/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

An analysis of whether AI will destroy, displace, create, and accelerate jobs in the coming years like other waves of automation, or if this time is different

Pretty much everyone in tech agrees that generative AI, Large Language Models and ChatGPT are a generational change …

Benedict Evans

Context & Ripple Effects

This piece closes a loop that opened six months earlier, when ChatGPT and generative ML marked a step change in use cases and raised the tolerance-for-error question that any job-displacement claim has to answer. Since then the coverage has moved from what the models are to whom they touch: Stratechery mapped the impact on Apple, Amazon, Meta, Google, and Microsoft in January, and by March experts were cautioning that short-term promise and peril are more modest than the fervor suggests.

First-order effects

  • Companies deploying generative coding tools get more output from experienced programmers while eliminating tasks currently done by junior developers — the WSJ reporting frames this as a cost-saving substitution, not just augmentation.
  • The named big-five platforms (Apple, Amazon, Meta, Google, Microsoft) face pressure to decide whether LLMs are a threat to embed or a capability to sell into every workflow.

Second-order effects

  • If firms stop hiring juniors because seniors plus AI cover the work, they cut off their own pipeline of future senior engineers — a knock-on labor-market effect the productivity gains don't price in.
  • Vendors of developer tooling and 'answer engine' replacements for search compete to own the interface layer where this displacement actually happens, shifting value toward whoever distributes the models.

Third-order effects

  • Whether this wave differs from prior automation turns on the error-tolerance question flagged in the December step-change analysis: cognitive work with low tolerance resists displacement longer than the hype implies, stretching the transition over years rather than months.
  • A sustained junior-hiring pullback would restructure how the industry trains expertise, making apprenticeship-by-task something companies must deliberately rebuild rather than get for free.

The trend: Generative AI is entering its distribution phase, where the binding question shifts from model capability to which tasks — and which rungs of the career ladder — the tools absorb first.

Discussion

  • @fudge.org Jay Cuthrell on bluesky
    In the late 90s, a MBA graduate assured me Microsoft FrontPage would make “web people” obsolete.  Bruce Sterling's “spicy auto complete” aside, AI models combined with networked robotics will eventually impact everything that is manually intensive and artisanal.  [embedded post]
  • @stevesi Steven Sinofsky on x
    AI and the automation of work // @benedictevans with a deeply historical explainer on how AI will transform work and jobs. He does this better than anyone writing today. https://www.ben-evans.com/... [image]
  • @newley Newley Purnell on x
    .@benedictevans on ChatGPT, AI, and the automation of work: “We don't have AGI, and without that, we have only another wave of automation, and we don't seem to have any a priori reason why this must be more or less painful than all the others.” https://www.ben-evans.com/...
  • @benedictevans Benedict Evans on x
    I wrote something about AI, ChatGPT, productivity and jobs. https://www.ben-evans.com/...
  • @dok2001 Dane Knecht on x
    I got stuck on the fact that Evernote according to one report is still top workplace shadow IT app. https://twitter.com/...
  • @aridavidpaul @aridavidpaul on x
    Imo, AI the first “tool” in human history that's not a tool. Even the first gen models we have today contain a feedback loop - they produce outputs based on what humans want. That's a full production cycle w/feedback enabling ongoing optimization and evolution. /1 https://twitter…
  • @willcritchlow Will Critchlow on x
    Connecting two things I read: 1. https://mxphi.com/... 2. https://www.ben-evans.com/... Right now, the best applications of ChatGPT look a lot like NP tasks: hard to do, but easy to check.
  • @eswarpr @eswarpr on x
    @stevesi @benedictevans I'm with @benedictevans on this. Working with 3 entrepreneurs on 3 distinct ideas only made possible by GenAI. I find it to be a very App Store moment, with a little bit of first mover advantage but much more of the usual are-we-solving-a-real-problem clas…
  • @_jeremyflores Jeremy Flores on x
    1/ The problem with this article is that it skips over a problem most people are skipping over: you don't need AGI for the current techniques and approaches to be extraordinarily dangerous to the labor force. Short 🧵: https://www.ben-evans.com/...