Recommendation · MoneyZach Lloyd advises that most organizations can build a simple version of a software factory easily, but building one that actually scales requires a lot of supporting infrastructure you end up wanting — so he recommends focusing on your own product rather than building that infrastructure yourself. He frames building scaled factory infrastructure in-house as an anti-pattern for most organizations, while noting Uber built an internal version demonstrating large organizations can do it.
hedged
AI Engineer · Software Engineering Is Becoming Factory Engineering — Zach Lloyd, Warp
Recommendation · MoneyJeff Smith argues that the best place to implement significant database security is at the database itself. If security is built only into tools, a zealous engineer can pick up another tool, point it at the database, and go straight in. Putting security at the database layer means more times than not you're going to be okay.
AI Engineer · Your AI Agent Wants Your Data. Here's How to Lock It Down - Jeff Smith, Oracle
Recommendation · MoneyJeff Smith recommends giving LLM agents read-only access to the database, but stresses that read-only alone is not enough. Even with read-only access, agents may still read data you don't want them to see, so you must also lock down what they can do and see once they're inside.
hedged
AI Engineer · Your AI Agent Wants Your Data. Here's How to Lock It Down - Jeff Smith, Oracle
Recommendation · BodyBakken endorses running fast enough to get a real running effect in terms of running economy, but not going too hard and incurring the negative effects of going above threshold.
Garage Strength · The Genius Who Solved Endurance Training w/ Dr. Marius Bakken
Recommendation · MoneyRichmond Alake recommends that the audience implement agent memory within their agents, and that memory be treated as something to structure and model deliberately — not an afterthought.
AI Engineer · Everyone Has the Same AI Models. Here's What Sets Agents Apart - Richmond Alake, Oracle
Recommendation · MoneyRichmond Alake recommends engineers head toward building or adopting an intelligence plane for their agentic systems, noting some audience members are already implementing intelligence-plane components (knowledge, memory, governance, continual learning) and others will start after the talk. To differentiate agents built on the same models, he says to build an intelligence plane — knowledge, memory, governance, and continual learning — because information alone doesn't equal intelligence.
AI Engineer · Everyone Has the Same AI Models. Here's What Sets Agents Apart - Richmond Alake, Oracle
Recommendation · MoneyRichmond Alake recommends that developers and engineers of agent systems own their own memory and engineer the capturing of traces, building it into a system that allows continual learning to be possible.
AI Engineer · Everyone Has the Same AI Models. Here's What Sets Agents Apart - Richmond Alake, Oracle
Recommendation · MoneyBuild a company harness by running AI over your documented async history so it can extract and formalize your decision-making principles; this is where Supabase is investing most in AI usage.
Latent Space · Supabase’s Plan for a Billion AI-Built Apps
Recommendation · MoneyRoland Gavrilescu says once you validate that you have great taste and your users believe you have great taste as well, that's when you promote to the next version of an agent recipe. The secret is you keep doing this loop over and over again — continuously codifying your taste into an agent that can reproduce the same service or product for other people. The secret of building good loops is whether someone can iterate on your system the way you would.
AI Engineer · The Loop Is the Product — Roland Gavrilescu, Introspection
Recommendation · MoneyLakshya A. Agrawal advises bringing actionable side information and surfacing as much domain-specific information as possible to optimizers, because optimizers of the future will be able to work with it.
AI Engineer · Beating RL With Reflection: GEPA and Optimize Anything — Lakshya A. Agrawal, GEPA
Recommendation · MoneyLakshya A. Agrawal advises using a language model or agent to reflect on the entire rollout trace — what worked and what did not — instead of relying only on a zero-or-one reward signal.
AI Engineer · Beating RL With Reflection: GEPA and Optimize Anything — Lakshya A. Agrawal, GEPA
Recommendation · MoneyRoman Ugarte recommends delegating day-to-day work tasks to AI bots to focus on higher-leverage activities, and delegating completely — trusting the bot to finish without oversight rather than partially supervising. He argues that when you only 90% trust a teammate you still carry the mental load, so the aim is 100% trust to truly delegate.
Lenny's Podcast · How we built Grok Bot in a month | Roman Ugarte (SpaceXAI)
Recommendation · BodyThe creator advises prioritizing the face pull before leaving the gym, an exercise most people relegate to last or skip entirely, because what you do after your workout determines whether it was great or a missed opportunity.
ATHLEAN-X™ · Do This AFTER Every Workout (NON NEGOTIABLE!)
Recommendation · MoneyHassan El Mghari recommends having an expert human in the loop when building AI agents in specialized spaces, because domain expertise catches what engineering checks miss.
AI Engineer · Right Answer, Wrong Reasons: AI Forecasts Costco - Hassan El Mghari & Omar El Mghari, Together AI
Recommendation · MoneyHassan El Mghari recommends having models write structured data (like formulas in JSON) instead of having them do the math and do everything all at once.
AI Engineer · Right Answer, Wrong Reasons: AI Forecasts Costco - Hassan El Mghari & Omar El Mghari, Together AI
Recommendation · MoneyAamir Shakir recommends first making search really good before scaling up your model or data.
AI Engineer · Better Search Took This AI Agent From 36% to 98% Accuracy - Aamir Shakir, Mixedbread
Recommendation · MoneySnyder recommends that founders be the ones doing the selling, especially in a startup during zero-to-one product discovery. For non-founders doing zero-to-one work, he recommends focusing on trying to sell the thing and learning through selling. Getting lost in zero-to-one work is a real risk that selling-focused learning protects against.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Recommendation · MoneySofia Jakovcevic recommends closing the loop between what coding agents produce in the IDE and what is actually happening in the observability backend, by connecting coding harnesses and coding agents with what is actually happening in the production database. This gives agents full understanding of what changes are actually making.
AI Engineer · Your Agent Fix Just Broke Something Else - Sofia Jakovcevic, Arize AI
Recommendation · MoneyThe first stage of building a personal brand is to do something epic — in terms of volume or a quality outcome — before anything else.
Alex Hormozi · How I Would Build a Personal Brand in 2027 (If I Had to Start Over)
Recommendation · MoneyFind big moments and big external goals that you can drive towards — things that scare you and where you actually have real stakes — as belt moments.
Alex Hormozi · How I Would Build a Personal Brand in 2027 (If I Had to Start Over)