is unlikely, highlighting AI labs' role in creating negative narratives to excite investors or unwind post-COVID hiring. Economists, he notes, are generally skeptical of mass joblessness.
  • Klein cites ASU professor Eldar Maximov's research, indicating that employment grew faster in occupations heavily adopting computers than in those that did not. Cost reductions from task automation led to new demand, expanding overall occupations.
  • Klein also uses a personal example, noting his podcast team grew significantly after starting with one researcher, leading to more extensive and challenging episodes. He observes that enthusiastic AI adopters are working harder than ever due to increased possibilities.
  • Macroeconomic data does not support the doomer narrative, with the unemployment rate at 4.3% in March 2026 and 4.4% in March 2020, and stable average hourly earnings. Demand for software engineers, despite AI exposure, continues to accelerate, up 18% since May last year.
  • Merzmik Ahmed argues that AI increasing demand for software engineers is now a tech consensus. He applies Jevons' paradox, explaining that cheaper "digital bricks" (code) make previously expensive projects feasible, thus expanding demand for builders.
  • Greg Eisenberg predicts an unprecedented explosion of entrepreneurship as intelligence becomes cheaper and displaced workers create new businesses out of necessity and opportunity. Stripe Atlas data supports this, with 100,000 incorporations and Q1 up 130% year-over-year.
  • Roger Karma in The Atlantic notes a market shift: from concerns over AI infrastructure lacking profitability six months ago, to fears of insufficient data centers today. This change is driven by revenue catching up to hype, particularly with the rise of "Claude code."
  • Sam Altman, OpenAI CEO, stated on May 1st that the goal is to augment and elevate people, not replace them, and believes jobs doomerism is likely wrong long-term. He also noted that even with powerful AI, he has "never been busier."
  • Economic commentator Noah Smith views Altman's statements as a "huge messaging pivot" for OpenAI and the AI industry, which previously explicitly stated goals of replacing humanity. This shift suggests a more constructive dialogue about AI's impact.

Big Tech (2)

  • Morgan Stanley raised CapEx forecasts for five hyperscalers to $805 billion this year (from $765 billion) and $1.1 trillion next year (from $951 billion). Andreas Steno Larsen notes that the backlog of demand for additional capacity is rising even faster than CapEx spend.
  • Mag Seven companies spent over $400 billion in CapEx in Q1 this year, but their reported and projected backlog stands around $1.3 trillion, indicating significant unmet demand. David Sacks argues AI CapEx will contribute a 2% tailwind to GDP growth this year, potentially 2.5% this year and over 3% next year per Morgan Stanley.

AI Infrastructure (1)

  • The Associated Press reports that construction unions are collaborating with tech companies to overcome community opposition to data centers. Rob Bear of the Pennsylvania Building and Construction Trades Council emphasizes data centers create significant construction jobs.
#Agents#Enterprise#Startups#Labor#Macro

AI Inequality • May 17

  • Nathaniel Whittemore defines Q2 2026 as the AI 'second moment,' shifting from chatbot assistants to workable agentic systems, with stakes marked by billions of weekly users and $650 billion in planned capex.
  • Claude Code revenue grew from $1 billion to $2.5 billion annualized in two months. Anthropic's enterprise share reached 70% of first-time buyers, and its overall revenue run rate hit $19 billion.
  • The quarter saw rapid frontier model releases: GPT-5.2 Codex, Genie 3, Opus 4.6, GPT-5.3 Codex, Sonnet 4.6, Gemini 3.1 Pro, Nano Banana 2, and GPT-5.4, with no single benchmark winner across common tests.
  • OpenClaw became the most starred open-source project on GitHub. Nvidia launched Nemo Claw as an enterprise-grade wrapper, and Anthropic integrated its features into Claude Code and Claude Co-work.
  • Enterprise AI shifted from pilots to production, with 40% of enterprises predicted to have working agents by end of 2026. Pulsia, a fully agentic company, reached $6 million annualized revenue with a single founder.
  • Survey data shows 71% of practitioners used vibe coding, 62% used agentic automation, and the average respondent uses 3.5 models. ROI shifted from time savings (13.6% of use cases) to increased output and new capabilities.
  • Customer service AI adoption is mature with 91% of businesses experimenting, but 64% of customers prefer no AI in interactions. Legal AI adoption lags, with only 15% of tasks using AI despite 80% capability.
  • HR AI deployment grew 320% in 12 months, from 19% to 61%. Finance AI adoption faces data quality obstacles, with 91% of firms reporting low impact.
  • Generative Engine Optimization (GEO) market was under $1 billion in 2025, projected to reach $34 billion by 2034. Sales AI use cases are mature, with 63% categorized as 'primetime' for most organizations.
  • Nathaniel Whittemore argues the capability overhang - the gap between AI's potential and deployed value - is widening, increasing the disparity between leading and lagging companies.
Also from this episode: (1)

AI & Tech (1)

  • Anthropic and the Pentagon clashed over terms for Claude's use, leading to Anthropic being designated a supply chain risk and a subsequent lawsuit. ChatGPT's agreement with the Department of War triggered a 775% surge in one-star reviews.
#Agents#Enterprise#AI Infrastructure
This Week in Startups
This Week in Startups

Why is Gen Z hates AI? • May 18

  • Jason Calacanis advocates for 'delightful scale' companies making $500k to $5M annually as a viable path, arguing necessity and modern tools like ChatGPT lower the barrier to entry.
  • Jason Calacanis advises against using AI alone to file patents due to the high-stakes, specialized legal nature of IP, though it can be used with a human-in-the-loop for initial work.
  • Calacanis believes running AI models locally on hardware like Mac Studios is the future for privacy and cost reasons, not a fad, especially for sensitive corporate data.
Also from this episode: (13)

AI & Tech (6)

  • Jason Calacanis argues college graduates fear AI not out of simple anxiety but from feeling betrayed by tech leaders like Eric Schmidt, who they believe have bad intent about job displacement.
  • Calacanis says this generation has used ChatGPT for two years to complete degrees and understands the technology well, which fuels their cynicism about future job prospects.
  • Hosts note a stark gap between business excitement about AI and average consumer sentiment, particularly among recent graduates entering the workforce during economic uncertainty.
  • Calacanis believes the only way to escape a permanent underclass in the AI economy is to start a company, as being a worker for someone else makes you a cost center.
  • The Information reports Anthropic and OpenAI generate 89% of all AI startup revenue, creating a duopoly that is consolidating power.
  • Hosts note that six months ago, Anthropic and OpenAI represented 4.5% less of total startup AI revenue, indicating they are rapidly gaining market share.

AI Infrastructure (3)

  • A University of Utah study projected US data center construction jobs will peak and then decline, with total direct operations jobs reaching only about 65,000 by 2030.
  • Calacanis highlights that tech companies cut 100,000 jobs in the first part of the year, a figure already surpassing the total data center jobs projected for 2030.
  • Calacanis argues AI model providers like OpenAI and Anthropic are likely selling tokens at a massive loss currently, similar to Uber and Lyft's early subsidized growth phases.

Education (1)

  • A viral New York Times essay by Stanford senior Theo Baker claimed cheating with AI is omnipresent on campus, dissolving the foundations of liberal arts education faster than the workforce.

Startups (1)

  • Hosts discuss the risk for application-layer AI startups, as major labs gaining access to their usage data could quickly build competing features and erase their market.

Society (2)

  • The shooting incident in Austin highlighted the privacy-versus-safety debate around Flock Safety's license plate readers, which the city council had removed but a neighboring county used to capture suspects.
  • Calacanis suggests potential safeguards for surveillance tech like Flock include strict audit trails, biometric access logs, and enforced data retention policies of up to 36 months.
#Startups#Labor#Big Tech#Regulation

Related Stories