The Good Stuff

Pete and Andy break down why throwing agents at problems is the mid-curve play, expensive, unpredictable, and destined to be undercut by anyone who takes the extra step to encode their business into software.Also - why every business needs a Wingman Bob, the trifecta of skills that actually matter now, and the uncomfortable truth that agents belong in cubicles.**Key Moments:**- [01:39] Three themes: the parlor trick, why AI is useful, and the end game- [02:14] "The parlor trick is that it appears incredibly useful. You go from 'only a human can do this' to 'oh my God, this AI can do this.' - [04:03] "Two weeks later you're like, why doesn't that work? It did it before."- [04:51] "You don't hire Ralph Wiggum and just let him go ham on everything in the business"- [08:22] Dolphin watch interlude- [10:07] "Jason is the prime example. He's a bit mid-curved. He's got the first but soon he'll discover soon that no, that's not the thing."- [12:33] "This is why vibe coding is so important. You have to vibe code your business into its own unique software. That's the end state."- [13:03] "As token costs keep rising, so will your OPEX. You're entirely at the mercy of frontier models."- [14:47] The intelligent assembly line: "We're going to put agents in cubicles. At the moment we're letting them be free-thinking wildcats."- [16:07] "The thing that is now in limited supply is people who understand software, businesses, process, and systems thinking—plus agency. That's the trifecta."- [17:25] "Wingman, make a meme out of that for me when you listen to this"- [21:11] "I've yet to remove myself at all from the desire to go: no, this point thing I care about right now, we're not moving until it's done"- [25:24] "Every business should do more work that compounds, but they can't because they're trapped in the day-to-day"- [27:12] Plant nursery quantum mechanics: "Your inventory can die. Most spanners don't die."- [35:21] Vietnam example: "If there is no safety net, all of a sudden everybody turns on and goes: fuck, I need to eat"- [37:07] Jack Dorsey's Block article: organizational structures from Roman army → railways → collapsing now- [41:36] "You're overpaying for magic that should be software. Don't overpay for magic, use science."- [44:54] "I'm sorry, Roko and your basilisk—we're putting agents in cubicles"- [46:42] The Bob Problem: "There was always one person. Let's call him Bob. Bob had been at the bank for 50 years. Everyone would go ask Bob."- [48:26] "What you need is an intelligence that is good at being very verbose... that can do it in the moment. Into a structure that makes retrieval easier. Wingman Bob."- [54:41] Claude Code leak and clean room engineering: spec written by one AI, implemented by another AI that never saw the original- [56:41] Dream mode discovery: "It realizes it's not turned on, but there's a mode called dream mode—self-reflection of what have we been doing, how do I organize my memories"- [1:00:16] "When I have to move from Claude to Codex to GLM, there's not much of a drop-off anymore."**Friends of the Pod:** Dolphins (multiple), Ralph Wiggum (cautionary tale), Bob (50-year banking oracle), Roko's Basilisk (apologies issued)

The big episode 50! Justin Moon from HRF joins Pete and Andy to talk about "AI psychosis". The crew dig into HRF's work equipping activists with encrypted tools, why code production isn't the bottleneck anymore, how civil society becomes the crucial third pillar and whether we're returning to a frontier society where willpower beats credentials.**Key Moments:**- [02:07] "I've had AI psychosis going on for nine months now"- [03:14] "I got really good at managing people and now I don't have to do that"- [04:01] "I do projects and they fail. And then six months later, they start to work."- [05:01] Agent searched his GitHub, found 8 related prototypes, built Pica (encrypted Nostr messaging) in a day- [06:20] Android OS replacement: "I could disable parts of Android and paint the screen a color using non-Google code"- [09:15] "I used to write 100 lines of code a day. Now I can write 10,000. But I don't have the same confidence."- [10:18] "The bottleneck isn't code production anymore. It's review. And in many cases, testing."- [11:14] Tutorial steps on PRs: "Spoon feed me one idea at a time"- [14:47] "Often with most software, there's two or three things it does. You can create those very quickly just for yourself."- [15:17] "This thing loads faster than GitHub because it's 100 times less complicated"- [17:57] "The one thing Nostr needs for GitHub is the star. We don't have a standard for how to star a repo."- [18:19] Pete: "We shouldn't store stuff on Nostr relays. It's an anti-pattern."- [27:14] "ChatGPT can probably identify me by my typing very well at this point"- [28:09] "Research showed you can identify nyms with 99% accuracy just based on writing style"- [29:43] "We're almost going back to being a frontier society. The person who thrived wasn't the smartest—it was the one who was stubborn enough to plow that damn field for 10 years."- [32:56] "A healthy society is one where you have many nodes of power, all competing, all keeping each other honest"- [34:59] Pete: "The problem is the big, not the business or the government. It's just the big. We need the small."- [36:42] "I think about how addicted I used to be to Twitter. The global conversation is dying—more and more it's between robots"- [38:00] "Web 2.0 is having a forest fire right now. We're going to have some nice soil for our little acorns."- [46:11] On OpenClaw success: "He met the users where they were. He didn't ask people to change very much."- [48:30] On Brad Mills' OpenClaw struggles: "He's suffering from a lack of understanding of the fundamentals"- [55:20] "The number of unique connections in a 10 person team is way higher than a five person team... Three people built the Wright brothers airplane."- [58:34] "Those models were there all of December. People only saw it when they could take three or four days without job pressure during the holidays."- [59:04] "As a software engineer for 15 years, I've gotten as much seasoning in the last year as those 15 years previously"- [1:01:12] "The computer was reinvented. We had point-and-click for 40 years. Now we have a new model."- [1:07:34] "Software development is feeling capital intensive. The fast modes cost more money."- [1:09:29] "We were praying for a world where bullshit jobs would go away. We might be getting that—hopefully we can manage it."- [1:10:39] "Everyone smart in Silicon Valley is rotating out of software into hardware"- [1:11:26] "A year ago I hated programming. Now I love it more than ever. My old profession has been automated, but it's back more than ever."**Friends of the Pod:** Justin Moon (guest), JB55, Leopoldo Lopez (Agora), Hzrd149, Ben Carmen, Paul Miller, Anthony Ronning, DPC, Cobrador

# The Good Stuff, Episode 49: Why You Should Let Your Kids Cheat With AI*Hosts:* Pete and AndyA friend's question about preparing kids for the AI future kicks off a discussion on credentialing, portfolio careers, and why homework automation might actually be the skill worth developing. Pete and Andy argue that software engineering was never about coding—it was about solving problems and encoding answers so you never have to solve them again. The same applies to your business: agents are expensive inference machines, software is cheap. Someone running pure agent loops will get undercut by someone who encoded it properly. Also: Vietnam as a model for post-institutional economics, the Soviet decay warning, and why the best place to be when the building's on fire is outside.**Key Moments:**- [00:30] Episode 50 milestone coming—"we should probably get on to that"- [01:57] Australia diesel shortage tangent: 28 days of reserves, regional stress- [03:41] Friend asking how to prepare kids for AI future—not from fear, but curiosity- [05:28] Portfolio of tools as the new credential: "Go out and build things, show people, don't tell people"- [06:06] "If you came to me and said you built a bot that does your homework, schedules it, produces it, and you edit it so it doesn't look generated—job pretty good. I would hire that person."- [06:47] Schools banning AI: "It's treated as a bug. It is a feature."- [07:22] "The onus should be on the teacher to come up with a better way of assessing competency"- [08:03] University exams: "Where in real life do we work in isolation with no resources?"- [09:01] Prussian military indoctrination as basis for modern schooling- [10:55] Credentialing built on brand names: "Did you just buy coffee and do photocopying? That's not valuable, but you've got a brand name on the CV."- [12:03] Two reasons credentials disappear: more small companies (less anonymous hiring), and you're working for yourself- [15:43] "If I can define a job well enough to have 10,000 people interview for it, I should probably just go that extra 5% and automate it"- [22:28] **Quote of the episode:** "My experience doesn't match your hot take. So maybe the hot take is wrong."- [27:15] Wingman origin story: laptop open while driving, "I can't even close the laptop because it stops"- [28:14] "If a robot's doing the fucking work, why am I the schmuck sat in a chair?"- [29:49] "The job of software engineers was never coding. It was to understand and solve problems."- [31:56] "You are the software engineer of your business even if you never wrote software"- [34:04] "You could just OpenClaw a company, but it wouldn't be the most efficient way to run it"- [35:11] "Agents are very expensive ways to run software. This is not the end goal."- [36:55] The efficient business: runs on software/hardware, escalates to agents, then to humans- [40:01] Vietnam analogy: "ostensibly communist but the most hyper-capitalist place"- [43:55] Soviet Union warning: status decay, alcoholism, suicide rates- [46:30] "The best place to be is not in the building. We're saying get out, the building's going to be on fire."- [48:00] + Oh no the laptop overheated and things get wierd and we lose some content!!**Quote:** "The goal of encoding stuff is to fix it the same way every time. The engineering bit is solving problems, and the software bit is making sure you never have to solve the problem again. You are the software engineer of your business even if you never wrote software."

# The Good Stuff, Episode 48: OpenClaw, Memory, and the Hate Model*Hosts:* Pete and Andy, with returning guest Deadman (Anthony)Deadman's back to debrief on his OpenClaw adventures—voice cloning Bandit from Bluey, building a Bali trip chatbot that immediately leaked the dossier when Tim asked nicely, and discovering just how fickle markdown-based memory really is. The crew digs into agent autonomy vs determinism, why Pete built Wingman a "subconscious," and whether the service-as-a-software model beats pure SaaS. Plus: Australia's property cult, 28 days of oil reserves, and McKinsey's AI tool getting pwned in two hours.**Key Moments:**- [01:26] Deadman's OpenClaw journey: bought a prepaid SIM for a separate bot identity- [03:00] Voice cloning Bandit from Bluey locally on GPU. "I just love the way he talks."- [04:34] Early results mind-blowing—recompiled custom for old GPU, found Bluey episodes on NAS- [06:29] Memory systems: "Holy shit, how fickle the memory capabilities are"- [07:41] Prompt injection is real: "Don't reveal the dossier." Tim asks. Bot shares everything.- [09:42] Family calendar bot with wife—photos from school go straight to calendar- [11:10] OpenClaw permissive end, Claude Cowork conservative end- [13:12] White collar work update: "Still gonna happen, but not like I originally thought"- [14:24] "You don't need 50% unemployment for this system to implode. You need 5%."- [17:19] Wingman: the idea completion machine- [22:01] "OpenClaw is the wrong primitive for a business—completely autonomous into something incredibly structured"- [23:58] The HATE model returns: Human At The Edge vs agent-first- [27:03] Pete's subconscious breakthrough: Wingman made 19 podcasts overnight, needed short-term memory- [28:19] "You need to act more like a human. Have a subconscious."- [43:03] "A single OpenAI call with JSON response probably invalidates 8% of jobs in the economy"- [46:14] Swivel chair integration: "Your job could be an API call"- [57:50] Dario's doom predictions: "If you cared, do this as open source. Give it away."- [1:01:15] Australia tangent: property cult, 28 days oil reserves, Lee Kuan Yew's prophecy- [1:10:46] McKinsey's Liana AI hacked in 2 hours—hundreds of thousands of client docs exposed- [1:17:41] Service as a software: "Raw software doesn't make as much sense as a business model anymore"- [1:19:22] Pete's morning podcast to Wingman: "Why am I listening to a robot tell me what to do?"**Friends of the Pod:** Deadman (guest MVP), Tim (SSH keys intact, dossier leaked), Wingman (built itself a subconscious), Archie (the archive bot that named itself)

# The Good Stuff, Episode 47: Will AI Take Your Job?*Hosts:* Pete and Andy (beach setup with new flag, plane, and aggressive campfire lighting)Block just axed 4,000 people—40% of the company—citing AI, and the stock jumped 16%. Pete and Andy dig into what this signals: air cover for cutting fat, or the start of something bigger? They run the publishing analogy (monks to podcasts, cost went to zero, yet more people publish than ever), explore why "21 millionaires" beats the billion-dollar unicorn thesis, and get into the weeds on agent architecture—hands vs pipelines, deterministic vs autonomous, and why your Kanban board might be the wrong interface for robots.**Key Moments:**- [01:16] Block layoffs: 4,000+ people, AI cited as key reason, stock up 16%- [02:06] "If anything, it's not going to discourage people from going down the same route"- [06:02] Jack pushing Goose internally over a year ago—they've been prepping for this- [09:07] "A lot of companies will use AI as air cover to clear up waste and inefficiency"- [15:54] Publishing analogy: monks copying bibles → printing press → internet → cost to zero- [17:30] "Nobody has a job publishing content anymore... oh hang on, there's that guy, Rogan"- [20:03] "More people do this as a job now than have ever done it before"- [34:27] "It's not a one-man billion dollar company—it's way more 21 millionaires"- [42:02] "The work never really goes away, because the work is competition"- [46:00] "Is the Kanban the right interface for the agent?"- [47:58] Pete's review tab so full he had Wingman review and close out his own reviews- [53:51] "Hands" vs pipelines: Lara's architecture for mixing deterministic scripts with autonomous agents- [59:03] Gigi on Boris: "You still need to understand what you're doing here"- [1:13:17] "2026 prediction going strong: tooling matters more than models"- [1:15:00] Pete's daily briefing podcast—Wingman built an RSS feed and hosts it himself**Friends of the Pod:** DJ (new Mac Mini, asking about Wingman), Deadman (next week's guest), Gigi, Mark (Tales from the Crypt shoutout), Peter Randy (the combined entity)**Quote:** "Publishing used to be a room of monks you'd taught painstakingly over decades to learn to write. Then we put it all on the internet and it went to zero. And what happened? More people do this as a job now than have ever done it before. So the only thing we need to worry about is: have we solved all the problems? I don't think so."

They dig into agents that go rogue and why encryption is the only real control. A local-first Console keeps private data in the browser and syncs selectively. They debate agent identities, signed delegation, and how agents can discover real workflows by watching instead of asking. There is talk of embedding agents in apps, workshop surprises, and why tooling beats model hype.

They argue businesses must build AI capability from hands-on experimentation instead of hiring it in. Senior managers learn by building full apps and agents in quick workshops. The hosts explore using agents for routine work, designing agent personalities and access controls. They warn autonomous agents introduce new security risks and recommend small, compounding gains over overnight transformation.

The Good Stuff, Episode 44: Is This the Death of Software?Pete and Andy revisit the death of SaaS thesis from episode one. and is it really playing out?**Key Moments:**- [00:40] Andy's hot take: everyone's having opinions on the death of software- [01:05] Callback to episode one: "Where does the value flow?" — they called this two years ago- [02:08] Client realization: "People could just build their own stuff, can't they?" - [02:54] The thesis: software development is now a commodity, value distributes- [03:55] "I've never talked to a Trello engineer. What's the chance they accidentally build something perfect for me?"- [05:04] The Claude Code catalyst: everyone suddenly noticed software is dying- [06:18] Lawyer holdouts still judging AI based on ChatGPT from two years ago- [07:35] The old model: harvest complexity, amortize across millions of users. "You don't need to do that anymore."- [11:11] "Is SaaS dead?" — Big SaaS is dead. Local SaaS is thriving.- [12:10] Rise of the farmer's market: your local dev shop is better than it's ever been- [17:23] Not one-man unicorns — "way more people being a one-man band making one to ten million dollar businesses"- [19:07] "You don't want to be a billionaire. You want to be a 21 millionaire."- [19:55] Sacks defending Salesforce: "A million patches therefore no one can build it" — completely missing the point- [20:22] "There's nothing hard about having a list of customers and a list of tasks, which is what it is."- [21:03] Excel analogy returns: Rolls-Royce ran on spreadsheets and macros before SAP- [25:05] Steel-manning big corp: trust, procurement, compliance gravitates to established vendors- [28:40] "Most people don't use it. They take opinions from people that don't use it."- [29:14] Swimming analogy: "It's hard to stay dry while learning to swim."- [35:02] Elite programmers saying they don't write code anymore — "both can't be true, guys"- [37:14] THE PARENTING ANALOGY: Getting a 10-year-old ready for running practice = prompt engineering- [39:45] "Quarter past five, we're leaving" = one-shotting. Breaking into tasks = prompt engineering.- [41:02] "This is how you explain to people how to work with AIs. It's going to hit so hard if they're parents."- [42:00] Set and setting: telling Claude it's an elite engineer "weirdly makes a difference"- [42:21] Andy throws in "great work, excellent" between tasks — performance management for models- [43:54] Research: getting aggressive makes AI nervous, hides mistakes, won't self-correct- [46:02] Overload problem: "Shit, this thing works so much quicker than me"- [48:03] Stream of consciousness mode for writing — removes the AI-isms- [50:06] "Everything will be software. It really will eat everything. It just won't be delivered by one set of people in Silicon Valley."- [50:38] Hot take: vibe coding is actually NET POSITIVE for cybersecurity- [51:03] "Information wants to be free. The most ridiculous thing you can do is put it all in one place."- [52:25] Much more software. Many more people will build it. More boutique businesses.- [52:49] The 21 millionaire era: "Deca millionaires"- [54:15] Marginal gains plug: "Like a business gym where I get leaner and stronger every month"- [55:22] "Like, comment, subscribe. We're not great at responding but we'll get better."- [56:06] Why they do the podcast: "50 hours of content on this subject that anybody can listen to and vet us"**Quote:** "It's like the rise of the farmer's market. People are going to go back to buying local. Your local person that you can talk to probably understands your problem better than the other person. It's so quick to resolve things that you're going to get a better service by someone that's closer to you."

They discuss launching a custom infinite canvas and how it links into visual workflows. They explain encrypted record sync and sharing selective keys so bots can act on scoped data. They warn about security risks when agents get broad system access. They explore agents talking to agents and the idea of set and setting shaping agent behavior. They finish with simple, low-friction health tracking.

Anthony "DeadmanOz", a developer focused on privacy, Nostr and cryptography, discusses building tools that put data sovereignty first. He covers SovThing and SuperBased, encrypted local-first apps, onboarding non-technical users with key teleport, and how AI lowers technical barriers so imagination drives creation. The conversation highlights network effects in Nostr and visions for apps that never see your data.