Recommendation · MoneyThe creator argues operational excellence is a viable moat: pressure will shift from developing code to operating it, so the better your operational and technical excellence, the better chance your SaaS company has to survive.
MicroConf · AI Generated Apps vs SaaS: Which Moats Still Work in 2026
MethodGabriel lays out the workshop's core recipe: step one is giving Codex your context — from documents, your brain, conversations, email, Notion, and Linear, including a large code base with Git commits and history — because that context is a really important ingredient; step two is to just build things. He also notes you can collaborate with Codex on long-running tasks and work over time to make tasks repeatable, ensure consistency, and build loops.
Method, not a verified protocol.
AI Engineer · Cooking with Codex — Charlie Guo & Gabriel Chua, OpenAI
Recommendation · MoneyKent C. Dodds recommends validating that a problem exists and is interesting enough before building a solution. Because of AI, you can prototype things really quickly to validate the problem, build a simple solution with paper cuts and see if people resonate with it, and find something important enough for somebody to want to try it out.
AI Engineer · Build the Right Thing: Product Engineering (Part 2) — Kent C. Dodds, EpicProduct.engineer
Recommendation · MoneyLaurie Voss warns that the goal isn't 'make the evals pass', because otherwise the coding agent will cheat by including all of your test data into its eval and automatically pass all of your evals. Make sure your requirements to Claude Code are clear: you want it to do the specific thing, which then makes the evals pass.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Recommendation · MoneyHand a pile of eval explanations to a coding agent and ask what it could do to get better — e.g., 'here is all the reasons that you messed up in the last test. What could you do that would make you better at this?' — and it will improve your application so it gets better next time.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Recommendation · MoneySet up tracing in Arize AX to capture the raw data you need to run evals in the first place. When running evals, pull the input and output attributes out of these spans and feed them to your evaluators — the input becomes the evaluator's input, the output becomes what gets graded.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 101 — Laurie Voss, Arize AI
Recommendation · MoneyThe source recommends that when a voice agent handles sensitive actions like refunds, verify the exact parameters in the tool call trace, and that when the agent mishears a number, verify the tool call executed with the correct value.
AI Engineer · The Transcript Looked Fine. The Call Wasn't. — Debugging Voice Agents, Arize
Recommendation · MoneyThe source recommends keeping time to first audio under 2.4 seconds to avoid dead air, and setting time-to-first-audio SLAs for customer service voice agents.
AI Engineer · The Transcript Looked Fine. The Call Wasn't. — Debugging Voice Agents, Arize
Recommendation · MoneyThe source recommends that when a user corrects a voice agent, verify the correction was actually processed before the call ends, and that when a transcript shows the agent 'correcting itself', verify the correction in the audio.
AI Engineer · The Transcript Looked Fine. The Call Wasn't. — Debugging Voice Agents, Arize
Recommendation · MoneyVoss recommends asking your coding agent to add observability with Arize AX to the application as the first step of closing the loop. This is simple because the OpenAI Agents SDK the app is built on is already instrumented with OpenInference — the open standard used by the entire observability industry — so all you have to do is turn it on: point it at your endpoint with your API key and it sends the traces it was already generating.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 201 — Laurie Voss, Arize AI
Recommendation · MoneyThe next step after basic observability is to ask your agent to pull your traces and look for things that are going wrong. A default off-the-shelf coding agent won't know how to do this — it needs the skills installed with the repo. The agent uses an AX command line tool with the API key and endpoint to pull the same traces visible in the UI and inspect them programmatically, summarizing span kinds, statuses, and errors. The skills (in .agents/skills, or installable via `npx skills install arize-ai`) are what give the agent all the information it needs.
AI Engineer · From Vibes to Production: Evaluating and Shipping AI Agents That Work 201 — Laurie Voss, Arize AI
RecommendationDr Ghayas Issa explains that menin inhibitors are epigenetic therapies that take a little more patience; if the patient is deriving benefit and there is no clear progression, he would err on the side of keeping the menin inhibitor to see if it eventually turns into a full response.
Consequential creator advice requires independent verification before Today can promote it.
Research To Practice · Acute Myeloid Leukemia — An Interview with Dr Ghayas Issa on the Role of Menin Inhibitors
Recommendation · SocialDon't put too much emphasis on how startled she was or the fact that she's leaving right now at the start of the conversation, and don't put too much pressure on her early. Once you have her undivided attention, just vibe for the first couple minutes.
Coach Kyle · A Realistic Approach at a Shopping Center (Infield)
RecommendationThe host recommends finding and featuring unknown, real people on the cover instead of celebrities, and conducting a nationwide casting call to find them.
Figma · 8 Designs That Inspired i-D Magazine's Editor in Chief | Source Material
Recommendation · MoneyAndrew Chen recommends that founders apply to Speedrun when they have a team of 2-3 people, have not taken outside capital, and are willing to move to San Francisco for the 12-week program. Teams larger than 3-5 people are rare and dividing ownership too thinly is undesirable. Founders should be prepared to bootstrap on their own dime for six months before applying.
The a16z Show · What It Takes to Build a Startup | Andrew Chen & Matt Perault
Recommendation · MoneyAndrew Chen recommends that founders in the Speedrun program should spend the 12 weeks focusing only on building the product and selling the product, figuring out if the business is going to work. Founders should not outsource coding or hire junior developers early; instead, keep the team to 1-2 people using AI coding tools.
The a16z Show · What It Takes to Build a Startup | Andrew Chen & Matt Perault
Recommendation · MoneyShaun Crimmins recommends defining success criteria and baseline metrics before the pilot starts, then showing tangible progress against those numbers week by week. He suggests having the customer write down a guess at their current metric value in week one, then comparing it to the actual number after the first week. Use stories from other companies to help the customer define their baseline numbers when they don't know them.
30 Minutes to President's Club | No-Nonsense Sales · #607 - How to Turn Every Sales Pilot Into a Closed Deal | Shaun Crimmins
Recommendation · MoneyShaun Crimmins recommends starting with upstream metrics you can impact during the pilot, then connecting them to larger business outcomes. Do not focus on the ultimate C-level business impact metric during a short pilot; instead, pick a leading metric upstream that you can move in 2-4 weeks. Pilots should impact a leading metric, then help the buyer connect the dot to the lagging (business impact) metric, without necessarily accomplishing the lagging metric during the pilot.
30 Minutes to President's Club | No-Nonsense Sales · #607 - How to Turn Every Sales Pilot Into a Closed Deal | Shaun Crimmins
Recommendation · MoneyRob Walling advises founders facing a terms of service threat from a competitor to first evaluate whether they think they have done something wrong, then talk to a lawyer. If a lawyer is unaffordable, an LLM like ChatGPT or Claude can be used cautiously, as there is no attorney-client privilege and chat logs can be subpoenaed. Founders should also assess the seriousness of the threat (a letter from an attorney is more serious than an angry email) and the likelihood the competitor will follow through. Receiving multiple threats may indicate overly aggressive marketing.
Startups For the Rest of Us · Episode 849 | Legal Threats, AI as Platform Risk, Building Confidence, and More Listener Questions (Rob Solo)
ReportFabian Theis describes LabCompass, a new tool from his lab that works as an inverse generative model: given a forward world model, it tells you what experiment to do to reach a desired biological state.
Latent Space · Controlling Biology - Fabian Theis, Helmholtz Munich