Mitchell Bryson

Today in AI — 24 May 2026

Today's top AI news — curated links and commentary on the stories that matter for product builders.

·3 min read

Several AI commerce players want to sit between users and checkout. Google, OpenAI, and Klarna are racing to own the buying moment, while Cerebras drew a strong market response for owning more of the inference layer underneath.

The checkout wars heat up

Google introduced Universal Cart for AI-powered cross-merchant shopping. Klarna launched an AI-powered Shopping Search app inside ChatGPT. Fast Company frames Google and OpenAI's moves as part of the race to build agentic commerce. For product builders, the question is which checkout surface becomes worth watching first.

Cerebras cashes in on speed

Days after its record-breaking $5.5 billion IPO (stock up 108% on debut), Cerebras demonstrated its chips running the trillion-parameter Kimi K2.6 at 981 tokens per second, 6.7x faster than GPU clouds. The market valued speed-of-inference as a standalone business, and Cerebras immediately delivered proof.

Agents working unsupervised

Alibaba's Qwen 3.7 Max ran autonomously for 35 hours to optimise code for Alibaba's custom chip. Telegram shipped bot-to-bot chats and chat automation. Together, the two releases push towards more continuous agent workflows.

AI goes to work

OpenAI launched a $4 billion Deployment Company to embed AI engineers in enterprises. Anthropic launched Claude for Small Business with QuickBooks and PayPal integrations. One story is aimed at enterprise deployment; the other is aimed at small-business adoption.

Culture, health, and the boundaries of AI

The Academy ruled AI-generated actors and scripts ineligible for Oscars. Anthropic partnered with the Gates Foundation on a $200 million global health initiative. Ferrari and IBM used AI to turn casual F1 watchers into superfans.

If you are building a product today, the infrastructure beneath you is consolidating fast: someone wants to own your checkout, your deployment, and your inference.

Stay up to date

Get notified when I publish something new, and unsubscribe at any time.

More news

Today in AI — 25 May 2026

Today's top AI news — curated links and commentary on the stories that matter for product builders.

The machines found the bugs — now nobody can fix them fast enough

Three stories from the past 24 hours point to the same uncomfortable truth: AI can now find software vulnerabilities far faster than humans can patch them, and the gap is widening on both sides of the line. Anthropic's Project Glasswing update reports over 10,000 critical flaws discovered in a single month — with open-source maintainers begging Anthropic to slow down disclosures because they cannot keep up. Meanwhile, Google confirms the first documented case of criminals using AI to build a zero-day exploit, and Aikido's research reveals that Google's own credential revocation takes up to 23 minutes — a window Google has declined to close. The security equilibrium that held when humans found bugs at human speed is breaking apart.

The cart is the product now

Google launched a Universal Cart for cross-merchant AI shopping, Klarna built a Shopping Search app inside ChatGPT, and a Fast Company investigation found Google, OpenAI, and others racing to build agentic commerce. The AI industry's newest commerce prize is the shopping cart itself: the place where items get added, compared, and purchased. The platform that controls the transaction controls the game that actually pays.