Prime Radiant/Machine Cards
DMDayCANONclass card

Google DeepMind AI Lab (2014)

infrastructure pace layer · 2014–ongoing

lifespan: 50 yrs · motor: push

Class card for the Google DeepMind research-and-deployment AI lab — the unified successor to DeepMind Technologies (founded London September 2010 by Demis Hassabis, Shane Legg, and Mustafa Suleyman; Google acquisition January 2014 ~$500M) and Google Brain (founded Mountain View 2011 by Jeff Dean, Geoffrey Hinton, Andrew Ng; merged into Google DeepMind April 2023 under Hassabis as CEO). Operational telos: scientific AI breakthrough (DeepMind heritage — AlphaGo, AlphaFold, AlphaZero) + commercial AI deployment (Google Brain heritage — Gemini model series). Post-merger (April 2023): both missions consolidated under one org. t_active_start = 2014 (Google acquisition date — the moment the machine's DM-Day form begins; DeepMind Technologies 2010-2014 encoded in DM-Dawn snapshot). Motor: push (DM default — independent capability-push without responding to a Progress telos). Substrate incorporeal + cognitive + semiotic: the operating substrate is ML training pipelines, model weights, protein-structure databases, and the scientific-AI-as-machine-intelligence narrative that frames DeepMind's contribution to global AI discourse. Heavy corporeal substrate (Google TPU clusters, GCP datacenters) externalized to parent machine:google-search-advertising-1998. AlphaGo March 2016 defeats Lee Sedol 4-1 (Seoul) — canonical DM-Day inflection for AI-as-strategic-resource framing. AlphaFold 1.0 CASP13 2018; AlphaFold 2 CASP14 Nov 2020 (Nature 2021); ~200M+ protein structure predictions (July 2022 database). Nobel Prize in Chemistry Oct 9 2024 (Hassabis + Jumper for AlphaFold; David Baker shared). Gemini 1.0 Dec 6 2023 (Ultra/Pro/Nano); Gemini 2.0 Flash Dec 11 2024. dm_current = divergentism (rare positive case): fragmentation LOW, narrative coherent post-AlphaFold Nobel 2024, scientific capacity HIGH. Unlike OpenAI (late_modernity / energetic zombie) or Meta AI (postmodernity), Google DeepMind exhibits genuine DM-Day capacity with a coherent scientific mission plus commercial deployment path. w0 atlas seed: fragmentation LOW, heretic MEDIUM. V0.2 GAP: tep6_ai_biotech paradigm primary contested (AlphaFold straddles tep5_ict + tep6_ai_biotech; encoded as tep_paradigm_primary=tep5_ict, tep_paradigm_secondary=tep6_ai_biotech, contested=true). heavy_or_light: light — operational grammar is incorporeal (models, publications, algorithms); heavy compute substrate held by Google/Alphabet parent.

Machine type

incorporeal

Plasticity

plastic

Substrate

incorporeal cognitive semiotic

Wave source

wave-9-atlas

Inputs

  • google_alphabet_compute_infrastructure_tpu_gcp
  • alphabet_capital_research_budget
  • world_class_ml_researcher_labor
  • protein_databases_pdb_uniprot_for_alphafold

Outputs

  • alphago_alphazero_game_ai_series
  • alphafold_protein_structure_database_200m_predictions
  • gemini_model_series_api_and_consumer
  • academic_publications_nature_science_peer_review

Landscape pressures

  • llm_commercialization_vs_scientific_mission_tension (65% intensity)
  • compute_arms_race_capital_intensity (85% intensity)
  • regulatory_agi_governance_planetary (68% intensity)

Intra-era couplings

Cross-era couplings

State variables

heretic_density
0.55
CANON
push_fragmentation_count
2
CANON
narrative_coherence
0.80
CANON
gravitational_weight
0.88
CANON
real_virtuality_saturation
0.82
CANON
black_box_depth
8
CANON
plasticity_demand
0.85
CANON

Phase snapshots

DM-Dawn2010–2014complex
DM-Day2014–2026complex

Notable instances

  • DeepMind Technologies founding — London, September 2010 (2010) — Founded by Demis Hassabis (Cambridge BA, UCL PhD), Shane Legg (IDSIA), Mustafa Suleyman (Oxford). Early investors: Found…
  • Google acquisition of DeepMind — January 26, 2014, ~$500M (2014) — Google acquires DeepMind Technologies January 26 2014 for reportedly ~$500M. Hassabis + Legg remain as lab co-leads. Dee…
  • AlphaGo — March 9-15, 2016 (DeepMind Challenge Match, Seoul) (2016) — AlphaGo defeats Lee Sedol 4-1 in the DeepMind Challenge Match, Seoul, March 9-15 2016. First AI to defeat a world-champi…
  • AlphaZero — December 2017 (2017) — AlphaZero (Dec 2017) masters Go, Chess, and Shogi tabula rasa via self-play reinforcement learning — single algorithm, n…
  • AlphaFold 1 (CASP13 2018) + AlphaFold 2 (CASP14 Nov 2020; Nature 2021) (2018) — AlphaFold 1: CASP13 2018 (dominant score). AlphaFold 2: CASP14 Nov 2020 — Jumper + Hassabis lead; Nature paper July 2021…
  • Google DeepMind merger — April 20, 2023 (2023) — Google DeepMind merger April 20 2023 (Sundar Pichai restructure). DeepMind + Google Brain merge into single org under Ha…
  • Gemini 1.0 — December 6, 2023 / Gemini 2.0 Flash — December 11, 2024 (2023) — Gemini 1.0 Dec 6 2023 (Ultra/Pro/Nano tiers); Gemini 2.0 Flash Dec 11 2024. First post-merger frontier model series. Has…

Sources

  • Hassabis, Demis (2024). AlphaGo and AlphaFold interviews, DeepMind blog posts 2016-2024 · 88%
  • Jumper, John et al. (2021). Highly accurate protein structure prediction with AlphaFold (Nature) · 95%
  • Silver, David et al. (2016). Mastering the game of Go with deep neural networks and tree search (Nature) · 95%
  • Silver, David et al. (2018). A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play (Science) · 92%
  • Nobel Committee for Chemistry (2024). Press release: The Nobel Prize in Chemistry 2024 · 98%
  • Alphabet Inc. (2024). Annual Report 2024 (Form 10-K) · 92%