Principle · MindWith Bill, Theo tried to create intimacy immediately; with Mike, he let it build incrementally. Choosing when to share something can either deepen the conversation or kill it completely. You know you've shared the right amount when the conversation keeps going versus getting shut down.
Charisma on Command · When Someone Intimidates You, Do This...
Creator perspective · MindBy repeating what Mike said back to him or paraphrasing it, Theo makes Mike feel like he's not only being heard, he's being understood. In Theo's case, reflective listening helps him stay present in the conversation and focus on Mike rather than on himself.
Charisma on Command · When Someone Intimidates You, Do This...
Creator perspective · MindThe narrator describes two complementary moves: zooming out a level when a conversation gets too intense takes the pressure off without shutting the topic down (as when Theo asked to go back to the quarters), and zooming in gradually — asking a more personal question only after the person has engaged with wider topics, watching their body language for passion — is how Theo and Mike got to a deep place early.
Charisma on Command · When Someone Intimidates You, Do This...
Principle · SocialSean builds rapport with Will through sharing vulnerability himself. When we share something vulnerable with someone, we're usually looking for a particular response, whether it's validation or reassurance or empathy. Modeling positive vulnerability for someone like Will is one of the most generous things you can do in a relationship, but it's also dangerous. If you want to practice sharing vulnerability, share something that you're comfortable getting nothing back from. If you need someone else to reassure you or open up in return or respond a certain way, you're putting your emotional state in their hands, which can make it much harder to stay grounded. Sean can respond calmly to Will's insensitive remark because he didn't start talking about his wife looking for something from Will.
Charisma on Command · The Psychology That Disarms Emotional Games
Principle · SocialWhen someone provokes you, it's nearly impossible to not react in the moment — that's outside of your control. What is in your control is taking time to process things after the fact. After someone gets under your skin, go home and process what just happened. Processing not only helps you make sense of it, it also increases your empathy for the other person. Grounded charisma comes from slowing down and looking for the underlying pattern, then figuring out what the actual conflict is instead of just reacting to what seemed to happen on the surface.
Charisma on Command · The Psychology That Disarms Emotional Games
Principle · MoneyUse LLMs to help generate plans, but keep the plans themselves deterministic and explainable to the customer — especially for recovery scenarios. Never use a probabilistic mechanism for data recovery in a cyber scenario, because you cannot afford to guess.
The Official SaaStr Podcast: SaaS | Founders | Investors · SaaStr 876: Shipping Enterprise AI Agents with the CPOs of Rubrik, Glean, and Harvey
PrincipleCharlie Guo warns that giving a Goal vague or non-verifiable completion criteria leads the model to technically complete the task but not in the way you expected; be really specific with the criteria.
AI Engineer · Cooking with Codex — Charlie Guo & Gabriel Chua, OpenAI
PrincipleGabriel Chua advises against pasting sensitive credentials like deployment keys directly into chat — ask Codex for a secure way to set them instead — and warns that failing to provide credentials or permissions for external services like Vercel and Convex blocks the agent's critical path and stalls the whole task.
AI Engineer · Cooking with Codex — Charlie Guo & Gabriel Chua, OpenAI
PrincipleGabriel Chua recommends chaining Codex primitives — automations, plugins, inter-thread communication, and work tree threads — together when you want to build a software factory, a system that builds individual pieces of software on an assembly line rather than a single piece of software.
AI Engineer · Cooking with Codex — Charlie Guo & Gabriel Chua, OpenAI
ReportCharlie Guo explains that Codex hooks are configured via config.toml in the Codex home directory, and recommends using them to introduce deterministic behavior at checkpoints: logging conversations, scanning inputs/outputs for security (e.g., detecting pasted API keys), checking tool calls against an approved whitelist, and running custom validation checks when a turn stops.
AI Engineer · Cooking with Codex — Charlie Guo & Gabriel Chua, OpenAI
ReportThe FBI built a secure messaging app and marketed it exclusively to organized criminal groups (Operation Trojan Shield / ANOM), leading to arrests in Australia and the US, with most of those criminals indicted and imprisoned. The host uses this as the most illustrative story about creepy VPNs, noting that VPNs and secure messaging tools can be compromised by the very entities that market them as safe — and that free VPNs on app stores are similarly suspect ('if it's free, there's nothing free').
The a16z Show · Rebuilding the Internet for Privacy | Barrett Lyon on DoxxNet
Creator perspective · MoneyKent C. Dodds's central thesis: since AI agents make implementation easy, the scarce, valuable skill shifts to deciding what to build and whether it is worth building. He notes the entire workshop depends on this slide.
AI Engineer · Build the Right Thing: Product Engineering (Part 1) — Kent C. Dodds, EpicProduct.engineer
Principle · MoneyKent C. Dodds warns that when people get an idea, they tend to jump straight into over-specifying the solution — mobile app, website, integrations — before the potential user has even agreed the problem is worth solving. He endorses validating the problem with users first, before committing to a specific solution shape.
AI Engineer · Build the Right Thing: Product Engineering (Part 1) — Kent C. Dodds, EpicProduct.engineer
Principle · MoneyMulti-agent systems add handoffs — one agent decides to pass work to another — adding another level of complexity. Check whether the triage agent routed correctly and whether the specialist agent handled the handoff gracefully. Each layer adds new ways things go wrong, and non-determinism means you need more evals because errors can cascade.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Creator perspective · MoneyAgents making a series of tool calls and decisions have increasing chances of mistakes at every step in the chain, so you must judge intermediate steps as well. Check whether your agent is using the right tool, passed the right arguments, and knew when to stop. Getting stuck in loops is a real failure case that must be evaluated for.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Principle · MoneyBinary scores are easy for LLMs to produce, but code evals don't produce explanations because they're just code. A real judge explanation (a travel planning agent told it gave destination info but no cost estimates for a budget-travel question) makes the eval actionable: you have a concrete failure and know what to fix in the prompt, turning evals into a debugging tool rather than just a scoreboard.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Report · MoneyLaurie Voss describes the full production loop: the application produces traces, online evals grade them and produce labels and explanations; eval labels feed into monitors that can alert the team when something goes wrong; the team investigates, finds failing traces, saves them as a regression dataset, improves the agent, runs an experiment to verify, and ships the fix.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Principle · MoneyCascading failures are worse than obvious ones: an agent asked about Tesla the car company returning a report on the 18th century inventor did every step 'correctly', so no unit test would catch it — you must read inputs and outputs.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Principle · MoneyOne important principle is to grade what the agent produced and not the path that it took. Checking that the agent followed a specific sequence of tool calls in a specific order is a common instinct that practitioners have found too rigid in practice — if you grade the path, you punish creativity, because agents regularly find valid approaches the eval designer didn't anticipate. Instead, grade whether the output had the right information and whether the final state matched what you wanted. The ticker check doesn't care how the agent found the data — it just checks whether the ticker is in the output, which is the right level of abstraction for most code evals.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Creator perspective · MoneyThe speaker described failure patterns visible only in traces: an agent writing a file without permission, the write silently failing, and then claiming the report was saved — producing a useless output referring to a nonexistent file; also an agent stuck in an endless loop of web searches.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI