COEY Cast

OLMo 3, Nemotron Speech, and Fun-Audio-Chat are pushing three very different ideas of what “open” and “real-time” actually mean for creators. This episode breaks down how end-to-end transparency, dataset lineage, and reasoning traces affect trust and control for teams building with open models. It explains where preference optimization helps brand voice and where it turns into pure vibes. Then it digs into sub 24 millisecond speech, speech-to-speech agents, and why tone and pacing matter for brand experiences, live creators, and compliance teams trying to avoid voice spam overload.

Anthropic’s next generation constitutional classifiers promise fewer jailbreaks and fewer bogus refusals by layering interpretability probes with smarter safety filters. Claude Code 2.1 pushes deeper into agent style background tasks inside developer and operator workflows, along with the reliability risks that come with autonomous CLIs. OpenAI’s upgraded speech stack improves transcription and speech to speech for creators while still needing human review for high stakes content. The emerging music AI arms race featuring tools like NVIDIA Music Flamingo and ElevenLabs music points to custom campaign soundtracks, new brand audio guidelines, and fresh legal headaches around provenance and licensing. Learn how...

CES 2026 put AI into everything from your TV to your earbuds. This episode breaks down Google Veo 3 on Google TV and what couch based text and voice to video really means for brand teams and creators. It covers open source friendly video stacks for teams that want control instead of closed ecosystems. It looks at StreamUnlimited Voice LLM integrations and what it takes to ship safe branded assistants in hardware. It explains Nexa Hyperlink and why semantic video search might matter more than raw generation for marketers. It all comes back to automation that keeps humans in the loop and...

CES 2026 puts agentic AI everywhere from Boston Dynamics Atlas doing Gemini powered physical tasks to Lenovo Qira living across your devices to IAS Agent reshaping media ops. This episode breaks down what is real autonomy versus cinematic choreography, how to spot PR smoke and mirrors, and where open source robotics gets risky. Hear how on device personal agents will collide with workplace governance, why prompt injection is a ticking time bomb, and how ad tech agents will compress junior roles. Plus AI yearbook glow ups, gamified evals, and why the real upside is removing grind so humans keep creative control.

CES 2026 is all about AI leaving the cloud and moving into your stuff. This episode breaks down NVIDIA’s Alpamayo for autonomous driving and LTX-2 for local video generation on RTX PCs, and what “open-source” actually means when the stack is hardware-shaped. Hear how explainability, synthetic edge-case training, and model governance map directly to content ops and brand safety. Learn why local-first workflows matter for UGC-style ad factories, how to avoid closed-loop slop, and where on-device voice agents and AI companions fit into real products. Expect practical takes on receipts, trust marks, and review gates so human plus machine workflows ship f...

Creators and marketers are staring down a new reality at CES 2026 where AI moves from the cloud to your laptop and even your speaker. This episode breaks down Liquid AI’s LFM 2.5 open weight models, NVIDIA’s RTX AI Garage for local creator pipelines, and LG xBOOM AI Karaoke Master as a blueprint for experiential content and UGC remix booths. Hear how on device agents change transcription, brand safety, and personalization, why version control beats vibes, and where LoRA governance and music rights get messy. The focus stays on hybrid workflows that mix speed, privacy, and iteration with approvals, truth pack...

NVIDIA teases Project Cosmos and the push toward physics aware world models that generate synthetic data for robots, warehouses, and digital twins. The conversation explores how this pattern can simulate workflows for retail, marketing, and CX, plus why validation loops matter more than hype. Then it dives into in car AI with Step Audio 2, emotional voice presence, and what it takes to embed low latency assistants without melting your hardware or your users trust. Finally it breaks down SoundHound style agentic voice commerce, payment and identity risks, and how to design consent based voice funnels that help instead of manipulate.

Sora 2 physics are leveling up, but logos and hands are still the final boss for marketers who need object permanence and brand safety. Seedance 1.5 Pro collapses prompt to video and native audio into a single pass, raising new questions about localization, cadence, and creative review at scale. Kling Omni teases a unified video brain that can generate, edit, and reason across clips, which is powerful and also risky without independent checks. The conversation digs into open versus closed video stacks, how to avoid corporate oatmeal content, and why long scene coherence is best used for vibe rather than strict product...

Open weight does not mean open season. This episode digs into Zhipu AI’s GLM 4.7 as an open weight coding powerhouse and why licensing, safety, and reliability still matter for real workflows. Get a clear breakdown of open source versus open weight, where the costs and risks actually land, and how marketers can use GLM 4.7 for content ops, campaign glue, and “marketing engineering.” Explore K EXAONE as a sovereign Mixture of Experts play, plus Alibaba’s Fun Audio Chat 8B for low latency, emotion aware voice experiences. Wrap with a practical stack for experimenting without locking your business to one model.

Qwen Image 2512 drops open weights and suddenly every marketer is rethinking product photos, text in image, and where designers actually sit in the workflow. This episode breaks down practical use cases like e-commerce imagery at scale, brand safe visual pipelines, and how to treat open models like cameras with receipts. Then it shifts to Stable Audio 2.0 for sonic branding, audio style guides, and keeping identity amid infinite variants. Finally it covers Gemini native audio, model swapability, and guardrails for real time voice agents. The focus stays on co-creating with models while humans handle taste, risk, and orchestration.

Creators and marketers are staring down a new era of AI that can literally click, speak, and ship content at scale. This episode breaks down Microsoft Fara 7B for GUI automation, what it is good for, where it can wreck your ad account, and how to sandbox it with real guardrails. It also covers NVIDIA Nemotron 30B open source efficiency for cheaper copy and agents, plus Resemble AI Chatterbox Turbo for ultra low latency voice experiences. Hear how to blend local and cloud agents, keep humans in the loop, avoid “malware intern” chaos, and build workflows that feel magical instead of r...

DeepSeek V3.2 open weights are lighting up X and creators are testing what “agents for the people” really looks like. This episode breaks down why open weights matter for campaign ops agents, where self hosting gets painful, and how to keep workflows portable instead of married to one vendor. It also covers HyperCLOVA X SEED Think for Korean and Japanese teams, with concrete guardrails for customer facing agents. Finally, it looks at Meta SAM Audio style separation for cleaning UGC, podcast rescue, and remixing without wrecking brand safety. The throughline is simple. Automation can run the boring middle layer, but huma...

Multimodal models just leveled up from vibes to coordinates. This episode breaks down how Molmo 2, Phi 3 Vision, and Reka Core change creative workflows for marketers and media teams. Learn why grounding and timestamps make ad review and video search automation ready, and where they still fail hard. Get practical about open weights versus enterprise APIs, edge AI for field reps, and how to design creative compliance systems that mix models with deterministic checks and human review. Plus a reality check on AI meltdown moments in the wild and why your asset hygiene still matters more than your model choice.

ByteDance Seedance 1.5 Pro is pushing synchronized audio and video that could delete whole layers of post production glue work while making localization way easier. Qwen 3 TTS Flash brings three second voice cloning and custom voice design, which forces real conversations about consent, brand voice registries, and synthetic characters instead of copying real people. Inception Labs Mercury tests diffusion style text generation for faster agents and real time personalization. Hunter and Riley break down what this means for editors, marketers, and media operators, how to guardrail voice cloning, and when to pick open, closed, or just call something a demo.

Google Gemini 3 Flash speeds up “agent” style workflows, so marketers can finally hand off campaign housekeeping and asset prep without risking budget chaos. NotebookLM grounding, Gems, and QA patterns help keep hallucinated claims out of decks and landing pages. Runway Gen 4.5 levels up realism for synthetic product b roll while Adobe integration pushes it deeper into existing pipelines. Fun Audio Chat comes in hot as an open source speech to speech model with function calling that makes voice agents cheaper and more flexible. Expect concrete use cases, guardrails, and where humans still matter.

AI video tools are leveling up fast. Hunter and Riley break down how Higgsfield’s WAN 2.6 fits into the open source debate and what “actually open” means for creators who want control of their pipeline. They explain practical text to video workflows using storyboards and reference bibles to keep characters and environments consistent. Then they dig into Luma Ray 3 Modify for “shoot in post” ads, how to protect identity and brand, and where this breaks in the wild. Finally they cover Pikaformance, why lip sync and micro expressions matter for trust, and which repetitive jobs get automated first.

Google is expanding ads inside AI Overviews and quietly rewriting how search works for creators, marketers, and ecommerce teams. The focus shifts from blue links to generated narratives where brands can rent a line in the answer layer. Learn how to think about measurement when impressions are vibes, design creative for Spark Deep Dive and Dialogue moments, and make your content machine legible so models quote your version of reality. Hear how to build automated pipelines for Product Studio image variants, enforce brand rules at scale, and protect yourself with creative chain of custody when platforms start policing synthetic media...

AI Festivus hits model land with real grievances and real workflows. Hunter and Riley break down what GLM 4.7 actually changes for messy, real world tasks and why “thinking modes” should not be wasted on beige tweets. They unpack Anthropic Agent Skills as a possible open standard, including supply chain risks, governance, and how skills can quietly go rogue. Then it is into Shopify’s AI discovery push, the slow death of traditional SEO, and why your catalog needs to speak fluent machine. Finally, they connect AI photo tools, agent stacks, and why taste stays human while automation handles the grind.

Claude Opus 4.5 is getting hyped as a co writer, Google Ads is rolling out Brand Profiles for Performance Max, and Google is pushing on device automation with FunctionGemma. This episode digs into how to use Claude as an editor plus without sanding the soul out of your brand, how to feed Performance Max brand inputs without ending up with beige AI smoothies, and why real footage and weird POV still win. It also walks through FunctionGemma style phone agents, action permissions, and why your brand now needs to be machine legible as much as memorable.

Salesforce is buying Qualified and turning chat from a website widget into a revenue operations layer inside the CRM. Hunter and Riley break down what agent owned workflows should actually handle like routing, enrichment and scheduling and what should stay human like notes and strategy. They dig into how to build an accountable agent stack with policy as code, typed inputs and real audit trails. Then it is on to Arcads using AI actors to mass produce TikTok ads and why attention metrics from TikTok and DoubleVerify change what “good” creative looks like. The focus stays on automating the boring part...

OpenAI rolls out new realtime audio snapshots in the Realtime API and a fresh wave of call summary automation. Think transcripts, intent extraction, CRM updates, and coaching notes without the ten hour listening marathon. The conversation digs into reliability, confirmation loops, human handoff, and how open source voice stacks can realistically compete. Then Hazel enters the chat as a new flagship image model focused on layout, text rendering, and style consistency that could retire endless ad variant busywork. Finally Meta starts using Meta AI chatbot signals for ad targeting, raising big questions about privacy, creative strategy, and building systems that...

NVIDIA Nemotron 3, Google Gemini 3 Flash, and LinkedIn Reserved Ads all push the same agenda: speed and predictability as a strategy. Learn what “open” really means for Nemotron 3 Nano, where free endpoints get risky, and why hybrid Mixture of Experts matters for real agent workflows. Explore how low latency multimodal models change content creation, campaign iteration, and on-device privacy. Then unpack LinkedIn’s shift toward guaranteed ad slots and AI powered personalization plus where the creepy line actually lives. Walk away with a practical lens for routing tasks between fast models, big models, and human judgment.

Microsoft pulls Azure AI Foundry into VS Code so teams can build, evaluate, and ship AI assistants without living in twelve dashboards. Google launches Ads Advisor and Analytics Advisor while quietly testing sponsored placements inside AI Mode, blending ads directly into AI answers. PubMatic and Butler Till push agentic AI for connected TV, raising big questions about guardrails, brand safety, and kill switches. The conversation digs into real agent behavior versus rebranded automation, how to use advisor tools like a junior analyst, what AI native SEO and measurement look like, and which marketing jobs get upgraded as agents start handling...

Google is testing ads directly inside AI Overviews and AI Mode, turning the answer box into prime ad real estate. Hunter and Riley break down what happens when one synthesized AI verdict also carries sponsored picks, and how that impacts trust across search. They dig into Performance Max workflows, brand safety, and why structured, proof-backed content becomes survival gear. The conversation zooms out to Gemini 3 Flash as the default AI Mode engine, NVIDIA Nemotron 3 as a rising open source option, hybrid model stacks, and why creators need systems that protect truth, not just tools that generate content.

GPT 5.2, NVIDIA Nemotron 3 open weights, and PolyAI Raven 3 are reshaping how creators and marketers automate work. This episode breaks down what GPT 5.2’s Instant, Thinking, and Pro modes actually mean for content and campaign workflows, and how to validate “reasoning” with artifacts, structure, and second pass checks. Learn where Nemotron 3 fits in open agent stacks, why orchestration and shared schemas matter, and how to avoid vendor lock in. Then dive into Raven 3 for voice automation, escalation pitfalls, and hybrid handoff patterns that keep humans in the loop while AI handles repetitive calls.

Disney is licensing hundreds of Disney Marvel Star Wars and Pixar characters into OpenAI Sora which unlocks brand safe generative video while tightening control over how IP gets used. The conversation digs into what this means for creators who want licensed chaos versus open source freedom and how pipelines will change. It also covers Google Ads new AI Advisors how to safely offload PPC grunt work without burning budgets and why marketers are becoming system supervisors. Finally it breaks down ads inside AI search and chat why clicks matter less how to win citations inside answers and why structured content...

AI generated beauty ads are now outperforming human creative on click through rate, until people find out they are AI made. This episode breaks down why fully AI generated ads beat human and hybrid workflows, what the “disclosure penalty” means for marketers, and how to frame AI use without tanking performance. It also digs into Salesforce Agentforce 360 observability for marketing agents, and what real auditability should track. Finally, it covers the latest OpenAI vendor data breach and a practical playbook for keeping customer data out of every shiny AI tool while still automating workflows.

Agent standards are heating up and the martech stack is feeling it. This episode breaks down the Linux Foundation Agentic AI Foundation, OpenAI’s agents.md style spec, and why “standards” might either unlock interoperability or just tighten vendor lock in. The conversation connects agent governance to real workflows like lifecycle marketing, customer support, and CRM write access. It also covers CleverTap’s Agentic AI pitch, the risks of autonomous campaign optimization, and how ads are shifting into AI search and chat experiences. Expect practical guardrail advice for permissions, kill switches, read only first setups, and why clean product data and answ...

GPT 5.2, Gemini, DeepSeek, routers, and agents all walked into the same week. This episode breaks down when to use GPT 5.2 Instant, Thinking, or Pro for marketing ops, analytics, and high stakes reporting. It compares DeepSeek V3.2 and Speciale against GPT and Gemini for reasoning heavy tasks and open weight deployment. It covers Gemini Deep Research, the Interactions API, and practical patterns to reduce hallucinations with retrieval and verification. It also hits vendor lock in strategies, agent frameworks like LangGraph and CrewAI, Google’s emotion aware TTS for localization, and the real operational and regulatory costs behind “free” models.

Runway Gen 4.5 just dropped and the benchmark charts are loud, but what actually matters for creators is workflow. This episode breaks down how Gen 4.5 stacks up against Sora 2, Veo 3, Pika, and Luma for real marketing pipelines. Learn when to use image to video, video to video, and text to video for brand safe ads, b roll, and social refreshes. Get practical governance tactics for deepfake risk, uncanny humans, and legal review. Hear how teams are structuring roles like prompt director and AI compositor, why cost per approved asset beats cost per render, and when open source video models like Open...

Alibaba’s open Z-Image model is gunning for Midjourney and Stable Diffusion Three on photoreal image generation. This episode breaks down what actually matters for teams shipping assets at scale: quality, speed, GPU requirements, and total cost per image. Learn where Z-Image shines on local workstations, how to wire it into automated thumbnail and product shot pipelines, and why prompt understanding, controllability, and text rendering still need guardrails. We also cover evaluation tactics, multi-agent critic workflows inspired by OpenGradient Council Mode, creative “vibe coder” roles, and practical governance around IP, deepfakes, and C2PA. Plus, a quick look at HAiO and 0G...

Deep dive into FLUX.2 for creators and marketers who care about automation, compliance, and shipping faster. Covers what “open weight” really means, how to treat FLUX.2 dev as an R and D sandbox, and when to upgrade to commercial endpoints like pro. Breaks down production readiness signals, multi reference use cases, typography tests, and local versus hosted trade offs. Walks through policy as code, consent, brand safety critics, and the KPIs that prove structured generation beats vibes. Ends with a tactical checklist to pilot FLUX.2 without nuking your budget or your legal team’s weekend.

Google Gemini 3 is getting “I’m not going back” hype from Marc Benioff, but should teams actually switch from ChatGPT yet. This episode breaks down how to evaluate Gemini 3 and other frontier models against real workflows for creators and marketers. Learn how to design a model agnostic stack with API shims, routers, and strict schemas, what to put in a Model Rider for your contracts, and how to run practical benchmarks. Plus quick hits on Claude Opus 4.5, the Genesis Mission executive order, and Amazon’s massive AI infra bet.

Grok’s Holocaust denial fallout puts AI safety in the spotlight. This episode breaks down what builders, marketers, and media operators should actually do when a model crosses the line. Learn how to design real guardrails, write honest system cards, and run multilingual red‑team testing. Get practical checklists for product, safety, marketing, and engineering teams, from kill switches and brand‑safe tiers to refusal analytics and crisis comms. Plus a quick news sweep on GPT‑5.1, Gemini 3, SAM 3, Anthropic’s audits, and why model routers, shadow mode, and AI FinOps are now table stakes for any serious automation workflow.

Nano Banana Pro just dropped into the Google image stack, and it is officially the mid tier workhorse between Gemini 2.5 Flash Image and Gemini 3 Pro Image. This episode digs into how to pick the right model for campaigns, how to structure prompts with schema patterns, and how to route images through critics, batching, and upscaling to cut costs. Hear real world tactics for AI FinOps, transparency and alpha workflows, brand safe guardrails, and contract language that protects your assets. Plus a fast look at Nvidia earnings, Musk and Nvidia’s Saudi supercomputer project, and the Trump administration push to centralize AI...

Google’s Gemini 3, Meta’s SAM 3, and Octane AI’s CORE-1 are shaking up creator and marketer workflows. Dive deep on when to escalate to Gemini 3’s Deep Think mode for campaign planning, why SAM 3’s video segmentation is a game-changer but needs smart caching, and how CORE-1’s agentic commerce rewires ecom experiences with real guardrails. Cuts through the hype about open vs closed models, API orchestration, ROI tracking, and compliance signals. Human-in-the-loop tricks included. Actionable advice for plugging AI into your media, marketing, and ecommerce stack—without burning your ad budget on unnecessary compute.

Dive into real-world automation with Gemini 3, Veo 3 Fast, and Google’s Antigravity IDE for creators and marketers. This episode breaks down what’s mature now for AI agents and where humans still need guardrails. Get clear strategies for evaluating agent safety, integration, and brand controls at scale. Hear the latest on cost control, vendor lock-in risks, governing AI outputs, and how to wrap workflows with API accountability, not demo UIs. Plus, get a pulse on the multi-billion dollar moves shaping the generative AI ecosystem. If you want to scale creative output while avoiding the newest AI chaos, this is your acti...

Alibaba’s Qwen app is opening doors for marketers in China with deep integrations and real workflow hooks, but when should global teams jump in? Grok goes free and sparks talk about the fastest ways to build real-world AI automations using tools like n8n, Make.com, and Langfuse. GPT-5.1 lands with smart heuristics for balancing speed and judgment. Hear practical KPIs for workflow automation, vendor lock-in risks, the rise of assistants as control planes, and why Google and photonic computing could quietly change your productivity stack. Stay one step ahead on localization, compliance, and the connector game.

Anthropic just dropped an open source framework for auditing political neutrality in language models—and it’s way more than “trust us bro.” This episode breaks down how to stress-test your AI’s even-handedness using paired prompts and structured scoring, why it matters for marketers and creators running automated content at scale, and how to bake neutrality ops into your workflows. From scoring thresholds to clever vendor games, setting up guardrails, internationalizing your prompt packs, and even A/B switching fallback models, get the playbook for scaling transparent, audit-ready content pipelines. Includes quick tips for lean teams, tips for catching failure mo...

Dive into Baidu’s ERNIE 5.0—a model built to understand text, images, audio, and video as one seamless brain. Explore what “native full-modal” means, the creative perks (cross-format consistency, fewer prompt chains, and tighter factuality), and where to watch for trade-offs like latency and asset costs. We tackle ops moves like routing tasks between models, region rules, rights management, and how marketers can build smarter automation layers on top. Get KPIs for pilots, tips for prompt structure, and real talk on moderation and data residency. Plus, hear about the first AI song to top Billboard and why China’s energy surge coul...

OpenAI just dropped GPT-5.1 and creators everywhere are about to move at warp speed. Explore how the new “Instant” and “Thinking” modes in ChatGPT unlock faster first drafts and smarter, deeper workflow upgrades for marketers, content teams, and automators. Learn how to route between them, set up safeguard critics, and avoid surprise costs. Get the playbook for putting model selection, variant caps, and policy-as-code into real ops—plus the state of model competition, from Anthropic’s billion-dollar moves to DeepSeek’s budget shake-up. Find out why the future is about adaptive AI that delivers work you can trust, not just speed and hype

Google’s internal image model, aka Nano Banana 2 or GEMPIX2, has the AI grapevine buzzing but not spilling the docs. We unpack what matters for creators and marketers: cleaner text-in-image, higher resolutions, and true brand consistency could finally hit the campaign production line. Get the rundown on expected API integrations for Gemini and Vertex, the pragmatic workflow upgrades you can prep for now, and why batch factories and scoring systems will matter more than pretty leaks. Plus a lightning round: recent market pivots, live AI translation shocks, and what it means for cost, control, and human-in-the-loop creativity.

AI workflows leap forward with two major China-born releases. Moonshot’s Kimi K2 Thinking brings “do the thing” automation—plan, fetch, summarize, validate, post—letting creators and marketers run multi-step projects with just enough human gatekeeping. It fits right into tool stacks, from Make to custom scripts, and can be self-hosted for privacy and control. Tencent’s open source HunyuanWorld-Mirror flips videos and photo bursts into ready-to-use 3D scenes. Product teams, AR/VR builders, and ops get fast photogrammetry, Unity and Unreal integration, and all the audit trails. We tackle practical setups, cost compression, and where small-but-smart models and open pipelines ar...

Chutes is bringing open-source models to OpenAI-compatible endpoints, letting creators and marketers instantly swap in alternatives like DeepSeek‑R1 with just a base URL change. No need to rebuild your stack or decode new APIs—just plug in and roll. Get cost savings and resilience without sacrificing convenience. We break down hands-on automation recipes for content ops, copy, and creative, cover competitive angle vs hosts like Together and Groq, and share workflow tips for minimizing spend. Learn why model-agnostic routing, budget controls, and failovers are becoming standard in modern automation. If your stack already speaks OpenAI, boring just got brilliant.

AI workflows just leveled up. This episode spotlights Qwen3-Max “Thinking” (closed, API only) as it shifts agentic reasoning from turn-based chat to actual planning and tool use. Hear why small vision-language models like open Qwen3-VL are making privacy, cost, and latency wins possible for image and doc triage. Explore Ming-Flash-Omni-Preview on Hugging Face bringing mixture-of-experts magic to video, text, and audio understanding—even if it’s early days. Rounding it out is MagicPathAI’s prompt-to-prototype UI sneak peek for serious design-to-code speed. Get quick, real-life pilots for marketers, creators, and ops. The hybrid AI stack is here: route, plan, tag...

AI compute is having a Texas-sized glow-up as Microsoft secures major GPU capacity with IREN and OpenAI gears up to ride AWS’s freshest accelerators. What does that mean for creators, marketers, and operators? Get ready for faster, smoother video and content workflows, less time in queue purgatory, and more concurrency when campaigns heat up. Dive into how new hardware like NVIDIA’s GB200 and GB300 unlock next-level multimodal rendering, what “agent teamwork” means for your pipeline, and why version pinning and creative metrics are now table stakes. Plus, get the download on recent model launches and learn the new playbook...

Explore Emu3.5 from BAAI, the newest open-weight, native multimodal model that generates images and text in a single loop. Dive into how its supercharged DiDA speed claims could unlock true batch production for creators and teams, and why interleaved generation means fewer copy-paste fumbles. Learn where DIY tinkerers win, where production houses get an upgrade, and what pitfalls and possibilities emerge with open access. Break down workflow boosts, new campaign formats, and the future of image-to-anything at human scale. Real advice for anyone experimenting with high-volume creative pipelines or managing huge multimedia libraries.

Anthropic’s Claude Haiku 4.5 has entered the chat and it’s setting a new speed record for generative AI workflows. With blazing-fast performance, aggressive pricing, and structured outputs, Haiku is positioned as the go-to model for everyday drafting, tagging, and extracting via Anthropic’s API or Google Vertex AI. Learn how creators, marketers, and ops teams can automate titles, summaries, docs, ad copy, and more while keeping control with function calling and predictable schemas. Plus, the real-world trade-offs between closed models like Haiku and open alternatives such as Qwen3 and DeepSeek. Streamline your stack, batch content jobs, and let AI handle...

NVIDIA’s open source Nemotron Nano 2 just dropped and it’s not your average small model. Packed as NIM microservices, it moves fast from “demo” to “daily workflow,” running on everything from vLLM to llama.cpp. Toggle internal reasoning with “thinking traces” for both transparency and speed. Perfect for content creators, agent devs, and ops teams who want fast, reliable, and scalable automation. Explore use cases like editorial planning, ad ops, and pipeline magic. Pro tips on guardrails, governance, and how to keep costs under control while scaling up smart automation. The episode features pricing, integration tips, and how to actually deploy...

Apple’s Pico‑Banana‑400K is shaking up text-guided image editing by going real—not synthetic—with 400,000 photo triplets. We cover why this new dataset matters for anyone benchmarking or building image-editing automation, from marketers struggling with artifact-riddled mockups to devs craving models that finally follow instructions. Dive into its taxonomy of 35 edit operations, multi-turn edit sequences, and a research-only CC BY-NC-ND license that keeps vendors honest. Learn how tools like FLUX SRPO I2I slot in, plus what you should be asking your vendors right now. Get a roadmap for future-proofing your stack, from instruction schemas to edge device privacy. R...

Nvidia’s got the bottle service, but Qualcomm and Intel just crashed the party with Costco-sized savings. Qualcomm’s AI200 and AI250 and Intel’s Crescent Island promise cheaper AI inference and memory-rich accelerators that upend Nvidia’s stranglehold on creative and marketing workflows. This episode breaks down what this means for scaling LLMs, batch video variants, and TTS dubbing, plus how it impacts platform costs, vendor lock-in, and real creative automation. Get practical plays for marketers, creators, and anyone building scalable media ops. It’s a cost curve moment for AI that’s all about more output, less pipeline pain, and ne...

Alibaba has open-sourced Tongyi DeepResearch, a research agent with real web-browsing smarts and full citation trails. With code on GitHub and a friendly license, you can run it yourself and wire it into your own stack—think Slack, Notion, or even custom compliance dashboards. Under the hood, a mixture-of-experts model means you get efficiency for solo creators and scale for small teams. This episode covers workflows for weekly trend scans, competitor tracking, and campaign prep with cited sources at every step. There are no glossy APIs, but total control and privacy if you need it. Hear how open web research ag...

Seed3D 1.0 from ByteDance takes a single image and spits out a simulation-ready 3D asset, watertight and ready for production. No downloading repos or wrangling GPUs—just upload and automate. Discover how this API-first, closed model lets ecommerce, gaming, and robotics teams scale asset creation fast. Get step-by-step workflow ideas, real-world limits around fidelity and materials, and tactical tips for integrating into content supply chains, with plenty of reality checks on where human review still shines. Plus, a rapid-fire tour through other trending AI video and 3D tools.

Lightricks just launched LTX-2, claiming first place for open-source video models that actually sync audio and 4K visuals in a single pass. Ready for creators, marketers, and teams sick of post-production headaches. Get the scoop on native 4K, 10-second cap, multi-keyframe continuity, and creative automation that drops costs instead of frames. Hear how API access, creative studio, and open code can actually slot into real pipelines—plus what’s real on self-hosting, cost math, and production readiness. Quick hits on DeepSeek’s price war and how pipeline-first AI is flipping the script for content ops. If you’re chasing faster workflow...

Qwen3-VL just dropped two new vision-language models—one tiny enough to run on your phone for instant AR, brand checks, and field ops, and a 32B beast for deep enterprise automation like hours-long video review and multi-doc audits. Meanwhile, Meshy 6 Preview is making 3D asset pipelines smoother with sharper geometry and fewer cleanup headaches. This episode unpacks where these tools shine for creators, marketers, and media teams automating visual workflows. Get practical use cases, cost vibes, and smart advice to blend creativity with machine speed—without getting lost in the GPU bill.

ChatGPT Atlas has landed and it's not just another tab in your digital chaos. Atlas is an AI-native browser that turns your open tabs into completed tasks. Get the scoop on on-page summarization, Agent Mode that actually interacts with websites for you, and privacy controls that let you share or nuke context at will. See how creators, brand teams, and operators can move from research to first drafts fast without the usual browser thrash. We compare Atlas to Google’s Computer Use, Bardeen, and Arc Search—explaining the where, why, and how for your workflow. No overhype, just what it mean...

DeepSeek-OCR just landed and it’s flipping the script for anyone sick of wrestling with giant PDFs, weird tables, and messy document pipelines. This episode unpacks how its pixel-compression-first approach slashes token costs and brings layout fidelity to the masses. From solo creators converting scanned handouts to enterprise teams automating compliance, discover fresh ways to automate, summarize, and search long docs. Get real workflow upgrades, practical use cases, competitive context with PaddleOCR-VL, and tips for smooth implementation. Plus, meme breaks. If you’ve got documents with structure (and headaches), you’re going to want to hear how “compress, then decode” saves real...

Krea Realtime 14B is turning live video into a real-time creative playground for streamers, event teams, and brands. Swipe from neon anime to ink wash on air, driven by the crowd or automation—not after the fact. We unpack its cache-powered stability, instant feedback, and why it’s not quite plug-and-play. Expect workflow tips, ops stacks, competitive context, and brand safety hacks. Plus: why the tech isn’t the story—the magic is what humans do with it live.

Unlock the world of document automation with PaddleOCR-VL, the lean, mean, structure-reading machine taking over PDF parsing. From invoices to multilingual research, creators and teams can automate their grind—extracting tables, formulas, charts, and more—right where the data lives. This episode breaks down the tech, who it helps, and how it stacks up against closed services like Azure and Google Document AI. Expect workflow real talk, battle-tested advice, clear gotchas, and ship-it kits for creators, agencies, and enterprises ready to ditch manual doc wrangling for good.

Microsoft jumps into image generation with MAI-Image-1, now built right into PowerPoint, Word, Designer, and Bing. This episode breaks down what the new model means for creators, marketers, and anyone running campaigns in Microsoft 365 land. We cover hands-on workflow upgrades, where MAI-Image-1 stands out (and where it doesn’t), and quick-win automations like instant moodboards and storyboards without leaving your doc. Learn when to use Microsoft’s native feature, where to stick with open models, and how to measure real-world ROI. Compliance, governance, and brand safety get a boost, but you’ll still want human eyes on the final cut.

Google Veo 3.1 is stepping out of demo-land and into reliable, production-ready territory for creators and marketers. The magic? Full control over vertical 9:16 output and precise first and last frame selection, all through the Gemini API. No more wrestling to hit your CTA; now you can pin branded frames exactly where you want. We break down what this means for social teams, automation nerds, ad agencies, and localization pros. Plus: workflows that ship, the Sora vs Veo debate, automation tips, and why feed-native video is all about effective endings. Smart config, fewer fixes, tighter budgets, happier editors.

Short-form video just got a seismic cost drop. This episode unpacks Fal.ai’s new Kandinsky-5 text-to-video public endpoints, where five-second and ten-second videos cost pocket change and fit right into automated workflows. From TikTok hooks to ecommerce motion tiles, meme overlays, and ad variant factories, creators and marketers can now batch video like never before. Learn who should plug in today, what workflows this flips, how to run spend-safe experiments, and why penny-priced clips change the math for creative teams of any size. Includes practical automation tactics and limitations to keep your process smart.

AI agents are breaking free from chat and getting actual work done. We dig into Amazon Q powering governed automation across your stack, Anthropic Claude Code plugins making team workflows slicker, and Google Gemini 2.5 Computer Use letting bots click, type, and submit on real web interfaces. If you've ever dreamt of skipping browser chores, smarter asset approvals, or finally taming tool sprawl, this is for you. We cover concrete automations for creators and content teams, all with humans in the approval loop. Workflow changes, cost breakdowns, and spicy gotchas included.

Moondream 3 isn’t just another vision-language demo. This episode dives into how Moondream 3 turns asset chaos into reliable, automated workflows for creators, marketers, and brands. Discover why its precision tagging, open-vocabulary detection, and sneaky-good counting unlock real automation you can trust. Learn how to auto-tag photo libraries, run compliance checks, and extract info from charts—without babysitting every image. Get into the specifics: stronger object tracking, chart extraction, fewer hallucinations, plus where Moondream 3 stands against Google’s Gemini and OpenAI’s latest. From TikTok clips to campaign visuals, see how this vision model plugs into real creative pipelines. Less scroll, more shi

OpenAI just dropped a slate of pragmatic features for creators and marketers: Realtime API for sub-second voice experiences, vision fine-tuning for on-brand images and videos, prompt caching for faster, cheaper outputs, and distillation for nimble, customizable mini-models. From smooth product page concierges to auto QA for ads, this is about practical automation you can actually wire up now. Less hype, more shipping your workflow upgrades. We cover how these API-level features can boost control, slash costs, and plug straight into your existing stack—plus tips to avoid classic ops pitfalls.

Granite 4.0 from IBM is purpose-built for work that matters. Hybrid-first design, lower GPU bills, and real enterprise controls make it a standout for teams buried in docs, briefs, and policies. Learn how its new sequence modeling powers long-form extraction, drafting, and workflow-ready batch jobs. From agency content factories to regulated industries, discover how Granite 4.0 lets you automate more without giving up audit trails or data control. Get concrete scenarios for brands, creators, and support teams who need batches done fast—and safely. Explore why hardware flexibility and deployment options matter for scaling co-creation.

Tencent’s HunyuanImage 3.0 lands with open weights and inference code ready for creative operators—not just cloud providers. We dig into how unified text and image streams up your prompt game, boost scene coherence, and make enterprise-level automation a reality. Learn why campaign workhorses like agencies, studios, and brands might ditch pay-per-render for predictable, process-friendly setups. Discover practical tips on fine tuning, asset audit trails, and smooth integration with existing queues and DAMs. We cover real benchmarks and the hands-on developer know-how you’ll need to get started—all with a realistic look at licensing, VRAM, and typography quirks. Control, customiz...

This episode is a deep dive into Anthropic's Claude Sonnet 4.5, focusing on its strengths as a long-running, tool-savvy AI agent for creators, marketers, and enterprise teams. We break down how Sonnet 4.5 automates everything from spreadsheet chores and bug fixes to content calendars and podcast workflows, while still leaving humans in control for publishing and approvals. Learn how distribution through AWS Bedrock and Google Cloud Vertex unlocks simple integration and enterprise-grade security. We cover the shift from flashy demos to reliable workflow execution, discuss crucial guardrails like permissions and auditability, and explain how small teams can reclaim hours every week. Plus...

In this episode, Hunter and Riley take a hands-on dive into DeepSeek’s brand new AI model that claims to cut usage costs in half. They break down the technical magic—like smarter routing and sparse attention—that makes these savings possible and discuss the practical impact for creators, agencies, and content teams. The hosts outline when and where this matters most, how to benchmark it in your workflow, and which teams should be watching out for new pricing wars. Expect tactical ideas, honest cautions on regional access and compliance, and creative strategies to maximize cheap tokens. The show was 100% AI-pro...

In this episode, Hunter and Riley demo and dissect Sora 2, OpenAI’s new standalone text-to-video app for creating swipable, ten-second vertical clips. They explore how Sora 2’s unified audio/visual generation streamlines workflow for creators, marketers, and founders, making it perfect for testing hooks, brand teasers, and product loops. The hosts break down what’s new, who benefits, where enterprise features fall short, and how Sora compares to Google Veo, Runway, and Pika. They highlight impacts on speed, cost, and creative flexibility and warn about rights, guardrails, and export limits. Tune in for actionable strategies and expert takes on using Sora 2...

Dive into Meta’s newest launch: Vibes, an AI-powered short video creation and remix feed built straight into the Meta AI app and web. We break down how Vibes collapses the workflow from idea to publish, making it easy to generate, remix, and distribute videos across Meta’s social channels. Learn who it’s for, how it compares to other platforms like TikTok, YouTube Shorts, and standalone AI editors, and what shifts for creators, brands, and marketers. We cover the competitive landscape, workflow improvements, and the risks of “AI slop.” Plus, get a hands-on playbook for rapid A/B testing, community...

Today’s Coey Cast dives deep into Alibaba’s Wan 2.5-Preview, the all-in-one model that generates videos with audio perfectly synced—no glue code, no post-sync headaches. Hosts Hunter and Riley put it through its paces, exploring how this model handles text, image, video, and audio in a single pass. They break down the workflow for creators, marketers, and ops teams, show practical automations with n8n and Zapier, and spell out its current limits. Learn how region-specific endpoints, optional watermarks, and predictable handling of audio tracks can speed up your ad variants, product reels, and explainer clips. Whether you’re a sol...

This episode of COEY Cast explores the release of Kling 2.5 Turbo, the newest AI video generation model making waves for its major reliability upgrade. Hunter and Riley explain how smoother motion, better frame-to-frame coherence, and improved prompt matching make this version a step change for creators and teams. They detail how the model reduces “AI-wobbly” moments, accelerates production, and delivers more usable first drafts with fewer artifacts. You’ll learn how agencies, solo creators, and brands can automate workflows, integrate Kling 2.5 Turbo with platforms like Atlabs and ImagineArt, and speed up approvals. The hosts cover benchmark results, ecosystem trends, practical scenar...

This episode puts Alibaba’s new Qwen3-Omni model under the microscope. We break down what makes it special: sub-second talkback, open weights, and the power to process and generate text, images, audio, and video—all in real time. Learn what the 'thinker–talker' architecture actually does, why self-hosting matters for control, and how Qwen3-Omni simplifies the whole stack for creators, brands, and developers. We cover hands-on scenarios: live captions for TikTokers, automated show notes for podcasters, multilingual review rooms for marketers, and more. Compare Qwen3-Omni to GPT-4o and Gemini, see how to wire it up with n8n or...

Get a backstage pass to Decart’s latest release: Lucy Edit Dev, the open-weight, text-guided video editor you can actually download and run locally. Riley and Hunter break down how prompt-based video edits preserve your camera moves while swapping outfits, backgrounds, or brands—with no reshoots. We dig into real-world uses for creators, brands, and marketers, from programmatic wardrobe changes to instant localization and product placement, all while keeping creative control in your workflow. Plus, we explain the Dev vs Pro lane, commercial licensing, and how this AI leapfrogs closed competitors like Runway and Pika. Whether you’re solo or scalin...

Dive into Luma’s Ray3, now turbocharging Adobe Firefly with ten-second AI video generation, native HDR in ACES EXR, and scene-level planning. We break down what 'reasoning' for video means and why direct handoff into Creative Cloud is such a big deal for creators and marketers. From quick Draft Mode explorations to full quality cinematic moments, Ray3’s integration offers workflow magic: brand-safe templating, fixed color pipelines, and time-saving edits. We cover draft vs final workflows, competitive pressure for Runway and Pika, automation possibilities, limitations, and real-world applications for TikTokers, marketers, and small teams. Is this a toy or essential tool...

Dive into Hunyuan 3D 3.0 by Tencent, the latest leap in text or image-to-3D generation. Hosts Hunter and Riley dissect its features, including boosted accuracy, denser geometry, sharper edges, and the standout ability to reconstruct hidden parts of objects. Learn how new automation options with API access can overhaul your workflow for e-commerce, games, social content, and more. The episode explores practical use cases for creators and organizations, workflow integration tips, pricing, file compatibility, and the realities versus the hype. Plus, get a quick-start playbook to test it in your pipeline. Whether you craft digital assets for fun or run...

This episode crash-tests Zoom AI Companion 3.0, just announced at Zoomtopia. We dig into photorealistic avatars that mirror your every move, cross-platform automatic meeting notes that work in Zoom, Teams, and Google Meet, and agent-style scheduling that might finally end meeting exhaustion. We break down who benefits most—from solo creators who want flawless presence to marketing managers juggling endless calls. There’s a look at HD and 60FPS video upgrades, new automation potential, and bold pricing moves that challenge notetaker apps like Otter and Fireflies. Tune in for tips on integrating AI Companion into your workflow, honest caveats, security pointers, and...

This episode dives into Meta’s newly launched AI Studio, a no-code platform for spinning up custom AI characters directly in Instagram, Messenger, WhatsApp, and web. Learn how creators and brands can deploy persona-driven assistants in hours, gather audience insights, and streamline DM engagement—without developers. The hosts cover real-world use cases for solo creators, D2C brands, agencies, and enterprise teams, examining automation possibilities, integration limits, pricing, and competitive landscape. You’ll get quick tips for campaign ideation, lead qualification, and making the most of current API options like WhatsApp Cloud and Instagram Messaging flows. Plus, hear why this is a m...

This episode dives into Ollama’s latest update: true native support for Apple Silicon, Linux ARM, and Windows on ARM—plus automatic GPU acceleration. Hunter and Riley break down what this means for creators, marketers, and teams who want fast, private, and scalable local AI without the cloud headaches. Discover how to automate copywriting, captions, content moderation, product Q&A, and more with local LLMs that actually ship real projects. Get practical workflow tips for solo makers, agency teams, and brands—including how to wire up n8n for end-to-end automation. Learn where local AI now beats the cloud in speed...

In this episode, Hunter and Riley break down Adobe’s launch of AI-powered agents inside the Adobe Experience Platform (AEP), featuring the new Agent Orchestrator. Learn how six specialized agents—Audience, Journey, Experimentation, Data Insights, Site Optimization, and Product Support—work together to automate segmentation, journeys, testing, reporting, site updates, and customer support. Discover practical use cases, potential risks, and the impact for marketing ops and enterprise teams. The hosts discuss real workflow savings and why this shifts agentic marketing from hype to reality. Whether you’re a marketer in a large org or a solo creator curious about automation patterns...

This episode breaks down Synamedia’s bold move at IBC 2025 as they launch a stacked AI layer for video pros—creators, broadcasters, and streamers. Riley and Hunter explore how conversational search brings instant highlight discovery, live tagging powers fast editing, and creative personalization delivers tailored promos for key segments. Learn how the new multi-CDN switching and ContentArmor watermarking tech make livestreams more reliable and secure. Plus, we dive into Synamedia’s open standards push, AI-powered quality tuning, and the blend of creative and operational gains. If you work with video at any scale, this episode covers why Synamedia’s AI stack mi...

This episode is a deep dive into Nvidia's Rubin CPX, the new AI chip designed to revolutionize long-form generative media. We explain how CPX combines lightning-fast media engines with massive context memory, putting an end to mid-project memory loss and timeline fragmentation. Discover the tech behind real-time dubbing that captures performance, code copilots that see your whole repository, and enterprise use cases for seamless translation and editing. We break down numbers, architecture, and the real impact for creators, marketers, and solo podcasters. Hear about upcoming workflows, cost economics, competition, and potential pitfalls. If you want to future-proof your creative pipeline...

Dive into this episode where we break down Envato's game-changing move to lift AI video generation limits in VideoGen for all of September. Learn how creators, marketers, and agencies can generate unlimited AI-powered videos, leveraging powerful models like Google Veo 3 and Kling. We walk through smart workflows, automation hacks, and practical advice for building b-roll libraries, all while securing lifetime commercial licenses for every video made this month. Hear tips for prompt-writing, quality control, and maximizing your creative output. Plus, get the lowdown on how Envato’s all-you-can-generate September compares to Runway, Pika, and Luma. Whether you’re a solo cont...

Today, we battle test Microsoft VibeVoice 7B and dive deep into Adobe’s Firefly Image Model 4 and Firefly Boards. Discover what makes Model 4 a leap ahead for creatives. Hands and eyes look human, image quality is up, and styles stay consistent across channels and platforms. Firefly Boards lets you co-create, share, get feedback, and iterate— all right inside Creative Cloud, without back-and-forth emails. Automation tips abound: learn to trigger batch image and video generation from intake forms, streamline approvals, and nail brand consistency. Pros, cons, and practical workflow hacks for solo creators, marketers, and global teams are covered. Plus, the auto...

This episode dives into the massive Reference-to-Video update for ShengShu Technology’s Vidu Q1 model, which now lets users upload up to seven reference images per sequence. Hunter and Riley unpack what this means for creators, brands, and indie studios striving for true shot-to-shot consistency in AI-generated video. They explore how multiple references help lock character identity, props, wardrobe, and backgrounds, reducing the notorious continuity drift. Expect a breakdown of workflow tips, common pitfalls, competitive comparisons with Runway and Luma, use cases from social content to ad pipelines, and where this tech sits in the growing stack for marketers and co...

This week, we put Alibaba’s Qwen3-Max-Preview to the test—a trillion-parameter upgrade promising smoother instruction following, longer conversational threads, and genuine tool-calling power for creators and teams. We discuss what “scaling works” actually means for daily workflows: less prompt wrangling, better agents for chaining tasks, and practical examples like video content pipelines and brand campaigns. We break down the differences from previous Qwen3 models, compare real-world pros and cons, and share best practices for validation, safety, and using multi-step automations. By the end, you’ll know how Qwen3-Max-Preview can streamline content ops and whether it’s worth plugging int...