- Product & Design Pulse
- Posts
- Product & Design Pulse v88
Product & Design Pulse v88
The Leverage Game ♟️
Malcom Jones
April 20, 2026
Welcome to this week’s edition of Product & Design Pulse, where we explore the latest in tech, product, design, and innovation! Last week was about leverage — who has it, who's losing it, and what it costs. Anthropic's Glasswing gambit bore immediate fruit as CEO Dario Amodei sat down with White House Chief of Staff Susie Wiles to discuss Mythos access, forcing the administration to negotiate even while the Pentagon's blacklist remains in place — a striking demonstration that building something too capable to ignore is its own form of political power. Meta, meanwhile, set May 20 for the first wave of 8,000 layoffs in what will be its largest restructuring since 2022, even as the company posted $60 billion in profit last year — confirming that AI-era headcount reductions are driven by reallocation, not distress. Ben Thompson tied the threads together by arguing that compute scarcity, not safety, is the real reason Anthropic is withholding Mythos from the public, and that the defining strategic question is no longer who builds the best model but who converts limited compute into the highest-value output. On the hardware side, Apple revealed four smart glasses prototypes that skip displays entirely in favor of camera-first AI wearability, while the U.S. government announced it will require data centers to disclose energy consumption for the first time — a regulatory milestone as the industry commits to multi-gigawatt expansions. The week's underlying message: in a world where compute, talent, and political capital are all finite, the winners are the companies that know exactly where to spend each one.
🎧 Audio Overview [BETA]
For those who don’t have time to read 😁 |
Last week…
Meta Sets May 20 for First Wave of 8,000 Layoffs, With More Cuts Planned
Reuters reports that Meta will lay off approximately 10% of its global workforce — close to 8,000 employees — on May 20, with additional rounds of cuts planned for the second half of the year as executives assess how AI capabilities continue to reshape internal workflows. The company is in a fundamentally different financial position than during its 2022–23 "year of efficiency" restructuring: it generated over $200 billion in revenue and $60 billion in profit last year, but is spending $115–$135 billion on AI infrastructure in 2026 and envisions fewer management layers as AI-assisted workers take over more functions. For the industry, this is the clearest signal yet that AI-driven workforce restructuring isn't a one-time event but a continuous process — and that profitability and layoffs are no longer contradictory when the goal is reallocating human capital toward compute.
Anthropic's Amodei Holds White House Talks on Mythos Despite Pentagon Blacklist
Axios reports that Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles — with Treasury Secretary Scott Bessent also present — to discuss government access to the Mythos model, even as Anthropic remains locked in litigation with the Pentagon over its supply chain risk designation. Multiple civilian agencies including Treasury, Energy, and CISA want Mythos for defensive cybersecurity work, and administration officials privately concede Anthropic's tools are "best-in-class" for national security despite some in Trumpworld viewing the company as ideologically hostile. For the AI industry, this is Anthropic's Glasswing strategy paying immediate geopolitical dividends: by making Mythos too capable to ignore, the company is forcing the government to negotiate on its terms — even while it refuses to grant the military unrestricted use of its models.
Did Neuralink Make the Wrong Bet on Brain-Computer Interfaces?
The Verge argues that Neuralink's brain-computer interfaces have so far only achieved brain-to-cursor control — a meaningful accessibility advance for paralyzed patients, but fundamentally distant from Elon Musk's promises of AI-merged superintelligence and superhuman cognitive abilities. Competitors like Synchron and Blackrock Neurotech are pursuing less invasive approaches with narrower but more achievable goals, while the scientific community has grown increasingly skeptical that Neuralink's current architecture can scale past basic input control given fundamental biological constraints like immune response, signal degradation, and brain plasticity. For product leaders watching the hardware frontier, this is a useful reminder that the gap between a compelling demo and a viable platform remains enormous when biology — not software — is the constraint.
U.S. Government to Require Data Centers to Disclose Energy Consumption
The Energy Information Administration told Senators Josh Hawley and Elizabeth Warren that it will launch a mandatory nationwide survey requiring data centers to disclose their energy consumption, including annual electricity use, behind-the-meter generation, cooling systems, square footage, and IT efficiency metrics — the first time the federal government will compel the industry to report this data. The EIA is already running pilot surveys in Texas, Washington state, and Northern Virginia, with a second tranche covering three additional states expected to finish by late September. For the AI infrastructure buildout, this marks a turning point in transparency: as companies like Anthropic, Microsoft, and Meta commit to multi-gigawatt compute expansions, the government is finally moving to measure what the industry has long treated as proprietary — and the results could reshape permitting, energy policy, and the economics of AI scaling.
Apple Tests Four Smart Glasses Designs as AI Chief Giannandrea Departs
Bloomberg's Mark Gurman reports that Apple is testing four distinct frame styles — including rectangular and oval designs in multiple colors — for its first smart glasses, codenamed N50, which skip displays entirely and instead use vertically oriented oval camera lenses to feed visual data to Siri and Apple Intelligence, with a possible late-2026 unveiling and 2027 retail launch. The glasses are part of a three-device AI wearable strategy alongside AirPods and a camera pendant, and Gurman also confirmed that former AI chief John Giannandrea is leaving Apple as the company pivots its AI approach. For product strategists, Apple's choice to bet on camera-first wearability over immersive AR is a deliberate rejection of the Vision Pro model — prioritizing the kind of everyday utility and distribution advantage that has defined every successful Apple product category.
Ben Thompson: Compute Scarcity, Not Safety, Explains Why Anthropic Is Withholding Mythos
Thompson argues that Anthropic's decision not to release Mythos broadly isn't primarily about safety — it's about opportunity cost: the company is already compute-constrained serving current Claude users, and directing scarce capacity to high-paying Glasswing partners generates far more value than offering Mythos to subscription plans that don't charge per-token. He frames this as the key strategic question of the AI era: in a world where compute is genuinely scarce, controlling demand — through the most compelling products — gives power over supply, which is how Aggregation Theory adapts to a constrained-resource environment. The essay also examines Meta's Muse Spark as evidence that the AI race is no longer about open vs. closed but about who can convert limited compute into the highest-value output — and suggests the biggest loser may be the "serially unfocused" OpenAI, which spreads its compute across too many products to dominate any single one.
