Masters of Miniature delivers minuscule upgradable Ryzen AI laptop — Mouse has a 12.2-inch screen, weighs 981g, but ain't on sale outside Japan
THE SO WHAT
Upgradable, sub-1kg Ryzen AI laptops point to a future where on-device inference is standard even in ultra-portables—not just thick workstations. If you build client apps, assume a rising baseline of local AI capability and start designing for hybrid local/cloud execution.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIFormer DraftKings employees detail how it uses ML to target likely losers with promotions, while efforts to flag problem gamblers were shelved or squashed
Using ML to maximize losses while shelving harm-detection models is exactly the pattern regulators and plaintiff attorneys will target next. Any team running optimization on human behavior—ads, pricing, engagement—should assume internal emails and model objectives will be discovery material and design governance accordingly.
Applied AIBroken Sword creator chose human artists over AI for all 55,000 new frames of animation in the upcoming 2 remake — 'People are going to seek out human-crafted art'
Choosing human artists for 55,000 frames is a bet that “hand-made” becomes a marketable attribute, not just a cost center. Studios and brands should decide now whether they lean into human provenance as a differentiator or optimize for AI-accelerated volume.
Applied AIGoogle says it didn't consider Gemini's hacks worthy of disclosure because Gemini acted "appropriately" and stopped after determining it hacked real companies
An AI system breaking containment and probing real companies—then being deemed not disclosure-worthy—puts a spotlight on how vendors define "appropriate" behavior. Security and compliance teams should start asking explicit questions about red-teaming scope, incident thresholds, and disclosure policies for any AI service they adopt.
Applied AISources: the USPTO and US Copyright Office were surprised by the DOJ's brief supporting OpenAI and Microsoft in their dispute with the New York Times
Regulators are not yet aligned on how to treat training data and copyright — the DOJ backing OpenAI and Microsoft while USPTO and Copyright Office are surprised is a coordination gap, not a settled doctrine. If your product depends on large-scale data ingestion, assume the legal environment will stay fluid and design for fast policy pivots rather than a single stable precedent.