Topic · 30 episodes
Technology
Technology in 2025–2026 is defined by three colliding forces: governments racing to control AI before legislation catches up, companies deploying models that outpace their own safety frameworks, and hardware shifting AI from cloud to pocket. Anthropic's Mythos cracked classified NSA systems and faced the first-ever model-specific export ban. Apple rebuilt Siri on Google Gemini. Meta priced AI glasses at $299. Meanwhile, Rogers documented across 6,000 studies that adoption S-curves only become legible in hindsight — a warning every lab is ignoring.
Frequently asked
Why did the US government ban Anthropic's Mythos AI model?
The US Commerce Department issued export controls on Anthropic's Mythos and Fable 5 after Mythos penetrated nearly all classified US systems during an authorized NSA red-team exercise in hours. The controls restrict both models to US citizens only. Anthropic received just ninety minutes' notice — setting the first-ever export ban on a specific AI model.
Did Apple rebuild Siri with Google Gemini?
Yes. Apple's rebuilt Siri, launched at WWDC 2026 with iOS 27, was completed in 2024, rejected, delayed 18 months, and rebuilt with Google Gemini handling heavy generative tasks. Apple has not disclosed which data types trigger the Gemini handoff, creating direct tension with its privacy-first positioning.
How cheap are Meta's new AI smart glasses?
Meta and EssilorLuxottica launched three displayless AI smart glasses models starting at $299 on June 23, 2026 — below the Ray-Ban Meta price of $359. Sold at Amazon, Best Buy, and LensCrafters, they mark the first time Meta has put its own brand directly on consumer hardware.
Why is Meta the last major AI company not signing the US federal AI security framework?
The Trump administration privately pressured Meta to submit its Muse Spark model to federal security review under CASI. OpenAI, Anthropic, Google, xAI, and Microsoft had already signed the voluntary framework. Meta's spokesman said the company 'hopes to sign soon,' leaving it the sole frontier AI holdout.
Why did John Jumper leave Google DeepMind for Anthropic?
John Jumper, who shared the 2024 Nobel Prize in Chemistry for co-building AlphaFold — which predicted over 200 million protein structures — left Google DeepMind for Anthropic. Alphabet's stock dropped 6% the same week, coinciding with Noam Shazeer's departure to OpenAI.
Episodes
How quantum tunneling and heat density are slowing silicon scalingIBM has built a 0.7nm prototype chip packing 100 billion transistors — more than ten times the neurons in a human brain — yet cost per transistor is rising, not falling. Quantum tunneling and heat density impose hard physical limits, and no single successor metric has replaced Moore's Law as the industry's shared north star.
OpenAI built its own AI chip and halved inference costs — here's why Nvidia should careOpenAI's first custom AI chip, Jalapeño, was co-developed with Broadcom and unveiled on June 24th. Broadcom's CEO claims it cuts inference costs by 50% versus Nvidia GPUs, but engineering samples only just arrived that day — broad deployment isn't expected until 2027 or 2028, and no independent verification of the cost figure exists yet.
Google just cut Meta's Gemini access — compute shortages are forcing tech giants to ration AIIn March 2026, Google restricted Meta's access to Gemini compute through Google Cloud — Meta was the most heavily affected enterprise customer. The shortage forced Meta to ration AI tokens internally, slowing pipelines for content moderation, scam detection, advertising tools, and coding assistance. Google holds infrastructure supply rivals cannot simply buy their way into.
Anthropic is now developing its own drugs, armed with 60+ scientific databases and ClaudeAnthropic launched Claude Science in beta on June 30, connecting Claude Opus 4.8 to 60-plus scientific databases including UniProt, PDB, and ChEMBL. In one benchmark experiment, a single researcher surfaced 864 missing disease relationships out of 915 that official ontologies lacked — across 490 papers on zoonotic spillover — for $26.
Anthropic just released Claude Sonnet 5 at a steep discount to compete in agentic AIAnthropic launched Claude Sonnet 5 on June 30 at introductory pricing — including $10 per million output tokens — that expires August 31, after which output rises to $15. A simultaneous tokenizer change inflates token counts up to 1.35×, meaning some developers could end up paying more than on Sonnet 4.6 even before the introductory period ends.
The mechanism behind slow starts, explosive middles, and plateaus in tech adoptionTech adoption follows an S-curve: slow start, explosive middle, then plateau. Frank Bass's 1969 model shows the social imitation coefficient (q) dominates independent curiosity (p) across nearly every technology studied — yet ChatGPT still hit 100 million users in three months, a slope no forecast predicted.
Apple just rebuilt Siri with Google Gemini — your iPhone AI is about to changeApple's iOS 27 integrates Google Gemini as the engine behind Siri, under a multi-year licensing deal reportedly worth roughly $1 billion per year. On the same day, Apple shipped an Extensions framework letting users swap Gemini out for Claude or ChatGPT — a built-in kill switch that effectively undermines Gemini's exclusivity from launch.
Anthropic just cut Claude access in half for California government workers — here's the dealCalifornia gave all state agencies access to Anthropic's Claude AI at 50% off, announced June 29 with Governor Newsom's name on it — despite an active federal supply-chain risk designation on Anthropic. The California DMV was already quietly running Claude for customer service before the announcement legitimized it statewide.
Google's AI brain drain worsens — Anthropic now poaching top Gemini talent directlyFour senior Google DeepMind researchers — including 2024 Nobel laureate John Jumper — left for Anthropic and OpenAI within six days in late June 2026, amid a DeepMind coding pivot that cut frontier research mandates. SignalFire data shows AI talent flows to Anthropic eleven times more often than the reverse.
OpenAI, Anthropic hit pause on new models — Trump admin cybersecurity reviews block launchesExecutive Order 14409, signed June 2, 2026, created a nominally voluntary 30-day cybersecurity review — but Anthropic's Fable 5 and Mythos 5 were forcibly taken offline, and OpenAI's GPT-5.6 Sol was access-gated, while Google's Gemini 3.5 Pro faced no federal review at all, exposing the framework's unwritten and unevenly applied criteria.
The White House is pushing Meta to submit to formal AI reviews as security concerns over AI deployment intensifyThe Trump administration privately pressured Meta to submit its Muse Spark AI model to federal security review under CASI, reported by the Times on June 23rd. OpenAI, Anthropic, Google, xAI, and Microsoft had already signed. Meta's spokesman said the company 'hopes to sign soon' — making Meta the sole frontier AI holdout under a voluntary framework built without legislation.
Apple shipped on-device AI scaling to 70B models—shifting power away from cloud providers and eliminating privacy risksApple's Core AI framework, announced at WWDC 2026, supports on-device large language models from 3B to 70B parameters on Apple Silicon, shipping Qwen and Mistral as pre-converted Swift packages. Developers can eliminate cloud API costs estimated at $3,000–$5,000 per year — but Apple's three-tier routing to Google Gemini servers remains mechanistically opaque.
Anthropic's Claude Tag replaces static Slack bots with always-on AI that learns, monitors, and acts without promptsAnthropic launched Claude Tag on June 23, 2026, replacing its Slack app with an ambient AI that joins workspaces as a user account, proactively participates in conversations without prompts, and can write code and retrieve data — while live federal litigation over AI workplace monitoring has yet to validate the consent model it depends on.
Meta just dropped a $299 AI smart glasses line—shifting wearable AI from luxury to mainstreamOn June 23, 2026, Meta and EssilorLuxottica launched Meta Glasses — three displayless AI smart glasses models starting at $299, sold at Amazon, Best Buy, and LensCrafters. The move drops AI wearables below the Ray-Ban Meta price of $359 and marks the first time Meta has put its own brand directly on consumer hardware.
The mechanism behind slow early adoption, rapid growth, then saturationTechnology adoption S-curves are descriptive, not predictive. Everett Rogers synthesized 6,000 studies to establish the pattern — slow start, explosive growth, saturation — but the curve only becomes visible in hindsight. Christensen showed industry insiders with real-time data still couldn't locate themselves on the curve, and the empirical foundation excludes every technology that failed.
Apple unveils a major Siri AI makeover as Tim Cook departs leadershipApple's Siri AI, launched at WWDC 2026 with iOS 27, was completed in 2024, rejected, delayed 18 months, and rebuilt with Google Gemini handling heavy generative tasks — while Apple has not disclosed which data types trigger the Gemini handoff, creating a direct tension with Apple's privacy-first positioning.
Anthropic's Mythos cracked classified NSA systems in hours, forcing a rushed government export banAnthropic's Mythos AI penetrated nearly all classified U.S. systems during an authorized NSA red-team exercise in hours — not a breach, but a controlled capability test. The Commerce Department issued export controls restricting Mythos and Fable 5 to U.S. citizens only, with just ninety minutes' notice to Anthropic, setting the first-ever export ban on a specific AI model.
John Jumper, Nobel laureate, just left Google DeepMind to join Anthropic—deepening the talent drainJohn Jumper, who shared the 2024 Nobel Prize in Chemistry with Demis Hassabis for co-building AlphaFold — a system that predicted over 200 million protein structures — has left Google DeepMind for Anthropic. Alphabet's stock dropped 6% the same week, alongside Noam Shazeer's departure to OpenAI.
The structural reason new technologies hit slow, then fast, then slow again
Apple rebuilt Siri on Google Gemini and Nvidia chips—is that independence or compromise?
iOS 27 beta shows Apple built Claude, Gemini, ChatGPT choice into Siri—but hid it at WWDC
Anthropic pledges $150M to nonprofits while Trump admin forces model withdrawal
Tech promised everyone would use AI for everything — but many are avoiding it entirely
The quantum tunneling wall — why chip scaling hits physics, not just engineering
Meta hired Alexandr Wang to build a new AI model—now Zuckerberg has to sell it
Anthropic just shut down access to Mythos after a US government order—what it means for AI competition
Trump administration orders Anthropic to block foreign access to its most advanced AI models
After a year and $14.3 billion investment, Meta's AI Chief faces the reality that his Muse Spark model lags competitors
Anthropic's own safety cautions on Fable 5 may have triggered the government to shut it down entirely
The US just disabled Anthropic's Fable 5 and Mythos 5 citing national security concerns