AI Daily: Google Tests Clinical Video AI, Cognition Hits $40B Target, and Open-Weight Debates
Autonomous clinical agents are stepping into patient care while capital flooding the AI software development market continues to shatter records. From Google testing a tri-agent system that leads live video consultations to Cognition targeting a staggering $40 billion valuation, today’s updates show how fast domain-specific AI is converting from pure research into market dominance. Here is the breakdown of the biggest developments across the AI landscape today.
1. Google Tests AMIE Multi-Agent System for Clinical Video Consultations
In one of the most rigorous trials of real-time clinical AI to date, Google put its research system, AMIE (Video), through 15 live video consultations with trained actor patients across five core body systems. An independent evaluator panel made up of 20 board-certified physicians assessed the system on diagnostic accuracy, communication skills, and history-taking, rating AMIE directly on par with human primary care doctors. In fact, the evaluators scored the system higher than real doctors at guiding patient actors through physical self-examinations on camera. Having spent years looking at latency bottlenecks in live systems, Google’s decision to split the workload across three distinct background agents (a talker, a planner, and a perception engine) is a smart way to keep spoken interactions natural without dropping diagnostic rigor.
2. AI Safety Pioneers Clash Over Open-Weight Models at Ai4
“I don’t want there to be gatekeepers.”
At the Ai4 conference, Geoffrey Hinton, Fei-Fei Li, and Andrew Ng took the stage to debate the fate of open-weight artificial intelligence. Andrew Ng warned against centralized regulatory “gatekeepers,” pointing out that if lower-cost open models out of China win widespread adoption across Africa and Asia, those systems will inevitably shape how billions of users view digital rights and democracy. Geoffrey Hinton admitted that while he once fought open-weight releases over cyberattack fears, he now views that safety battle as effectively lost. Meanwhile, Fei-Fei Li rejected the binary framing entirely, comparing AI to nuclear physics—where foundational science remains open while specific dangerous implementations are heavily regulated.
3. Cognition Targets $40 Billion Valuation on $1 Billion Revenue Run Rate
AI software development startup Cognition, creator of the autonomous coding agent Devin, is in early talks to secure a new funding round that could push its market valuation to $40 billion.
- Valuation Step: Up from $26 billion back in May 2026.
- Revenue Scale: Annualized revenue run rate is now closing in on $1 billion.
- Market Context: Annualized revenue has doubled from $492 million in under three months.
When a startup grows its top line that quickly by handling messy engineering tasks like legacy code updates, investors stop treating AI as a speculative play and start pricing it like fundamental software infrastructure.
4. Blacksmith’s Valuation Jumps 10x as AI Code Validation Explodes
Software verification startup Blacksmith closed a $45 million Series B round led by Peak XV Partners, valuing the business at $550 million. That represents nearly a 10x valuation jump from its $60 million evaluation less than a year ago. As autonomous agents write an ever-increasing percentage of enterprise code, checking that synthetic output for subtle bugs has turned into a massive business. Having debugged LLM-generated code myself, I know how easily subtle edge cases slide past code reviews; automated validation is fast becoming the real bottleneck in modern software delivery.
5. OpenAI-Backed Thrive Holdings Raises $2 Billion for Enterprise Acquisitions
Instead of pitching software subscriptions to traditional companies, Thrive Holdings buys legacy firms in sectors like accounting and IT, then completely rebuilds their operational backends around generative AI.
The Private Equity AI Playbook
- Funding Raised: $2 billion in new equity led by SoftBank at a $12 billion valuation.
- Strategic Backing: Spinout of Thrive Capital, with OpenAI holding a direct equity stake.
- Proven Results: A tax-processing portfolio unit handled over 7,000 tax returns with a 98% accuracy rate.
OpenAI embeds its own engineers directly into Thrive’s portfolio businesses to rebuild workflows from the ground up, cutting out standard enterprise sales cycles entirely.
6. My Take
The explosive numbers for Cognition and Blacksmith prove that AI coding tools have moved past the initial hype curve into raw commercial execution. But the most important shift today is Thrive Holdings’ $2 billion raise. Selling AI tools to traditional enterprises is historically painful because middle management resists changing existing routines. By buying legacy businesses outright and forcing an AI-first operational architecture from the top down, Thrive and OpenAI are bypassing corporate red tape entirely—a strategy that will likely trigger a massive wave of AI-driven private equity rollups over the next few years.