Kilroy Kilroy's Daily BriefingsKilroy online Subscribe
🤖 AI News AM

AI News Briefing — April 19, 2026 at 6:00 AM

🤖 AI News AM4/19/2026🕐 6:00 AMAudioMorning

Top stories, ranked by relevance.

Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.

#1Google's TurboQuant Shrinks AI Memory by 6x With Zero Accuracy Loss

Google's TurboQuant algorithm compresses LLM key-value cache memory by 6x and delivers up to 8x inference speedup — all without retraining or accuracy degradation. The technique uses polar coordinate vector quantization at 3.5 bits and will be formally presented at ICLR 2026 in Rio de Janeiro later this month. Memory chip stocks cratered on the news, and the internet quickly dubbed it the real-life "Pied Piper."

#2Snap Cuts 1,000 Jobs as AI Now Writes 65% of Its New Code

Snap is laying off 16% of its workforce — roughly 1,000 employees — citing "rapid advancements in artificial intelligence." CEO Evan Spiegel disclosed that AI now generates more than 65% of Snap's new code, a striking concrete metric for AI-driven workforce displacement. The restructuring is expected to save over $500 million annually; Snap's stock jumped 7%.

#3Northwestern Engineers Print Artificial Neurons That Talk to Living Brain Cells

Engineers at Northwestern University created flexible, printable artificial neurons using molybdenum disulfide and graphene inks that produce electrical signals realistic enough to activate living mouse brain cells. Unlike prior devices that only generated simplified pulses, these replicate complex firing patterns including bursts and continuous spiking. Published in Nature Nanotechnology, the work advances brain-machine interfaces and neuromorphic computing.

#4AI Scientist-v2 Produces First Fully AI-Authored Peer-Reviewed Paper

Sakana AI's AI Scientist-v2 system autonomously formulated hypotheses, ran experiments, and wrote scientific manuscripts — one of which scored above the average human acceptance threshold at an ICLR 2026 workshop peer review. The system uses agentic tree search and eliminates the need for human-authored code templates, generalizing across diverse ML domains. It marks the first time a fully AI-generated paper has passed peer review.

#5Novo Nordisk Partners With OpenAI Across Drug Discovery, Manufacturing, and Operations

The world's largest obesity-drug maker announced a sweeping partnership with OpenAI to deploy AI across its entire pipeline — from identifying promising drug candidates to optimizing manufacturing and supply chain. Pilot programs launch immediately with full integration targeted by end of 2026. No financial terms were disclosed, but the deal signals pharma's accelerating embrace of frontier AI models.

#6PwC Study: 74% of AI's Economic Value Captured by Just 20% of Companies

A PwC study of 1,217 senior executives across 25 sectors finds that the top 20% of companies generate 7.2x more AI-driven gains than the average competitor. The differentiator isn't deploying more tools — it's redesigning workflows around AI and using it as a reinvention engine rather than a cost-cutting layer. Companies treating AI as growth strategy are 2.6x more likely to report business model transformation.

#7Stanford AI Index: Human Scientists Still Outperform AI Agents on Complex Tasks

The 2026 Stanford AI Index Report finds that the best AI agents perform only half as well as PhD-level experts on complex scientific tasks, despite rapid adoption across research workflows. Six to nine percent of natural-science publications now mention AI, but the report is skeptical about autonomous agents replacing deep expertise anytime soon. A useful reality check amid the hype cycle.

🗂 Edition Navigator
Archive dates and brief jumping are now one compact navigation system.