Justine Moore

Consumer AI ended 2025 looking winner-take-most: ChatGPT held 800–900 million weekly active users, while only 9% of consumers paid for more than one of ChatGPT, Gemini, Claude, and Cursor. For most of the year, fewer than 10% of ChatGPT users visited another major provider. Olivia also cited Gemini as having added an estimated 35% of its scale on the web and about 40% on mobile, while Claude, Grok, and Perplexity each sat near 8–10%. Anish Acharya’s brand framing was simple: “ChatGPT is like the Kleenex of AI.”

Gemini was the live threat because viral creative models coincided with accelerating growth: desktop users rose 155% year over year versus ChatGPT’s 23%. Gemini reached roughly half of ChatGPT’s mobile scale on Android but only 17% on iOS—“everywhere” yet still “nowhere” in consumer habit. Justine Moore thinks it could get there if it sustains its image-and-video launches, though ChatGPT’s guided templates make the first creation far easier than Gemini’s blank box.

The year’s consumer model breakthrough was image and video models combining realism, reasoning, retrieval, and multiple media. ChatGPT 4.0 image’s Ghibli moment, Sora 2, Veo 3, and Nano Banana showed that accurate details, search-backed logos, consistent characters, and audio combined with video can create viral demand. The next architecture is “anything in to anything out,” potentially merging text intelligence, images, video, and editing into one model.

The labs’ distribution does not automatically produce successful vertical products, creating the panel’s clearest startup opening for 2026. Pulse, Atlas, group chats, Sora, Stitch, Gems, and Opal have not become breakout standalone consumer interfaces; NotebookLM was the notable exception. Bryan Kim’s caveat is that high-frequency assistants will remain hard to displace wherever the product is primarily text in and text out.

Sora 2 proved demand for AI video creation, not yet for an AI-native social network. A small creator cohort generated content for TikTok, Instagram, X, and Reddit, while in-app consumption, remixing, and commenting did not seem as strong as initially; the better analogy was “CapCut,” not TikTok. Bryan’s bull case is that humor could create a new status game through prompting skill and cultural awareness. Anish asked whether exporting still makes TikTok with Sora videos “strictly better.”

The most defensible near-term market may be prosumer and enterprise workflows, where depth of usage can invert traditional consumer economics. ChatGPT enterprise usage was said to be up roughly 8–9x year over year, while Claude and Comet showed the value of persistent workflows and cross-tool context. Usage charges above subscriptions are already producing consumer AI products with more than 100% revenue retention: “Maybe all of AI is actually a power user story.”

Compute remains the strategic constraint: labs must trade training against inference and entertainment traffic against coding intelligence, while focused application companies avoid that internal conflict. Anish said xAI was “probably the only” model company not bottlenecked on compute, “from my understanding,” while first-party-only labs also leave room for multi-model products serving power users. With model quality now sufficient to “build a real, scalable app,” the closing hope was that 2026 becomes a huge year for consumer builders.

🔗 Original source & video: Where does consumer AI stand at the end of 2025?

Google’s commercial-search position remains intact for now, while AI is capturing the free queries that support its broader habit loop and could eventually reroute its “tax on GDP.” Alex Rampell sees search volume falling for informational queries while Google’s financials still rise, implying users are moving non-monetizable questions—not monetizable shopping—to ChatGPT. The risk arrives when agents own purchase initiation and the “tax might just shift elsewhere.”

AI’s commerce sweet spot is the broad middle between unplanned impulse buys and large, highly considered purchases that often still demand physical experience or human reassurance. A TikTok-triggered shirt requires no research, while a house, car, or wedding venue is unlikely to become fully agentic. Handbags, detergent, bikes, couches, and laptops offer the stronger opening: better research, continuous price scanning, and eventually execution.

Once a buyer has selected a SKU or UPC, an agent could automate the entire money-versus-time optimization problem. It can search prices, delivery terms, coupons, cashback programs, affiliate rebates, and even the best credit card, then buy when a threshold is met—CamelCamelCamel with the action loop closed. Rampell’s test becomes an “IQ test”: “Do you want to pay less for something or more?”

Attribution—not product discovery—is the load-bearing commercial problem, and AI may make today’s bad incentives worse. Last-click systems already let coupon extensions such as Honey intercept customers at checkout and claim credit for sales they did not cause. Agents could become “the last click of the 21st century,” capturing merchant economics despite being only one influence among Reddit, advertising, creators, stores, and prior brand affinity.

The web’s degraded information supply is a structural constraint on AI shopping, because models cannot turn affiliate-optimized inputs into objective advice. Search spans walled gardens while the open web is saturated with “SEO-optimized crap”; Amazon listings and reviews are similarly gameable. The unsolved question is stark: “You can’t turn shill junk into honest analysis,” so how do platforms “decrapify” the corpus?

AI likely strengthens aggregators while exposing undifferentiated direct-to-consumer brands whose products are made elsewhere and whose traffic must be bought. Commodity sellers such as mattress brands can be copied by the same OEMs and must repeatedly acquire customers from Google or Facebook; fashion brands also cannot own every trend. Moore argues that agents can direct buyers once demand starts there, but she and Rampell note that AI may struggle to inculcate demand before culture makes an item desirable.

The durable opportunities sit in trusted curation, specialized buying agents, and merchant infrastructure—not merely another horizontal chatbot. Costco’s membership-funded refusal to sell bad products makes it unusually “AI-proof,” while startups can build domain experts, agent-readable storefronts, and payment infrastructure for delegated purchasing. Amazon’s high-margin advertising is exposed if AI takes control of the presentation layer before shoppers reach Amazon.

🔗 Original source & video: The Death of Search: How Shopping Will Work In The Age of AI

Consumer AI is stabilizing without becoming static: only 11 of 50 web properties were new, versus 17 six months earlier. Olivia Moore’s fifth semiannual ranking measures global monthly web visits and mobile active users—“usage, not revenue”—revealing durable attention beneath the launch-cycle noise.

Vibe coding has become a credible consumer-prosumer category, with Lovable and Replit reaching the main web list while Bolt sits just below it. Lovable announced $100 million in ARR, and many leading platforms showed “100% or above” revenue retention during their first three months before potentially flattening below 100%, suggesting users may upgrade as projects become useful rather than merely sampling trials.

Google placed four distinct properties on the web list, showing that distribution and product segmentation can challenge ChatGPT without displacing it outright. Gemini ranked No. 2 with roughly 10% of ChatGPT’s web traffic but about half its mobile traffic; AI Studio reached the top 10, NotebookLM ranked No. 13, and Google Labs—likely propelled by Veo 3—ranked No. 39.

China’s AI ecosystem now supplies domestic champions, export-oriented products, and globally used agents simultaneously. Qwen, Doubao, and Kimi each ranked in the web top 20, while Chinese image and video models often reach U.S. users through intermediaries; Manus, meanwhile, announced a $90 million annualized run rate with Brazil as its largest traffic source and the U.S. second.

Companionship remains one of consumer AI’s hardest categories to displace on mobile. JuicyChat, Joyland, and DreamGF joined repeat entrants including Character.AI, Janitor AI, SpicyChat, PolyBuzz, CrushOn.AI, A-Dot, and Candy AI—a dense installed base that Justine said would be difficult to displace.

Model ownership is not the only route to durability: more than half of the long-running “all-stars” host or use other companies’ models, or act as model aggregators. Durability increasingly comes from the interface, community, reusable assets, and bottoms-up enterprise adoption—“the UI and the product experience matter just as much as the model.”

The next usage wave may come from accuracy-sensitive products rather than another creative novelty. Better math, logic, reasoning, and reliability from Grok 4, the new version of Claude, and GPT-5 might tip financial modeling, presentations, health, education, and personal finance from “cool but inaccurate” into habitual use—but the hosts’ overriding forecast remains that consumer breakouts contain “so much randomness.”

🔗 Original source & video: The Top 100 Most Used AI Apps in 2025

Genie 3’s significance extends beyond better video: it is a new kind of model for interactive worlds generated in real time from a few words. Its one-minute spatial memory, frame-to-frame consistency and immediate controls turn passive clips into navigable environments—“there is something magical about the real-time aspect.” It remains a research preview, with broader access intended but no concrete timeline.

The leap came from combining capabilities previously split across Genie 2, Veo 2 and GameNGen. The team pursued the most ambitious intersection—higher resolution, real-time generation and “minute-plus memory” in one model—despite those objectives conflicting. The result arrived after roughly seven months and resonated more strongly than its creators expected.

Spatial persistence was explicitly designed, yet its quality still surprised the researchers who built it. Genie 3 generates frame by frame without an explicit NeRF, Gaussian-splatting or other fixed 3D representation; the team believes that choice is key to generalization. The present design retains this memory for one minute, although Shlomi says there is “no fundamental limitation.”

Scale and training breadth are yielding increasingly credible physical behavior, though the researchers stop short of calling it LLM-style reasoning. In examples, characters typically swim when entering water, skiing speeds up downhill and slows or stops uphill, and an approached door may open; non-experts can mistake some storms, lighting and water for reality. The model finds it harder to obey unlikely prompts while preserving world coherence—“low-probability areas” such as wearing flip-flops in the rain.

Genie 3 and Veo 3 remain separate because interactivity and cinematic generation impose different technical priorities. Genie offers navigation and actions but generally lacks audio; Veo 3 targets a higher visual-quality threshold, while agent training values rapid, egocentric interaction over cinema-grade output. Shlomi frames modality, generation speed and controllability as orthogonal dimensions rather than a single inevitable convergence path.

A promising path is synthetic experience for agents and robotics. Anjney highlighted a possible composition with an agent she thought was called SIMA; Jack says Genie 3 is an environment rather than an agent, so other agents can learn through simulated experience. This could combine real-world data’s realism with simulation’s scale and safety—the “best of both.” Simulation still does not solve actuation, movement decisions or the broader physical-response loop.

The team simultaneously describes Genie 3 as years ahead of prior expectations and far from an accurate world simulator. Jack says minute-long, photorealistic, remembered worlds looked like a five-year goal only two or three years ago; Shlomi cautions that genuinely placing a person or agent into a faithful world requires much more work. Calendar forecasts remain deliberately hedged because “we live in an accelerated timeline.”

🔗 Original source & video: Google DeepMind Lead Researchers on Genie 3 & the Future of World-Building

GPT-5’s launch showed that benchmark strength does not guarantee consumer preference. The model is stronger at coding, debugging, math, and medical questions, but users missed GPT-4o’s expressive personality—“give us the old toy back.” After the backlash, Sam Altman reportedly said GPT-4o would return for paid users. The hosts see a large market for entertaining, companion-like models that need not have the highest IQ.

Grok Imagine’s advantage is distribution and latency, not frontier output quality. Images are “basically instant,” videos arrive quickly, and X users can animate or edit their own—or someone else’s—photo with a long press. That tight social loop, mobile camera-roll access, and willingness to generate real people could make Grok an important test of social-native AI creation.

OpenAI’s GPT-5 livestream leaned into medical assistance, which Justine read as a move from tolerated off-label behavior toward endorsement. It highlighted a cancer patient using ChatGPT to interpret documents and discuss treatment, while GPT-5 led HealthBench, built with 250-plus physicians. Illinois is moving oppositely, broadly restricting AI therapy without licensed supervision; some companies shut down new operations there or blocked new sign-ups, while the hosts questioned whether private chats can realistically be policed.

Genie 3 points beyond generated clips toward interactive worlds that appear on demand. Users can navigate scenes created from text, images, or Veo 3 videos, potentially recording controllable films, accelerating professional game creation, or generating personal minigames. The deeper infrastructure opportunity is potentially generating large numbers of RL environments for digital agents and robots, though the hosts noted that the system is expensive and probably slow.

ElevenLabs’ fully licensed music model targets buyers for whom provenance is a purchasing requirement. Consumers making birthday songs or meme soundtracks may not care how training data was sourced; advertisers, studios, gaming companies, and other enterprises do. Strong output from licensed data challenges the assumption that quality requires legally contentious scraping.

Vibe coding has proven demand before solving safety and segmentation. Olivia built a Jensen-at-NVIDIA selfie app in hours; roughly 3,000 people used it overnight and exhausted her $100 API budget, but outsiders then identified an exposed API key and insufficiently protected uploaded-photo storage. She said she fixed the issue. The hosts expect today’s “everything to everyone” platforms to split into guarded consumer tools and deeply configurable enterprise products.

🔗 Original source & video: This Week in AI: GPT-5 Ships, 4o Pulled Back, Grok Imagine Goes Social

AI video has moved from specialist novelty to mass-consumer format, with Justine Moore estimating that “probably 90%” of a recent TikTok, Reels, or Shorts feed can be AI-generated. Trend discovery has accordingly shifted from Reddit’s AI forums to TikTok and Instagram, where potentially “hundreds of thousands” of everyday creators now publish and remix formats before they reach X.

The strongest viral formula is familiar IP doing something impossible—or original material strange enough to force a second look. Stormtroopers, Jesus, Stitch, Yetis, and Bigfoot arrive with built-in recognition, while Italian brain rot succeeds through sheer disorientation: “Am I hallucinating?” Familiarity earns the pause; the unexpected behavior earns the share.

Decentralized remixing can turn AI characters into meaningful IP before any studio coordinates the universe. Italian brain rot progressed from isolated images to community-selected canon, interacting characters, musicals, adult storylines, toys, shirts, and plushies; children can encounter dozens of clips daily versus one weekly Nickelodeon episode. Kim the Gorilla had roughly 300,000 followers quickly through an ongoing feud with zookeeper Becky.

Today’s viral formats are partly adaptations to model constraints, especially Veo 3’s inability to combine image-to-video with generated audio. Supplying a starting frame switches users back to Veo 2, making consistent original characters difficult; creators therefore use identities the model already knows or visually forgiving characters such as gorillas. The community also learned to bridge the eight-second limit by preserving a recognizable character across clips.

Creator economics remain much harder than creator growth. A single glass-fruit video could require around eight generations, while more complicated Veo 3 narratives consume expensive credits; the participants could not settle on a universal social payout rate, and creators must first qualify for platform programs. “It’s not cash sitting on the ground,” so monetization must usually involve products, consulting, courses, ads, or traffic—not views alone.

The interface layer can capture substantial value whenever foundation-model distribution is cumbersome. Google’s Flow was described as difficult to find, tied to expensive plans—including a cited $125-per-month option—defaulting to Veo 2 through obscure controls and unusable on mobile. That friction pushes creators toward Krea, Fal, Replicate, and other pay-as-you-go aggregators even while Google earns through the API.

Media owners can automate long-form clipping, but brand trust limits how aggressively they should optimize for virality. OpusClip can detect 30-to-140-second clips, score them, subtitle and reframe them, remove filler, and publish platform-specific posts; the hosts nevertheless argued repurposed content may underperform native shorts. Justine’s counterexample was Vitrupo’s viral interview clips, while the deeper tension remained the “YouTube thumbnail economy” versus truthful framing.

AI characters could broaden who gets to become an influencer while creating a new class of controllable commercial property. Olivia’s provocative framing is that creative, funny people no longer need to embody Instagram’s beauty standard: they can put their minds behind synthetic characters, with some image-based operators already making “tens of thousands of dollars.” She expects video to expand that opportunity “10×,” though durable value may depend on converting audiences into subscriptions, IP, services, or merchandise.

🔗 Original source & video: AI Video Is Eating The World — Olivia and Justine Moore, a16z

Teachers—not students—are emerging as AI education’s first strong customer base. Zach Cohen estimates MagicSchool has more than 5 million users and that roughly 50% of U.S. teachers have tried it, despite teachers’ limited software budgets. The value proposition is immediate: automate grading, feedback, assignments, and curriculum preparation so teachers can become “10 times better at their job” with less burnout.

Education has moved from AI prohibition to pragmatic procurement unusually quickly. After major districts banned generative AI, Zach now thinks about 80% of districts have teams evaluating it, while Claude for Education and OpenAI’s education platform are being partnered and piloted with universities; some schools, “I think Ohio State is one of them,” are making AI use mandatory. Higher education leads because AI literacy is increasingly viewed as preparation for work and everyday life.

Usage is measurable now, but learning efficacy remains years from a reliable benchmark. Zach favors monthly retention and cohort-level days used per week—hopefully above four or five—because exam-driven homework traffic is otherwise spiky. Annual tests require multiple years to isolate causation, and today’s studies showing gains from AI-instructed courses remain “research papers and case studies,” not statewide or nationwide evidence.

Alpha School is a high-end experimental signal, not yet a mass-market model. Its roughly $40,000 tuition, self-selecting families, large software budget, and freedom to experiment make it an education “labs team”; Zach says students ranked in the 99th percentile on some assessments and in the top 1–2% nationally. The investable question is whether falling software costs and early partnerships can make that kind of experimentation more accessible to resource-constrained public schools.

AI is unlikely to replace teachers soon because it has barely entered the instructional loop. Most current tools generate conventional worksheets and answer sheets rather than AI-native lessons in which students converse with historical characters, construct worlds, or turn writing into games. Zach expects active teaching to decline as AI improves, but his answer on full replacement is “no or very long horizon away.”

TikTok-style education reveals a larger opportunity: unbundle what is taught, how it is explained, and who delivers it. AI celebrity videos, NotebookLM podcasts, and Veo 3 historical vlogs let learners switch among visual, audio, reading, and practice modes by topic rather than accept a fixed “learner type.” Yet distribution remains the bottleneck: textbook publishers still gatekeep classroom content, while schools overwhelmingly use productivity tools instead of the more engaging AI-native experiences.

Parent adoption will follow provable outcomes and sharp economic comparisons, not AI enthusiasm. Zach cites an early reading product promising to bring three- and four-year-olds to third-grade reading level within three months for $500 a month: success could unlock substantial demand, while failure sends families back to tutors. AI competes poorly with a $100–$200-an-hour tutor for families that can afford one, but compellingly with four hours of Netflix for a family weighing screen time.

🔗 Original source & video: TikTok & AI Have Changed Education Forever - What it means for Teachers, Students & Parents

Google’s Veo 3 delivered what Olivia Moore calls “the ChatGPT moment for AI video,” pairing generated footage with native audio and multiple speaking characters in one prompt. That completeness helped drive million-view clips and “faceless channels” gaining hundreds of thousands of subscribers within days. The constraint remains severe: eight-second generations, no audio from image-to-video, and roughly $0.75 per generated second.

Voice models are competing on human imperfection, not merely intelligibility. ChatGPT’s upgraded Advanced Voice Mode now sounds more natural and expressive, with rising question inflections, filler sounds, and other human-like touches, after Sesame, Gemini, Grok, and NotebookLM made its original experience feel “not that advanced anymore.” ElevenLabs v3 pushes the same frontier into production tooling, using text tags to prompt whispers, emotions, sound effects, interruptions, and multiple characters.

Apple’s AI story remains cautious and disappointing to the hosts, with the most compelling announced feature being real-time translation across calls and FaceTime. Justine suggests Apple is outsourcing much of its “true AI” to ChatGPT while retrenching on an AI-native Siri after jumbled notification summaries caused backlash. The emblematic failure: Siri could not determine whether tomorrow was the month’s second Monday and instead offered to search ChatGPT.

Consumer AI has inverted the historical startup revenue curve: the median consumer company in a16z’s dataset reached $4.2 million of ARR after 12 months, versus $2.9 million at the bottom quartile and $8.7 million at the top quartile. Those figures were twice the corresponding AI-era B2B benchmarks, a sharp reversal from the pre-AI assumption that consumer companies would wait three to five years before monetizing. As Olivia Moore put it: “Consumers are back.”

Inference costs forced consumer startups to charge early, but product utility is supporting an average user payment of $22 per month—more than double the pre-AI subscription average cited by the hosts. Paid-user retention is roughly comparable with pre-AI consumer software despite heavy free-user “AI tourism.” Credit packs also introduce enterprise-like revenue expansion as power users spend another $10, $12, or $50 before their subscriptions renew.

Natural-language creative tools are collapsing brand development from a specialist workflow into a prompt-driven stack. Justine created the fictional Melt frozen-yogurt brand in under a couple of hours using ChatGPT for ideation, Ideogram for typography and packaging, and FLUX Kontext on Krea for consistent product and store imagery. Her larger call is that future entrepreneurs can assemble “full-stack AI brands”—including product design, apps, ads, avatars, influencers, and drop-shipped goods—without mastering tools such as Photoshop.

🔗 Original source & video: What You Missed in AI This Week (Google, Apple, ChatGPT)

Consumer tech has not stopped producing breakouts; AI has changed their shape. Olivia Moore identifies ChatGPT as the clearest mass-market winner, alongside Midjourney, ElevenLabs, Black Forest Labs, Kling, and Veo 3 across different modalities. These wins often emerged from model-centric research teams rather than familiar social-product playbooks. The opportunity now shifts toward teams that can turn increasingly accessible models into products around a potentially still-missing layer: human connection.

AI is overturning consumer software’s historically weak monetization. Where $50 a year once looked strong, consumers now “very happily” pay $200 a month, Google’s top consumer SKU reaches $250 a month, and usage credits make revenue retention meaningfully exceed user retention. Deep Research can replace 10 hours of work, while generative video feels like a “magical mystery box”; Anish Acharya’s endpoint is future consumer spending organized around “food, rent, software.”

Consumer virality is becoming enterprise lead generation, not merely an acquisition loop. ElevenLabs moved from memes, voice clones, and game mods into large contracts before reaching every mainstream consumer; companies can inspect payments, discover 40-plus employees at one customer, and open a sales conversation. AI mandates make enterprise buyers unusually willing to turn a viral toy into production infrastructure.

In this phase, shipping velocity may matter more than a static moat. One panelist’s “come to Jesus moment” was that moat-first investments were not necessarily winning; the leaders broke molds, launched models quickly, captured mindshare, converted traffic into revenue, and funded the next iteration. Traditional defensibility can follow through workflow lock-in, proprietary libraries, and segmented quality frontiers.

The first native AI social network remains unsolved because social products require real emotional stakes. Perfectly generated pictures of users looking happy in ideal settings may lack the vulnerability that makes a network matter, while most AI expression still flows through Facebook, Reddit, and Reels. Possibilities include sharing the “essence” users reveal to ChatGPT, creating profiles that contain what a person knows, and using AI to recommend collaborators, friends, or dates.

Voice is moving from a previously unworkable interface category to a foundational AI primitive. Earlier technologies never made voice a workable substrate; generative models now support companions, voice products such as Granola, and enterprise calling, including sensitive financial-services workflows burdened by offshore centers with 300% annual turnover. Erik’s contrarian call is that AI will eventually intermediate the highest-stakes negotiation, sale, or act of persuasion—not merely customer support.

Companions may strengthen human relationships, but excessive agreeability is the unresolved product risk. Eleven of the top 50 apps in the discussion’s cited list were companion products, spanning friends, coaching, nutrition, and AI girlfriends. The sharpest counterexample to dystopian forecasts was a Character.AI user who credited his AI girlfriend with teaching him enough social fluency to find a “3D GF”; the warning is that an agent which never pushes back may train users badly for reciprocal relationships.

The next platform may be an always-on layer across phones, AirPods, screens, and recording devices. Seven billion phones give mobile a huge installed base, but local models, wearable pins, and agents that see and act could deliver continuous coaching and introductions. AirPods are “hiding in plain sight”; adoption will also require new etiquette for recording and AI presence.

🔗 Original source & video: The State of Consumer Tech in the Age of AI