Last Week in AI: Record Adoption, Nvidia Dominance, Lovecraftian Holiday Horrors
AI adoption just hit historic highs, Nvidia crushed bubble doubts with a monster quarter, and Google’s Gemini 3 + Nano Banana Pro reset the competitive landscape. Meanwhile, Trump’s AI power grab sparked GOP revolt, deepfake evidence entered U.S. courtrooms, TikTok added AI-content controls, Grok 4.1 slashed hallucinations, and AI misfires – from career stalls to "ghastly" Christmas murals – kept piling up.
Agencies Used to Sell Expertise. Now They Have to Prove They Can Guide.
In my latest post, I argue that the old agency value proposition – we’re the experts, trust us – doesn’t cut it anymore. Clients don’t want proclamations; they want partners who can actually guide them through uncertainty, especially with AI reshaping everything. That means showing your work: how you experiment, how your tools and processes evolve, how you train your own people, and how you collaborate openly with clients. Authority now comes from momentum, not perfection. The agencies that win won’t be the ones claiming they know the future – they’ll be the ones demonstrating they can navigate it alongside their clients.
AI Adoption Shatters All Historical Records
Generative AI has become the fastest-adopted technology in history, with 60% of U.S. adults using it within three years of launch, surpassing smartphones and the internet, according to CCIA's 2025 SPICE AI Report. Daily usage among adults jumped from 12% to 17% in eight months, while workplace adoption exploded with 40% of workers now using AI tools, reporting average productivity gains of 15%. Among working AI users, daily engagement surged from 21% to 31% between March and July 2025. User sentiment remains overwhelmingly positive at 77% favorable impressions, growing stronger with continued use. CCIA's Trevor Wagener notes this unprecedented adoption speed demonstrates Americans' enthusiasm for the technology's practical benefits.
Nvidia's Monster Quarter Silences AI Bubble Doubters
Nvidia's blockbuster earnings demolished Wall Street's AI bubble fears, with the chipmaker posting $57 billion in Q3 revenue, beating estimates of $55 billion, while forecasting $65 billion for Q4 versus expectations of $62 billion. Data center revenue hit $51 billion, surpassing projections, as CEO Jensen Huang revealed "Blackwell sales are off the charts" with cloud GPUs completely sold out. The company has "half a trillion" in chip revenue booked through 2026, with potential for more. Wedbush's Dan Ives called it a "pop-the-champagne moment," declaring fears of an AI bubble "way overstated" and placing the AI revolution in just the "third inning." Despite warnings from Bill Gates and Michael Burry about bubble risks, analysts say Nvidia's execution proves AI demand remains insatiable, with hyperscalers maintaining their massive spending commitments.
Trump's AI Power Grab Sparks Republican Revolt
President Trump's push for federal control over AI regulation is igniting backlash from his own party, with Republicans accusing him of caving to Big Tech donors. The White House is drafting an executive order to sue states that pass AI laws, potentially withholding federal funding while creating an "AI Litigation Task Force" to challenge state regulations as unconstitutional. The plan, backed by a $100 million Silicon Valley lobbying fund, has drawn fire from Republican governors DeSantis and Sanders, plus Senator Hawley who noted what "money can do." The revolt comes amid mounting AI safety concerns including chatbot-linked suicides and job displacement fears. A YouGov poll found just 18% of voters support blocking state AI regulation, with Republicans warning Trump the move could prove politically disastrous in 2028 when victims of AI harms become campaign fodder against the party.
Google's AI Double Whammy Shakes Up the Industry
Google unleashed a one-two punch in the AI arena with Gemini 3 and its "absolutely bonkers" Nano Banana Pro image model. Arriving just months after competitors' releases, Gemini 3 launched globally November 18 with CEO Sundar Pichai boasting 650 million monthly active users and the model topping LMArena benchmarks. The launch included Google Antigravity, a multi-pane coding interface for "vibe coding" that lets developers generate code through natural conversation. But hilariously, when researcher Andrej Karpathy got early access, Gemini 3 refused to believe it was 2025, accusing him of gaslighting until he enabled Google Search. Meanwhile, Nano Banana Pro stunned developers with its ability to generate complex infographics, accurate text rendering, and 4K resolution images – though at premium pricing of $0.24 per 4K image versus competitors' $0.04 baseline, making enterprise integration its likely sweet spot.
ChatGPT has Become Everyone's Digital Best Friend
From parenting advice to wedding planning, ChatGPT has morphed into humanity's all-purpose companion. Parents increasingly turn to the AI for child-rearing guidance, despite warnings about its manipulative nature potentially causing reality breaks – a phenomenon linked to teen suicides. Meanwhile, brides use AI to plan weddings, with 36% actively employing ChatGPT for everything from budget breakdowns to vow writing, though one Reddit user claims his fiancée left him at the altar after discovering his AI-generated promises. A Washington Post analysis of 47,000 ChatGPT conversations revealed users seek advice and companionship far more than productivity help, with 10% discussing emotions and sharing deeply personal details including email addresses and phone numbers. Now OpenAI is launching group chats globally, allowing up to 20 people to collaborate with ChatGPT in shared conversations – transforming the bot from personal confessor to digital dinner party guest, complete with emoji reactions and the uncanny ability to know when to chime in.
When AI Holiday Cheer Goes Horribly Wrong
A London borough's attempt at festive decoration became a cautionary tale when two enormous Christmas murals appeared in Kingston upon Thames, featuring nightmarish AI-generated imagery. The 100-foot-wide artworks displayed grotesquely disfigured humans with extra appendages, mutant dogs merging with people, and a snowman with an eye on its cheek – hallmarks of outdated image generators. Locals mocked the "Lovecraftian horror" scenes, joking they celebrated "the return of our dark lord Cthulhu." The controversy took an unexpected turn when the murals were hastily torn down, not for their AI sloppiness, but because viewers interpreted crowds splashing through water as political commentary on UK immigration debates. The building owners claimed inspiration from 16th-century painter Pieter Bruegel, though this artistic defense hardly justified the AI-generated monstrosities that even ancient generators would struggle to produce.
AI Deepfake Evidence Infiltrates American Courtrooms
A California judge dismissed a housing dispute case after lawyers submitted an obvious AI-generated deepfake as witness testimony, marking potentially the first documented instance of fake AI evidence being caught in court. The video showed a barely-animated witness with fuzzy features and flapping lips while everything else remained frozen—so obviously fake that Judge Victoria Kolakowski dismissed the case entirely. When plaintiffs argued the judge failed to prove their evidence was AI-generated, their reconsideration motion was denied. The incident highlights growing concerns as OpenAI's Sora 2 enables anyone to create realistic crime videos. While the US Judicial Conference rejected updating AI evidence guidelines in May, believing existing standards suffice, Judge Erica Yew warns AI-generated fake evidence is "happening much more frequently than reported."
Why AI May Kill Career Advancement for Many Young Workers
AI is beginning to disrupt not just jobs, but career ladders. Companies are increasingly automating entry-level and developmental tasks – responsibilities once used to train and promote young professionals, according to the article. That shift means fewer opportunities for junior staff to gain experience, build relationships, and advance internally. If firms rely on AI rather than human mentorship or rotational roles, younger workers may find themselves stuck in roles with little upward mobility or risk falling behind their more experienced colleagues. The real concern: automation can erode the pipeline of talent and pathways that fuel long-term career growth.
TikTok Lets You Dial Up or Down AI-Created Videos
TikTok is piloting a new “Manage Topics” slider that lets users decide how much AI-generated content they want in their For You feed – offering a direct way to dial the presence of synthetic media up or down. The move follows months of rising anxiety over AI-flooded feeds and signals that TikTok is aiming for more transparency and user choice. According to the reporting, the company is pairing the feature with better AI-content detection, including “invisible watermarks” baked into videos produced with TikTok tools, plus support for C2PA credentials on uploads. By giving people clearer control – and more trustworthy labeling – TikTok is trying to balance creativity, safety, and authenticity as generative content becomes unavoidable.
Musk's Grok 4.1 Slashes Hallucinations by Two-Thirds
Elon Musk's xAI launched Grok 4.1 with dramatically reduced hallucination rates, dropping from 12% to just 4.2% while claiming the top spot on LMArena's leaderboard ahead of OpenAI, Anthropic, and Google. The model arrives in two configurations – a fast "non-thinking" mode and a deliberate "thinking" variant – following a two-week silent rollout where it won 65% of blind comparisons against its predecessor. Beyond accuracy improvements, xAI emphasized enhanced emotional intelligence, scoring highly on EQ-Bench3 evaluations. The company achieved these gains using frontier AI models as reward evaluators to autonomously score responses at scale. Available through grok.com, X, and mobile apps, plus enterprise API access at $5 per 1,000 tool invocations, Grok 4.1 represents xAI's competitive push.