AI digest: The agentic coding wars heat up
OpenAI and Anthropic battle for coding supremacy while enterprises throw serious money at AI infrastructure.
The big story this week is coding agents getting properly aggressive. Two major players just dropped significant updates that suggest we’re moving past basic autocomplete into full desktop takeover territory.
OpenAI turns Codex into a screen-watching coding companion
OpenAI massively expanded Codex to control Macs autonomously, generate images, and work on tasks for weeks without supervision. The update is a direct shot at Anthropic’s Claude Code. This feels like the moment coding assistants stop being helpers and start being proper colleagues. The always-on screen watching is either brilliant or terrifying, depending on your privacy tolerance.
Anthropic fires back with Claude Opus 4.7
Not to be outdone, Anthropic released Claude Opus 4.7 with major coding improvements. Interestingly, they deliberately reduced cybersecurity capabilities during training. That’s a fascinating design choice that suggests they’re thinking seriously about dual-use risks whilst pushing performance.
OpenAI goes vertical with GPT-Rosalind for life sciences
OpenAI launched GPT-Rosalind, their first domain-specific model targeting drug discovery and genomics. The promise is slashing 10-15 year development timelines through advanced biochemistry analysis. This signals OpenAI moving beyond general-purpose models into lucrative enterprise verticals. Expect more specialised models across different industries.
Enterprise AI valuations continue their rocket trajectory
Factory hit a $1.5B valuation for enterprise coding tools, whilst Upscale AI is reportedly raising at $2B after just seven months. The enterprise AI infrastructure gold rush is properly mental right now. These valuations suggest investors believe we’re still in the very early innings of AI adoption.