Release tension.

The article focuses on how leading AI labs are using exclusivity and staged releases of advanced models as both a marketing strategy and a security measure in a highly competitive industry. Topics include Anthropic's limited rollout of its "Mythos" model to control cyber-security risks, concerns ove...

Full description

Bibliographic Details
Published in:Economist Vol. 459; no. 9495; p. 64
Format: Article
Published: Economist Newspaper Limited 4/18/2026
Subjects:
Online Access:View this record in EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=193157014&site=ehost-live
header:
  @attributes:
    shortDbName: ssf
    uiTerm: 193157014
    longDbName: Social Sciences Full Text (H.W. Wilson)
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00130613
        ECO
      jtl: Economist
      issn: 00130613
      maglogo: N
    pubinfo:
      dt: 4/18/2026
      vid: 459
      iid: 9495
      pid: 161
      pub: Economist Newspaper Limited
    artinfo:
      ui: 193157014
      ppf: 64
      ppct: 0
      formats:
      tig:
        atl: Release tension.
      aug:
      su:
        Artificial intelligence
        Marketing strategy
        Security systems
        Computer network security
      sug:
        subj:
          Artificial intelligence
          Security Systems Services (except Locksmiths)
          Marketing strategy
          Security systems
          Computer network security
      ab: The article focuses on how leading AI labs are using exclusivity and staged releases of advanced models as both a marketing strategy and a security measure in a highly competitive industry. Topics include Anthropic's limited rollout of its "Mythos" model to control cyber-security risks, concerns over model theft and "distillation" by rival firms, and broader industry constraints such as limited computing power and rising demand for frontier AI systems.
      pubtype: Periodical
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: N
    holdings:
      @attributes:
        islocal: N