Deepmind was built to chase AGI, but its new chief just wants Gemini 4 out the door
What happened
Koray Kavukcuoglu, the new chief of Google Deepmind, is pushing to release the Gemini 4 model much earlier than the previously expected timeline of late 2023. Gemini 4 is already in the post-training phase and is running internally in Deepmind’s coding tool called Antigravity. This marks a clear shift in focus from the lab’s earlier ambition of chasing artificial general intelligence (AGI) under Demis Hassabis to prioritizing practical, trustworthy AI agents and faster product delivery.
Why it matters
Deepmind’s change in leadership style signals a move away from speculative AGI ambitions toward more immediate product milestones. This shift pressures the AI field by emphasizing deployable, reliable models rather than long-term AGI breakthroughs. Developers and businesses who track AI innovation will find Gemini 4’s accelerated deployment relevant because it suggests more concrete tools sooner, especially for coding and automation tasks. It also signals a strategic pivot at Deepmind after the quiet disappearance of Gemini 3.5 Pro and the departure of several top researchers to OpenAI and Anthropic, potentially weakening Deepmind’s position in the AGI race.
Koray’s dismissal of the AGI debate as “not the right conversation” and emphasis on trustworthy agents raises the bar on responsible AI deployment. For enterprises and product teams, this means new AI releases from Deepmind will likely focus on practical performance and safety, not just raw capabilities or theoretical milestones.
What to watch next
Observe how Gemini 4 performs upon release, particularly in real-world coding tasks and integration with developer tools like Antigravity. The success or failure of this model will test if Deepmind’s pivot delivers usable AI faster and whether it can recover some ground lost to OpenAI and Anthropic in talent and platform adoption.
Also worth monitoring is whether Deepmind continues deprioritizing the AGI narrative or if this is a temporary product-driven phase. How the lab balances ambition with practical deployment will influence investor confidence and industry dynamics around AI innovation.
AI Quick Briefs Editorial Desk