Principle · MoneyMolly Graham's Legos framework: when change or growth comes, the most important thing is to give things away and evolve with the company. When someone takes your project, the kindergartner instinct is to shove them away, but letting go means you can build something new, start from scratch, and learn. She advises fully separating your mind from handed-off work, trusting its future belongs to the receiver, and notes the huge freedom this brings. She jokes the best way to hand off is to 'chuck them at their face and run in the other direction' rather than carefully handing over with an instruction manual, and warns our tendency toward control is the thing to override.
hedged
Lenny's Podcast · Molly Graham: The grief, burnout, and opportunity hiding inside the AI transition
Principle · MoneyMolly Graham says your main job in the face of rapid growth and change is to make yourself irrelevant, because it is the only way you can be prepared for whatever is coming next. Inside a rapidly growing company, assume your role will transform faster than ever before and prepare accordingly. She suggests a thought experiment: what would you do if you believed your job was always going to exist but would look completely different every six years?
Lenny's Podcast · Molly Graham: The grief, burnout, and opportunity hiding inside the AI transition
Principle · MoneyMolly Graham says the Legos message is fundamentally about change and that a lot of hard stuff happens to you when you go through it. She urges leaders to normalize the emotional roller coaster for their teams instead of leaving people feeling broken, and says you can't lead right now without talking about grief and normalizing it. Feeling a wild roller coaster of emotions during change means all the right things are happening, not that something is wrong — these emotions are just your body and identity reacting to change. Change is scary and hard, but it doesn't necessarily mean the world is ending, and there's opportunity on the other side.
hedged
Lenny's Podcast · Molly Graham: The grief, burnout, and opportunity hiding inside the AI transition
Principle · MoneyMolly Graham tells managers and leaders to understand that they are role models — what they do and how they behave shows their team what good looks like. How you use AI trickles down, and your AI behavior sets the standard your team follows.
Lenny's Podcast · Molly Graham: The grief, burnout, and opportunity hiding inside the AI transition
PrincipleSophia Bush says that when you've been hurt, the impulse to retaliate comes from an older version of yourself shaped by your upbringing, and that the work you've done on yourself should override it. She frames it as having done too much work on herself to hurt someone when she's been hurt.
Call Her Daddy · Sophia Bush: Brooke Davis, Breakups & Beginning Again
PrincipleSophia Bush says that when recovering from a bad relationship or divorce, you have to untangle what you got wrong, look at why you wanted it in the first place, and reconnect with what you believe in before entering a new commitment.
Call Her Daddy · Sophia Bush: Brooke Davis, Breakups & Beginning Again
PrincipleSophia Bush says she learned to be open enough to say that something isn't minor and that she is not okay, rather than telling herself to buck up because others have it worse. She frames living the rest of her life differently — not looking fine while being miserable — as a deliberate decision.
Call Her Daddy · Sophia Bush: Brooke Davis, Breakups & Beginning Again
Principle · MoneyWhether a PM should row (do the work) or steer (direct the team) depends on the type of product and the phase of the product. Shift from rowing to steering when the project phase changes, because that is what makes the project go faster, and focus on whatever unblocks the team at that point in time.
Lenny's Podcast · Roles aren't converging—they're expanding | Tamar Yehoshua (Atlassian CPO)
Principle · MoneyAll of the AI-tool productivity gains work only if all the AI tools understand the context of your organization, which at Atlassian is built on the teamwork graph. Context graphs and an 'enterprise brain' are very real — you can now see the full context of your organization through tools, and Atlassian built a 'teamwork graph' that shows the full context of the organization.
Lenny's Podcast · Roles aren't converging—they're expanding | Tamar Yehoshua (Atlassian CPO)
Principle · MoneyOrca's principle is to be model agnostic and avoid taking a dependency on a single provider or harness.
Latent Space · Agents Prompt People: Using Agents for Long-Horizon Financial Work
PrincipleWhen a PR should have been escalated but got approved, that's a miss — it becomes a new rule. The re-tag miss was not a failure of judgment; it was a gap in the rules, a corner case nobody had written down, so it became a new rule. That's how the rules get better: the misses get found and become rules.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleRobinhood's answer is structured as three duties and three owners: scripts prove what can be proven, an agent judges what needs context, and people own the calls only they should make. They took a hybrid approach — scripts for the facts and AI for judgment — and did not want AI reviewing every change when some questions already have an authoritative answer. Three things to listen for in the talk: scripts for facts, an agent for the judgment, and people for the protected calls.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleA schema-changing PR can pass CI, tests, and the linter and still break downstream consumers, freshness SLAs, or expose data that was never supposed to be reachable. The gap between 'CI passed' and 'this change was actually safe' is the gap the data platform team has to close every day.
hedged
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleWhen a PR arrives at machine speed and reviewers lack context, use a hybrid reviewer rather than blocking on human review. Relying on a single review mechanism (all-human or all-automated) fails when PR volume outpaces reviewer capacity. Scaling PR volume to machine speed without scaling reviewer context creates a safety bottleneck in code review.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleWhen a change is scary or high-risk, always flag it for human review rather than clearing it. The flag is a signal for human judgment, not a final decision. Clearing scary changes automatically without flagging them leads to unsafe merges — this is one of the three closing lessons.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
Report · MoneyWith managed agents, running an agent boils down to a single API call. That call gets you the sandbox and the anti-gravity harness built out of the box. Jack endorses that you don't have to worry about managing infrastructure or scale — for more agents you just make more API calls.
AI Engineer · Build Your Own Team of Managed Agents - Jack Wotherspoon, Google DeepMind
Creator perspective · MoneyJack describes two problems with running agents locally: verification on your own laptop steals the browser and runs things on your machine, and a laptop simply doesn't scale for running agents. He also notes you don't want a customer-facing app running on your personal subscription. He endorses running verification in the cloud to remove that pain, and endorses managed/cloud agents as a great way to scale agent setups.
AI Engineer · Build Your Own Team of Managed Agents - Jack Wotherspoon, Google DeepMind
Principle · MoneyWhat used to be expensive is now cheap, and if you don't update quickly on that shift, you fall behind. Over-indexing on a technology decision made in June when it's now October means ignoring what feels like ten years of technological advancement.
AI Engineer · Why Waiting for the Next AI Model Can Pay Off - John Stecher, Blackstone & Scott Worland, Norm AI
Principle · MoneyRamakrishnan Lokanathan advises that when a model upgrade ships, you keep testing regression cases, known tasks, and past failures, because a model that is extremely better than the old one could very well be incapable of something the old one did very well, like setting a reminder.
AI Engineer · If Your AI Scores 100%, Your Eval Is Wrong - Ramakrishnan Lokanathan, eslams.com
Principle · MoneyIf an eval scores 100 out of 100, the eval is wrong: it signals the team is not ambitious enough and not testing enough, so the test must be made harder and pushed beyond current model capability to stress-test rather than make things easy.
AI Engineer · If Your AI Scores 100%, Your Eval Is Wrong - Ramakrishnan Lokanathan, eslams.com