Jul 30, 2026 – 5.00amIn April, the artificial intelligence arms race that had captivated markets and users since the public launch of ChatGPT in November 2022 suddenly took on a more sinister tone.Anthropic, the company behind the Claude models that were so good at vibe-coding that they caused the so-called SaaSpocalypse sell-off of software stocks, announced that its new model Mythos was being held back from public release.Subscribe to gift this articleGift 5 articles to anyone you choose each month when you subscribe.Subscribe nowAlready a subscriber? Paul SmithTechnology editorPaul Smith edits the technology coverage and has been a leading writer on the sector for 20 years. He covers big tech, artificial intelligence, business use of tech, the fast-growing Australian tech industry and start-ups, and national innovation policy from our Sydney newsroom. Send tips to @sayssmithy.99 on encrypted messaging platform Signal.Fetching latest articles
AI: the new arena for the cyber arms race
The US withheld access to new models from its allies for cyber testing. Now China is offering comparable models and experts are asking how Australia can stay safe.
TL;DRAI
Anthropic holds back Mythos from public release, marking a strategic shift toward caution in the AI arms race. For tech leaders, the slowdown in model launches signals a governance-first pivot: risk management and compliance now drive deployment decisions over pure technological velocity.
142 words~1 min read
