Creator perspective · MoneyElena Verna believes AI's most interesting effect on organizations is not replacing jobs but separating impact from headcount. Right now, how much you can ship determines how much impact you can drive, breaking the old reports-equals-impact equation. She urges everyone to think about this for themselves.
hedged
Lenny's Podcast · The rise of HI-ICs | Elena Verna (Lovable)
Creator perspective · MoneyJeff Smith argues that natural language to SQL is amazing but dangerous, because LLMs frequently run queries that look like they return the right data but don't. A SQL query that compiles and returns data does not mean it asked what you thought you asked. He would not trust Codex, Claude, Gemini, or any of these to decide what business branch to invest in based just on a prompt and a returned result.
hedged
AI Engineer · Your AI Agent Wants Your Data. Here's How to Lock It Down - Jeff Smith, Oracle
Creator perspective · MoneyCharlie Holtz rejects the 'software factory' framing for AI-assisted development, arguing it is the wrong way to think about the new tools. He points out that the 'feature factory' idea was tried ten years ago and failed, and that managing swarms of agents as factory line managers pushing buttons to pump out features is a failed model. In his view, working as a line manager in a 'dark factory' produces soulless, uncrafted, joyless software. He prefers an orchestra metaphor: a conductor waves the baton and a team of agents works in an intermingling of humans and agents.
hedged
AI Engineer · Orchestras, Not Factories: How the Fastest Builders Work — Charlie Holtz, Conductor
Principle · MoneyMasaba Gupta argues that the best and most unforgettable brands of the future will be based on culture and community, not product. She notes that brands today offer to be part of consumers' lifestyles instead of just saying 'here's a product, this is what it costs, this is what it does, you can buy it.' She adds that somebody who has a shit ton of money doesn't necessarily have taste — you can't buy taste — and that AI can never have taste because AI is a cold companion, and for taste to be authentic there has to be warmth in it. She also says the most unforgettable brands are the perfect bridge between art and commerce.
hedged
Figma · Config India 2026: What makes the House of Masaba feel like Masaba
Principle · MoneyMasaba Gupta argues that every brand has a body language — the storytelling and language that must stay the same across everything — and that this body language cannot change. She says a brand's body language almost introduces itself before you can say anything about it, and that this is how you protect a brand as it keeps growing and becomes a larger commercial success.
Figma · Config India 2026: What makes the House of Masaba feel like Masaba
Principle · MoneyKeep your point of view in mind continuously, because it shapes every question we bring to AI — and the willingness to continue needs a point of view and the reason to create, keeping you going beyond the first result that looks good enough.
Figma · Config India 2026: When everyone can make anything, how far can we take it?
Principle · MoneyBefore answering a client brief, question what innovation actually means, because the brief may not contain the real answer — as TeamLab did when a client asked for an innovative vending machine and they started by asking what innovation is.
Figma · Config India 2026: When everyone can make anything, how far can we take it?
Principle · MoneyAI can help discover and explore options we have not considered before, and those options can also reveal gaps in our original thinking.
Figma · Config India 2026: When everyone can make anything, how far can we take it?
Principle · MoneyWith AI, we can generate images and turn an idea into something tangible much faster than before — and more people can now make anything that was unreachable, so when everyone can make anything, the question becomes how far we can take it.
Figma · Config India 2026: When everyone can make anything, how far can we take it?
Principle · MoneyIdo Salomon argues that Warcraft-style complex UIs can be daunting for non-technical users, so a lower-denominator interface is needed now that agents are more capable — trying to find what makes stuff accessible.
AI Engineer · We're the bottleneck, but we don't have to be — Ido Salomon, MCP Apps
Principle · MoneyDiscord users are going to use Discord, Reddit users are going to use Reddit, and you basically have to do all of the above — just be where they are. Being everywhere was a lot of work but it worked out and was the right answer.
Latent Space · Supabase’s Plan for a Billion AI-Built Apps
Creator perspective · MoneyKevin Hou argues that scaling a product with model intelligence is difficult because it's hard to take away something users love and are familiar with in order to lead them down a potentially better path. He recounts that when chat was removed from Windsurf, users yelled at the team, but multi-step agentic research and execution became the new paradigm, and models advancing make the new paradigm better over time.
hedged
AI Engineer · Get Out of the Model's Way — Kevin Hou, Google Antigravity
Principle · MoneyRoland Gavrilescu defines an agent recipe as something that enables you to create reproducible frontier AI systems — a moat that keeps getting better over time and is not tied to any platform or provider.
AI Engineer · The Loop Is the Product — Roland Gavrilescu, Introspection
Principle · MoneyRoland Gavrilescu explains that when you put strong signals and verifiable work at the other end, you get workers or Claude Code agents. He endorses the principle that the quality of the signal determines the success rate of the loop, and the quality of the verifier calibrates whether that success is actually correct.
AI Engineer · The Loop Is the Product — Roland Gavrilescu, Introspection
Principle · MoneyRoland Gavrilescu endorses 'system distillation is the mode' — the ability to understand what went well and wrong in the first loop and know how to process that in the second one.
AI Engineer · The Loop Is the Product — Roland Gavrilescu, Introspection
Principle · MoneyRoland Gavrilescu's core principle: the loop is the product — the recurring agent loop, not a one-shot model call, is what constitutes the product.
AI Engineer · The Loop Is the Product — Roland Gavrilescu, Introspection
Creator perspective · MoneyRoman Ugarte describes the ultimate vision of GrokBot as having a team of AI bots that help you with your job and your life. AI teammates should feel autonomous: you can steer them in various ways but don't have to micromanage them. GrokBot should be the way you delegate a large portion of your day-to-day work so you can focus on higher-leverage things, and similarly the way you delegate low-leverage parts of your personal life.
Also discussed by Jediah Katz on AI Engineer (unclassified source — not counted until classified). Not counted as independent sources.
Lenny's Podcast · How we built Grok Bot in a month | Roman Ugarte (SpaceXAI)
Principle · MoneyIrina advises against a single vague metric like 'agent quality' because it hides interesting failures. Instead, use a balanced scorecard; the exact formula matters less than the shape of the scorecard. The scorecard also prevents the auto-research agent from gamifying the system.
AI Engineer · Long-Horizon Agents Need Experiments, Not Just Prompts — Erina Karati
Principle · MoneyJemima Abu defines accessibility as the practice of building products that people can use regardless of disability, and argues that a product people with motor, cognitive, or auditory disabilities cannot use is a useless product. She notes that while the audience may be sighted, for some people having their eyes shut is a permanent state.
AI Engineer · Teach Your AI Agent to Build Products Everyone Can Use - Jemima Abu, CAIS
Principle · MoneyWorkflow automation tools exist but rarely handle real-world complexity and ambiguity; a fixed flowchart breaks at the first wrinkle. Giving an agent the whole task and waiting for the next model to make it reliable is the wrong approach (the 'more agency' framing). The reflexive answer inside regulated firms — less agency, more rules, lock it down and hard-code every path — is also the wrong approach. In regulated verticals, reliability comes from where you place the boundary between deterministic parts and parts that genuinely need judgment.
AI Engineer · Reliable, Smart Workflows for Finance, Wealth and Insurance — Karthic Rao, Perpetual