Quill the owl mascot waving Nitido AI

Thursday, August 6, 2026

Today's stack is a bit of a shake up: Google's most legendary engineer is walking out the door, DeepSeek is quietly ditching its cheap price act, and Meta has some very uncomfortable ad moderation news to answer for.

Quill the owl illustration for today's headline

Jeff Dean, one of Google's most famous engineers, is leaving to start his own AI company

Jeff Dean has been one of the most respected names inside Google for two decades, helping build much of the technology behind its search engine and its AI systems. Now he is leaving, along with several other senior Google AI executives, to launch a new company called Discovery Loop. The plan is to use AI to speed up scientific breakthroughs, in areas like drug discovery and computer chip design. It is a striking loss of talent for Google right as it races against OpenAI and Anthropic to stay ahead in AI.

What this means for you: When top AI talent walks out the door to start something new, it is often an early signal of where the next big breakthroughs will come from. Worth keeping on your radar.

What this means for your business: Talent exits like this can reshape where cutting edge AI research shows up first. Watch for new partnership or licensing opportunities as spinout labs like this one get going.

Source: Wired AI

Radar 01

Google reshuffles its top AI leadership as Dean heads out the door

In a move that landed the same week, Google announced a major leadership change of its own. Demis Hassabis, who has run Google DeepMind, is stepping up to become chairman of DeepMind and chief scientist for all of Alphabet. He will keep leading Isomorphic Labs, Google's drug discovery AI unit. The changes concentrate decision making at the top as Google tries to keep pace with rivals.

What this means for you: Leadership changes at big AI labs often come with shifts in strategy and priorities, especially useful to track if you rely on their tools day to day.

What this means for your business: If your company builds on Google's AI products, watch closely for roadmap changes under this new leadership structure over the coming months.

Source: The Verge

Radar 02

DeepSeek, the cheap AI disruptor, is about to get more expensive

DeepSeek, the Chinese AI company that made its name by undercutting US rivals on price, now plans a significant price increase across its AI services. This is a notable shift for a company that built its reputation on being the budget option, and it may ease some of the pricing pressure DeepSeek has put on companies like OpenAI and Anthropic.

What this means for you: If you use a budget AI tool because it is cheap, do not assume that low price will stick around forever.

What this means for your business: Build room for vendor price hikes into your AI budgeting, even for providers whose whole pitch has been low cost.

Source: Bloomberg Technology

Radar 03

Meta ran ads containing AI-generated child abuse imagery

Wired reports that Meta ran more than 50 ads containing AI-generated child sexual abuse imagery across Facebook, Instagram, Messenger, and Threads, with some running as recently as this week. The ads slipped past Meta's own moderation systems, raising serious questions about how well AI-generated harmful content can be caught before it spreads to millions of people.

What this means for you: AI-generated harmful content is getting better at slipping past filters on major platforms, worth keeping in mind whatever platform you spend time on.

What this means for your business: If your brand advertises on these platforms, this is a fresh reminder to scrutinize ad placement safeguards and hold platforms accountable for what runs next to your name.

Source: Wired AI

Try This Today

Before you sign your next AI vendor contract, ask directly whether pricing is locked in or could change, and for how long. DeepSeek's surprise price hike is a good reminder that today's cheap AI deal is not guaranteed to stay cheap.

Quick Hits

  • AMD's data center revenue more than doubled year over year to $6.7 billion, driven by surging demand for AI computing power, while its gaming business took a back seat. [1]
  • Anthropic is building an in-house team to design its own custom AI chips, aiming to co-design hardware and models so Claude runs faster and cheaper. [2]
  • Researchers built MatrAIx, a system that simulates 8.3 billion AI personas to test how real people might react to new AI products and features, without running a costly human study every time. [3]

See all editions