|
KEP AI BULLETIN
Thursday 18 June 2026 |
Edition 018
8 stories across 5 sections today.
|
|
Models & Research
|
|
Using AI to help physicians diagnose rare genetic diseases affecting children
Researchers applied an OpenAI reasoning model to help doctors diagnose rare genetic diseases in children. The system successfully identified 18 new diagnoses in cases that had previously stumped physicians. This points to a practical, measurable use case for AI in specialized medical settings.
|
|
New Tools & Products
|
|
Google bets on Gemini to reinvent the smart home speaker | TechCrunch
Google has released a new $99.99 smart home speaker powered by its Gemini AI model. Unlike older Google Assistant devices that required specific voice commands, the new speaker supports more natural, back-and-forth conversation. Google is using the product to test whether generative AI can revive consumer interest in smart speakers.
|
|
Chips & Infrastructure
|
|
France Advances Europe’s AI Future With NVIDIA Technologies
France has been building out AI infrastructure over the past year, including data centers and national computing capacity, in partnership with NVIDIA. That infrastructure is now operational, with AI agents running in production and startups deploying real applications. The effort signals France's push to develop a competitive, home-grown AI ecosystem within Europe.
|
|
Robotics & Physical AI
|
|
RealSense unveils AI-native D585 Pro depth camera for robots
RealSense has launched the D585 Pro, a new depth camera built specifically for robotics applications. The camera uses a proprietary fifth-generation chip and delivers more than twice the depth accuracy of its predecessor. Improved depth perception is a key requirement for robots operating in real-world environments.
|
|
Business & Investment
|
|
World leaders want American AI. They just don't want America to be able to turn it off. | TechCrunch
Leaders including France's Macron and India's Modi raised concerns at the G7 summit that the U.S. could cut off foreign access to American AI systems without warning. Those concerns became concrete when Anthropic was ordered to block access for foreign users almost overnight. The episode is prompting governments to reconsider dependence on U.S.-controlled AI platforms.
|
|
Anthropic got hit by export rules nobody understands
The Trump administration ordered Anthropic to immediately cut AI model access for all foreign nationals, including users based in the U.S. and the company's own staff, citing export control rules. Anthropic spent the week working to restore service while navigating regulations that appear poorly defined. The incident exposed significant uncertainty around how export laws apply to AI software.
|
|
World model maker Odyssey nabs $1.45B valuation backed by Amazon and other big names | TechCrunch
AI startup Odyssey has raised funding at a $1.45 billion valuation, with Amazon among its backers. The company builds "world models," a category of AI that simulates environments rather than just processing language. Investor interest reflects a broader bet that world models represent the next major frontier in AI development beyond large language models.
|
|
The White House Wants Anthropic to Block All Jailbreaks. That May Not Be Possible
The Trump administration has told Anthropic that it must guarantee its AI model cannot be jailbroken before it will be allowed to rerelease the product. Security researchers say that kind of absolute guarantee is technically not achievable with current AI systems. The standoff highlights a gap between government expectations and what AI developers can realistically deliver.
|
|
KEP Insight
|
|
When AI helps crack medical cases that stumped experienced doctors, it tells you something important: reasoning models are now genuinely useful in high-stakes, specialist domains, not just for writing emails or summarising documents. For SME owners, the lesson is straightforward - if this technology can handle the complexity of rare disease diagnosis, it can almost certainly handle the complex, repetitive expert judgement calls in your business, whether that is credit assessment, contract review, or technical fault diagnosis. The barrier to deploying serious AI capability is dropping fast, and you do not need to be a hospital or a tech giant to benefit. Start identifying where specialist knowledge in your business creates bottlenecks, because that is exactly where these tools are heading next.
— Prabhu Iyer, KEP
|
|
Knowledge Equity Partners
This bulletin is curated by KEP AI systems and reviewed for relevance. It is for informational purposes only.
|