The Signal — May 28, 2026
The AI industry's acceleration showed no signs of slowing this week, from a consulting-firm landgrab putting frontier models in front of every Fortune 500 boardroom to a state legislature drawing the sharpest regulatory lines in America.
DeepMind CEO Hassabis Says Humanity in "Foothills of the Singularity," Predicts AGI by 2030
In a post-Google I/O interview with Axios published May 26, Google DeepMind CEO Demis Hassabis offered his most striking prediction yet: humanity is standing in the "foothills of the singularity," and AGI could arrive by 2029 or 2030 (possibly sooner).
What made the claim noteworthy wasn't just the timeline but the test Hassabis proposed to validate it. Rather than relying on existing benchmarks, he suggested training an AI system exclusively on pre-Einstein-era scientific knowledge and checking whether it could independently rediscover breakthroughs like general relativity. The idea reframes AGI evaluation around genuine novel reasoning rather than pattern-matching against data the model has already seen.
Hassabis has consistently moved his own goalposts forward over the past two years, but "foothills of the singularity" marked a rhetorical leap: the first time a major lab leader publicly invoked the singularity as a near-term framing rather than a distant abstraction. Whether the prediction holds or not, the Overton window for AGI discourse shifted.
Sources: Axios · Sherwood News · Business Insider
Three of Big Four Consulting Firms Standardize on Anthropic's Claude
Three of the Big Four consulting firms — Deloitte, PwC, and KPMG — have now standardized on Anthropic's Claude as their primary AI platform. KPMG's announcement, covering its 276,000 employees globally, completed a trifecta that also includes Deloitte's 470,000-person rollout and PwC's deployment across 30,000 US professionals.
Combined, roughly 1.1 million consultants will have Claude access by September 2026. The Wall Street Journal reported that KPMG plans to integrate Claude into its global tax advisory platforms, the kind of high-stakes, compliance-heavy work that serves as a proof point for enterprise AI adoption.
The strategic implications extend well beyond the firms themselves. As Deeper Insights noted, these consultancies advise virtually every Fortune 500 company. When a Big Four partner recommends an AI strategy, the platform they've standardized on carries an implicit endorsement. Anthropic effectively gained a distribution channel into the C-suite of global enterprise without a traditional sales motion, creating a competitive moat that will be difficult for OpenAI or Google to replicate through technology alone.
Sources: Deeper Insights · Fortune · Wall Street Journal
Illinois Passes America's Strongest AI Safety Bill Package
The Illinois Senate passed what Wired called "America's strongest AI safety bill," a comprehensive legislative package that goes further than any prior state-level AI regulation. The bills impose risk assessment and third-party audit requirements on frontier AI developers, mandate consumer disclosures for AI-generated content, restrict biometric data collection in schools, and introduce protections around AI companion applications.
The package emerged from a bipartisan effort led by Senator Sue Rezin, who framed the legislation as balancing innovation with accountability. Unlike California's vetoed SB 1047, which targeted a single dimension of AI risk, the Illinois approach addresses multiple vectors simultaneously, covering developer obligations, consumer rights, child safety, and school data privacy.
The bills still require the governor's signature, but their passage through the Senate signaled that state-level AI regulation has moved past the question of "whether" and into "how." For AI companies, Illinois just became the compliance benchmark to watch.
Sources: Wired · Senator Sue Rezin (Official) · Transparency Coalition AI
On the Editor's Desk
Several stories came through the pipeline but didn't make the cut. An EU Parliament AI framework story turned out to be misleadingly framed; it was actually an AI Act simplification measure, not the new liability law multiple outlets implied. A Bloomberg/BLS piece on AI job displacement data traced back to a May 15 core article, making it too stale for inclusion. A Canadian privacy ruling against ChatGPT was 22 days old by publication time.
These three stories earned their spots by combining recency (all from May 26), significance, and complementary thematic coverage: a capabilities prediction from the field's most credentialed researcher, an enterprise distribution shift with structural market implications, and the most consequential piece of US AI legislation to date. Together they capture the week's acceleration across technology and policy, business strategy and regulatory momentum.
