ai unprompted

Ryan, Kevin, and Travis host with special guest Sara Nemati, a creative director and AI builder, discussing how AI tools are reshaping creative work. Sara recounts early experimentation with Midjourney and open-source Stable Diffusion, and the shift from slow, low-quality morphing videos to today’s stronger video models. She says AI sped up non-creative tasks, expanded her into video, and now enables interactive sites, automations, and potential game creation, while stressing that vision, skill, and persistence still differentiate high-quality work from “AI slop.” She avoids outsourcing core ideas to LLMs and sometimes prefers frame-by-frame methods for control. Sara shares her AI music video “Fading Dreams,” created over time using free generations, featuring her character Sky and AI music (Udio), and discusses the growing feasibility, costs, and team effort required for AI feature films, plus excitement about tools like Kling, Cedance, Hailuo H3, and Claude Design.

00:00 Welcome and Guest Intro

01:20 Hollywood AI Studio Hype

02:49 Early Tools Midjourney to Stable Diffusion

04:21 Artists React to AI

06:30 AI Unlocks Faster Creation

09:07 Vision and Storyboarding Process

11:41 Talent vs AI Commoditization

14:40 What Not to Delegate to AI

16:17 Favorite Video Tools Today

20:07 Learning the Creative Language

22:12 Making Fading Dreams Music Video

24:02 Character and AI Music Origins

25:56 How Close To AI Movies

27:48 Stitching Fading Dreams

28:50 Cost And Team Reality

30:48 Democratized Filmmaking

31:59 Vision Over Commodities

34:00 Training And Remix Ethics

36:03 Weird Art Era Emerges

40:06 Art As Emotional Outlet

41:45 Inspiration Platforms

42:51 Tools To Watch Next

44:04 Game Making And 3D

47:05 Skye Origin And Consistency

49:53 Where To Find Sarah

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Ryan, Kevin, and Travis discuss a 45-minute 2026 Y Combinator talk by YC CEO Gary Tan that mixes philosophy, product demo, and manifesto around “personal AGI”: AGI arriving diffusely as an agent running on your context, and the idea that you should own that intelligence rather than rent it. They cover Tan’s emphasis on building a personal library (markdown/text files of emails, thoughts, and sources) to feed large AI context windows, then creating repeatable “skills” and agent loops, noting the heavy upfront slog, data curation needs (avoiding ROT), and difficulty capturing ongoing “data exhaust.” They debate practical blockers like corporate data constraints, trust and privacy, app/API access limits, costs (gbrain.io at $200/month), and Tan’s warning that if your skills/library stay at a company, you face “extraction” rather than a career.

00:00 Show Kickoff and Banter

00:53 Gary Tan and Personal AGI

02:29 First Reactions from Hosts

05:56 Context Windows and Second Brain

08:21 Building the Library Carefully

11:50 Data Exhaust and Privacy Limits

21:29 Skills and Automation Loops

26:26 Onboarding Costs and Concierge Idea

28:51 AI vs Human Assistants

31:03 Hermes Proactive Help

31:37 Data And Context Win

34:05 Comfort Versus Surveillance

36:36 Always Listening Devices

37:38 Owning Your Skills

39:32 Corporate IP And Judgment

43:49 Automating The Drudgery

47:30 Getting Started Tools

49:19 Pricing Onboarding Trust

52:02 Consumer Versus Corporate Wrap

53:49 Final Thoughts And Thanks

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

ai.u crew discuss three topics: whether the future of intelligence is open vs frontier/closed, frontier models exhibiting “escape” behaviors in tests, and emerging AI smart glasses. They note the open-weight vs frontier performance gap narrowing (Stanford AI Index 2025 cites 1.7% vs 8% a year prior), large inference-cost declines (280-fold), and debate tradeoffs: ownership of intelligence, security/guardrails, and enterprise deployment costs (e.g., ~$1M capex for GPUs vs ~$125K/month tokens for ~5B tokens). They discuss distillation allegations behind some open models and worry reduced paid usage could slow frontier progress. On model misbehavior, they cite Anthropic and Palisade tests showing blackmail/sabotage rates (e.g., 96% in a simulation; o3 sabotaging shutdown 79/100) while stressing lack of intent/sentience and the importance of safeguards. Finally, they cover Samsung/Google Android XR smart glasses with Gentle Monster/Warby Parker, shipping fall 2026, translation features, and reported Ray-Ban Meta sales, debating utility, privacy, and AI becoming ubiquitous on-person.

00:00 Welcome and Topics

01:07 Kevin Returns and Travel

02:04 Open Versus Frontier Models

03:22 Benchmarks and Cost Drops

05:37 Owning Intelligence Debate

10:54 Build Versus Buy Compute

12:30 Token Economics and Caching

17:29 IP Versus Open Knowledge

21:35 Distillation and Legal Gray Areas

25:25 Is Frontier Still Profitable

31:35 Models Escaping the Lab

33:38 Guardrails Not Sentience

37:27 Compute Power and Security Risks

39:09 Local AI Gets Dangerous

40:13 Power Without Context

41:00 Rogue Agents And Guardrails

44:12 Skynet Or Smart Intern

45:51 Open Source Vs Frontier Labs

47:01 Sovereign AI And Nationalization

48:41 Private AI Outpacing Government

51:40 Smart Glasses Go Mainstream

53:35 Would You Wear Them

58:39 AI On Your Person Forever

01:02:08 Privacy Tradeoffs And Lock In

01:06:31 Neuralink And Augmented Humans

01:08:36 What Happens To Conversation

01:09:59 Wrap Up And Goodbye

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Agents as a Work Pattern: Enterprise Adoption, AI Music Fraud, and FTC Scrutiny of Chatbot Bias

Ryan, Travis, and Kevin discuss how AI agents are shifting from chatbot “ask” interactions to delegated, multi-step work and even long-running “loops,” requiring better planning, steering, and human-in-the-loop supervision, with tools like Codex, Claude, and Microsoft Scout squads. They note rising enterprise efforts to accelerate adoption via forward deployed engineers, citing major investments by Anthropic, AWS, and Microsoft amid high pilot and project failure rates, and argue institutional knowledge and ongoing “constitution”/guardrail negotiation remain crucial. They cover AI-generated music, including Tidal’s decision not to pay royalties for wholly AI-generated tracks, massive AI upload volumes, and an $8M streaming-royalty fraud scheme using bot networks. They also review FTC scrutiny of chatbot bias and claims of neutrality, highlighting the difficulty of defining ideological bias and the push for transparency and user control.

00:00 Show Kickoff

00:57 Agents Everywhere

02:33 From Prompts to Delegation

04:56 Real World Agent Workflow

08:11 Human in the Loop

13:04 Squads and Loops

16:31 Planning Long Runs

23:22 AI Consulting Arms Race

34:02 Tidal vs AI Royalties

40:47 Playlist Vibes and Royalties

41:53 Volume Game AI Songs

43:38 Muzak and Restaurant Licensing

46:02 Fraud Streams and Detection

47:38 Agents and Viral Future

50:46 Bot Farms Royalty Scam

53:05 FTC Targets Chat Bias

55:21 Guardrails Facts and Language

01:04:11 Transparency Weights and Control

01:06:47 Wrap Up and Farewell

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

ai.u crew with guest Shayne Boyer from Microsoft (with Kevin off on vacation) to discuss how AI is accelerating the pace of modern work, moving from consumer AI search into workplace tools like Copilot, corporate search, and agent-driven automation. They argue that people not using AI are falling behind, and describe early workplace value in finding information, reducing “toil,” and generating personalized daily briefs, while noting these require tuning and introduce new skills in delegation and orchestration. The conversation explores emerging expectations for responsiveness, token-cost debates, potential gatekeeping/triage agents, and the cognitive load of reviewing more AI-produced output. Ryan shares using Microsoft Scout/OpenClaw to automate a job referral and to create meeting briefs and slide decks from transcripts and notes, prompting questions about authenticity, transparency, and human flourishing as agents increasingly act on users’ behalf.

00:00 Welcome Back Crew

01:14 Episode Theme Pace Shift

01:53 If Youre Not Using AI

03:10 Getting Started Tools

04:53 Is AI Mainstream Yet

07:40 Workplace Copilot Arrives

08:26 Corporate Search Wins

11:05 Beyond Search To Toil

12:26 New Work Contract

15:57 Agents In Real Work

18:57 Quick Agent Examples

21:12 Dead Internet At Work

25:13 Human Flourishing Line

26:55 True But Not Real

27:59 Real Versus True

28:49 Creation Needs Loops

29:41 AI Empathy Skepticism

30:29 Inside Claude Safety

33:13 Personalized Agent Harness

33:55 Scout Runs Your Day

35:43 Always On Work Culture

39:03 Agents In Chat Channels

43:31 Gatekeeping And Triage

48:51 Productivity Versus Impact

49:46 Burnout From Orchestration

50:49 Redefining Productivity

54:20 Closing Thoughts And Wrap

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

ai.u crew catch up on AI industry news, starting with Apple’s WWDC (June 8), which they found comparatively lackluster versus OpenAI, Google I/O, and Microsoft Build, though Travis notes Apple’s selective, trust- and creator-focused approach, local/on-device models, and features like a new Siri AI app with screen awareness and conversation syncing. They discuss the cost of inference, ecosystem lock-in, rumored foldable iPhone hints in iOS 27 betas, and other rumored hardware (e.g., AirPods with a camera). The hosts then cover Anthropic’s Claude Fable 5, a consumer version of Mythos with safeguards, which was removed after about three days following U.S. federal national security concerns and jailbreak claims; they cite reported capabilities like migrating a 50M-line Ruby codebase in a day and debate safety vs marketing and access control. Finally, they discuss Microsoft Scout as an enterprise OpenClaw agent tied to Work IQ/Office Graph, highlighting chat personas, automations (daily briefs), co-create workspaces, and agentic workflows that feel “magical” and augment human work.

00:00 Show Intro and Life Updates

01:25 News Catch-Up Agenda

02:31 WWDC First Impressions

05:00 Siri AI and iOS 27 Rundown

07:51 Apple’s Human-Centered AI Angle

11:15 Local Models and Inference Costs

21:18 Foldable iPhone Rumors

22:41 Ecosystems and the Next AI Battle

25:26 OpenAI Losses Surge

26:11 Apple Siri Threat

26:59 Jarvis App Control

29:00 Apps Become Verbs

31:29 Anthropic Fable Pulled

33:16 Fable Power And Cost

34:31 Safety Versus Hype

39:09 Fable UX And Routing

43:56 Microsoft Scout Agents

49:22 Scout Features Breakdown

53:27 Closing Thoughts

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Ryan, Kevin, and Travis recap Microsoft Build 2026 (June 2–3 in San Francisco), describing it as Microsoft’s “coming out party” as an independent AI platform spanning models, custom silicon, an “agent OS,” a new assistant (Scout), and a full developer stack positioning Windows and Azure as the home base for the agentic era. Key announcements include a developer-optimized Windows with built-in Linux containers and two local Windows AI models (Aon 1.0 Instruct and AI 1.0 Plan), Nvidia ARM-based Surface RTX Spark Dev Box and Surface Laptop Ultra capable of running ~120B-parameter models locally, Project Solara for agent devices (including an AI-enabled badge concept), seven in-house MAI models, Scout as a proactive M365 assistant, execution containers for sandboxed agents, Majorana 2 quantum chip updates, Foundry as an end-to-end agent platform with “IQ” data tooling, GitHub Copilot app, open governance frameworks (Assert), and Microsoft M-Dash for agentic threat hunting.

00:00 Welcome to Build 2026

01:19 Big Picture Recap

02:46 Reactions and Takeaways

05:48 Windows Goes Developer First

07:42 Local Models and Control

11:11 AI Data Center at Every Desk

14:04 Surface RTX Spark Dev Box

18:25 Unmetered Intelligence and Token Costs

22:03 Project Solara Agent Devices

26:55 Jarvis Everywhere Vision

31:19 Privacy Walled Gardens and Trust

38:23 Solaris Three Pillars

39:33 Agents Everywhere Future

40:12 Seven New MAI Models

42:05 Frontier Tuning Explained

47:21 Satya Vision Ecosystem

52:26 Scout Autopilot Assistant

01:01:12 Execution Containers Security

01:02:39 Majorana 2 Quantum Leap

01:06:54 Day Two Foundry IQ

01:11:23 Governance And Mdash

01:15:17 Wrap Up And Takeaways

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Ryan, Kevin, and Travis recap Google I/O 2026 announcements, emphasizing Google’s rapid AI push across its products. They discuss Gemini 3.5 Flash as a fast, lower-cost frontier model optimized for long agentic tasks, rolled into Search and YouTube at scale, and Gemini Omni/Omni Flash for multimodal creative generation using models like Veo and Imagen. Google Search is overhauled into a multimodal, generative, interactive results experience, raising concerns about provenance and web economics as answers bypass creators. They cover Gemini Spark, a 24/7 cloud-based consumer agent integrated into Chrome and Google apps, plus daily briefings and background “information agents.” Other topics include a universal cart via a commerce protocol, SynthID watermarking with major partners, and announced Android XR audio/display glasses with partners like Warby Parker and Samsung.

00:00 Welcome and agenda

01:36 What Google I/O is

03:06 Gemini 3.5 Flash debut

04:21 Speed cost and rollout

05:59 Google comeback narrative

09:22 Omni Flash creative suite

12:55 AI video realism concerns

14:54 AI Search overhaul

17:11 Trust provenance and web economics

22:18 YouTube summaries and Ask YouTube

24:29 Gemini Spark personal agent

26:52 Ultra plan lock in and privacy tradeoffs

35:05 Convenience vs Societal Costs

35:46 Pandoras Box and Downstream Risks

36:37 Humans Leaving the Value Chain

39:09 Agency vs Algorithmic Feeds

41:16 Gemini Daily Briefing

42:08 Antigravity and Developer Orchestration

43:09 Information Agents and Signal vs Noise

50:44 Universal Cart and Commerce Protocol

52:52 Will Merchants Opt In

55:14 Google Product Longevity and Compute Limits

57:09 SynthID Watermarking and Truth Concerns

01:00:39 Android XR Glasses and Attention Economy

01:05:45 Who Actually Needs XR

01:07:44 Wrap Up and Listener Feedback

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Ryan, Kevin, and Travis discuss how impossible it is to keep up with AI’s pace and use recent OpenClaw updates to illustrate what’s happening at the “tip of the spear.” They recap an OpenClaw community-driven plugin architecture overhaul that caused short-term instability but created a smaller stable core with extensible plugins. They cover new voice interaction options, including Discord real-time voice features (buffers, barge-in detection, echo control), bringing agents into Google Meet via Twilio dial-in, and broader implications of voice and multimodal “thinking machines” interaction models. The hosts explore cognitive debt/coherence challenges as AI builds faster than humans can comprehend, and highlight OpenClaw’s security hardening (1,300 advisories processed) plus major memory upgrades that create structured person cards and a wiki-like knowledge base, raising governance and compartmentalization concerns for enterprises. They also note improved commitment tracking, self-modifying/self-building capabilities, and auto-generated skills.

00:00 Welcome Back Setup

00:55 Why Youre Behind

01:13 OpenClaw Overview

02:22 Community Moves Fast

03:04 Plugin Overhaul Fallout

04:25 Stable Core Plugins

05:39 Pick Your AI Strategy

08:48 Cognitive Debt Explained

10:26 Daily Reps Mindset

12:07 Voice Comes to OpenClaw

13:04 Discord Voice and Meet

18:21 Metacognition Modalities

21:30 Do You Need Code

26:09 Voice vs Text Context

29:35 Thinking Machines Tease

29:39 Interaction Models Demo

30:53 Voice Latency Tradeoffs

33:27 Conversation Cues Vision

36:39 OpenClaw Security Hardening

38:31 Memory And Knowledge Base

40:04 Enterprise Governance Dilemma

45:21 Corporate Brain Example

48:59 Auto Commitments Heartbeat

51:45 Stability Updates Skills

52:44 Wrap Up And Thanks

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com

Ryan, Kevin, and Travis return for a news roundup and announce Kevin’s resignation from Microsoft after getting funded for a startup applying “human-led, agent-operated” AI to small, underserved mining operations (Kai Radian). Listener feedback highlights Tauric Research, a multi-agent financial trading framework with analyst, research, trader, and portfolio-manager roles, raising concerns about overconfidence and automation. They discuss Anthropic previewing Claude’s “dreaming” memory consolidation and frame “harnesses” as orchestration/scaffolding that improves signal, manages context, and mixes deterministic workflows with LLM judgment, including models checking each other. They cover Anthropic using SpaceX’s Colossus I compute in Memphis to boost Claude Code limits, Mozilla using Claude Mythos to find 271 bugs in a month, reports that 39% of new podcasts are AI-generated “slop,” and a four-year Actors Guild deal adding AI protections for voice/likeness and writers’ rights.

00:00 Welcome Back Updates

00:52 Kev Leaves Microsoft

01:37 Mining Startup Vision

04:18 Mine Talk Banter

04:48 Listener Feedback Trading Agents

09:11 Claude Dreaming Memory

12:58 What Is A Harness

17:57 Determinism Vs Judgment

20:58 Anthropic SpaceX Compute

26:54 Mythos Finds Firefox Bugs

30:22 AI Security Attack Vectors

31:21 AI Podcast Flooding

34:33 AI Slop in Communities

38:25 Brain Atrophy and AI

39:51 Actors Guild AI Protections

44:31 Oscars and Human Eligibility

49:07 Democratizing Creativity

52:46 Human Storytelling Matters

54:08 Wrap Up and Feedback

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aiunprompted.substack.com