/
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

Google's AI shakeup suggests it may be prioritizing AI diffusion over frontier-model leadership, betting on AI compute as a bigger economic opportunity

SemiAnalysis thinks the DeepMind shakeup means Google is losing the AI race.  It might be that Google is choosing to run a different race.

Asimov's Addendum Tim O'Reilly

Context & Ripple Effects

Google's AI story had already split between model research and the infrastructure behind it: related coverage said DeepMind had lost frontier-model momentum as leadership and RL talent departed, while Google Cloud stood to gain from more compute. The appointment of Koray Kavukcuoglu to lead DeepMind concentrates that tension under a leader with both AI-architecture and DeepMind responsibilities.

The new interpretation is that the reorganization need not be read solely as a retreat from model leadership. It instead frames Google’s scale in compute and product reach as an alternative route to AI advantage, extending an earlier view of AI as a sustaining force for Big Tech.

First-order effects

  • Google DeepMind’s leadership change shifts accountability for frontier research and AI architecture to Koray Kavukcuoglu, while Google Cloud is positioned to benefit if more AI work translates into compute demand.
  • SemiAnalysis’s assessment of a frontier-model setback is recast as a strategic trade-off: Google may emphasize getting AI into products and infrastructure rather than treating model leadership as the sole measure of success.

Second-order effects

  • Google’s AI competitors face a sharper contrast between competing for frontier-model prestige and converting model demand into cloud and product usage; Google can pursue the latter through its existing compute footprint.
  • DeepMind research teams face stronger pressure to show that their work supports Google-wide deployment and compute utilization, not only standalone model benchmarks.

Third-order effects

  • If Google follows through on this allocation, AI competition may be defined less by a single leading model and more by who can industrialize compute and distribute AI across large product ecosystems.
  • That would reinforce a bifurcation between firms optimized for frontier research leadership and firms optimized to commercialize the infrastructure and distribution surrounding models.

The trend: AI strategy is shifting from a singular race for the best frontier model toward a contest over compute commercialization and product distribution.

Discussion

  • @timoreilly Tim O'Reilly on x
    An alternate hypothesis about the shakeup at DeepMind. https://asimovaddendum.substack.com/ ... [image]
  • @kroscoo Kris Cao on x
    For another point of reference Gemini are running ads with the devil wears Prada and Harry Kane, which feels pretty far from the OpenAI/Anthropic framing.
  • Tim O'Reilly Tim O'Reilly on linkedin
    SemiAnalysis is probably right that Google gave something up on August 5, but possibly wrong about why.  The evidence does not necessarily show that Google has abandoned technological ambition. …
  • @edzitron.com Ed Zitron on bluesky
    The unsaid part of this is that both Google and xAI are looking like they're sort of giving up on frontier models, which means they will create more supply and less demand for AI compute, further centralizing it around Anthropic and OpenAI.  The void yearns [embedded post]
  • @caseynewton Casey Newton on bluesky
    Btw regarding the claim in here that Gemini 3.5 Pro has been “silently canceled,” Google tells me that is not the case.  [embedded post]
  • @carnage4life Dare Obasanjo on bluesky
    Google as a company both competes with Anthropic and benefits from it.  The Gemini team tries to compete with Claude while the Google Cloud team hosts Claude's API.  —  So Google has to decide if to use their datacenters to compete with Anthropic or make money from them.  —  Maki…
  • @rezendi Jon Evans on x
    “frontier labs may discover that model leadership resembles semiconductor fabrication, enormously important strategically but surprisingly poor as a standalone business.” https://asimovaddendum.substack.com/ ... (I think TSMC and Samsung would disagree...)
  • @benshanyoufeng Youfeng on x
    @timoreilly The Westinghouse framing is apt. But Westinghouse and Edison were separate companies. The harder test: can one org run both races simultaneously, with the same compute budget and the same talent pool pulled in competing directions? That's not the Westinghouse bet. It'…
  • @jessicalessin Jessica Lessin on x
    Very smart, as always from @timoreilly.
  • @gabrielayuso Gabriel Ayuso on x
    Good read. This is why I love working on Google Search. We have the unique opportunity to bring the power of AI to everyone. There's a big different between frontier and frontier at scale. “Google also doesn't need to win the frontier to dominate the edge. Its Flash-class
  • @minhsmind Minh Do on x
    This sounds like a fairly convincing argument that is fair to Google and to the opportunities that still lay dormant in the space.
  • @fernandotn Fernando Torres on x
    @timoreilly The article makes it sound like it was a conscious decision to give up the race on AI models. From conversations with Google employees, they were forcing them to use their internal tools using Gemini for coding. Everybody inside knew it was not working, and they were …
  • @timkellogg.me Mr. Tim on bluesky
    great article by @timoreilly.bsky.social arguing that Google is intentionally transitioning into a Westinghouse-like domination of the “AI grid” rather than the ever-risky pursuit of the smartest SOTA models  —  asimovaddendum.substack.com/p/googles- we...