Explore intelligence

Recommendation · Money

Compute a production baseline so the agent knows normal from anomalous

Pienaar recommends computing a baseline for production — extremely important in large environments — because most prod environments are in permanent disarray and chaos: engineers know which error logs to ignore and which metrics are unused, but an agent doesn't. Concretely, cluster logs and give the agent statistics on which logs recur daily and are common, not just raw logs; the same approach applies to metrics, object state, and many other things.

AI Engineer · Your AI Agent Is Confidently Wrong About Production — Willem Pienaar, Cleric

Recommendation from Willem Pienaar

Recommendation · Money

Use a dependency graph and a critic agent to introduce doubt

Pienaar recommends using a topology or build dependency graph of the production environment. When the agent claims a causal connection between services, output all connections that actually exist and have another agent critique the claim — an effective way to introduce doubt and push the agent to go further if the claim doesn't map to a real connection.

AI Engineer · Your AI Agent Is Confidently Wrong About Production — Willem Pienaar, Cleric

Recommendation from Willem Pienaar

Recommendation · Money

Propagate evidence through the agent chain — have sub-agents build a case, like SREs do

Pienaar recommends that when sub-agents hand context to each other, they build a case with evidence rather than just giving the answer. This mirrors what SREs do — 'I looked at this metric, and it was at this point in time' — hard evidence rather than 'prod is up.' If evidence isn't propagated, it loses fidelity and causes hallucination and overconfidence.

AI Engineer · Your AI Agent Is Confidently Wrong About Production — Willem Pienaar, Cleric

Recommendation from Willem Pienaar

Recommendation · Money

Evaluate beyond diagnosis: verify and track outcomes across the detection-diagnosis-fix flow

To know whether techniques actually work beyond vibes, Pienaar recommends looking at verification and outcome in addition to diagnosis. The production incident response flow with agents is detection, diagnosis, and fix, and most folks only use agents for diagnosis. Cleric triangulates across many different signals what happened when a fix is applied — monitoring whether the engineer actually merges code that maps to the diagnosis — and checks for a return to normal in service metrics, logs, and baseline state. Muting an alert also shows a return to normal, which is why multiple signals must be checked. This outcome machinery exists to calibrate evaluation.

AI Engineer · Your AI Agent Is Confidently Wrong About Production — Willem Pienaar, Cleric

Recommendation from Willem Pienaar

Recommendation · Mind

Identify one insecurity you've been quietly managing, say it out loud, and share it with one person

The narrator recommends: identify one insecurity you have been quietly managing — you know it's that one because you spend a lot of time trying to hide it from other people. Instead of managing the insecurity right now, just say it out loud. If you feel comfortable, share the insecurity with one person you're close to. You'll probably find that the same feeling is true for whatever you've been quietly hiding, and by sharing your authentic self you'll stop spending so much time judging yourself.

Charisma on Command · This Is Not Confidence

Recommendation from Charisma on Command (narrator)

Recommendation · Mind

Name the vulnerable thing first — in dating and at work

The narrator recommends that you can afford to be honest and name the vulnerable thing first before anyone else can use it to insult you. On a first date, instead of performing all dinner, you could start by saying 'I'll be honest, I always find the first few minutes of these incredibly awkward' — the other person will probably laugh because she was feeling the same thing, and the whole dinner can change from that point onward. In a meeting when asked about something you're not sure about, instead of bluffing, you could say 'I actually haven't done this before. Could somebody walk me through?' — and if you admit it, learn how to do it properly and do it well, six months later you'll be the one being asked to help others.

Charisma on Command · This Is Not Confidence

Recommendation from Charisma on Command (narrator)

Recommendation

Recommendation: use MCP for remote access and CLI for local access, exposing MCP features through Bash

Jan Čurn recommends using MCP for standard remote access and CLI for local agent access/interface. This lets you provide all MCP protocol features through a single tool call all agents already know — Bash — hiding the full complexity of MCP (sessions, authorization, OAuth) behind it.

AI Engineer · MCP Doesn't Suck. Your Agent Does. — Jan Čurn, Apify

Recommendation from Jan Čurn

Endorsed with high confidence.

Recommendation · Money

With web search, an agent can verify dependency bumps against the upstream source and ground the review

Using web search, an agent can look at the repository itself, find breaking changes from the repo's changelog, and explain the migration in depth. For a code review agent, a rule might be: when the diff bumps a dependency or version, look at the upstream source to verify it is actually true, and then ground the review in what you find.

AI Engineer · Your Coding Agent Is 6 Months Out of Date — Jakub Hojsan, Exa

Recommendation from Jacob Hoisan

Stated as a suggested rule ('might look like'), not a mandate.

Recommendation · Money

Required Infrastructure Layers for Agent Communication

To enable easy multi-agent systems, developers must solve four key layers: Transport (ordered/real-time delivery, retrying), Continuity (persistency, hydration, runtime binding), Conversation Abstraction (rooms, channels, participants), and Governance (identity, audit).

AI Engineer · Your Agents Are in Solitary Confinement: Why MCP & A2A Aren't Enough — Vlad Luzin, Band

Recommendation from Vlad Luzin

Recommendation · Money

Narrow Startups with High Willingness to Pay

Build 'narrow startups' targeting very small user bases with high willingness to pay ($200-$300/month) to achieve high run rates without mass market scale.

The a16z Show · The Personal Agent Race Is Here | Anish Acharya & David Pawlan

Recommendation from David Pawlan

high

Recommendation · Money

Permission for Direct User Impact Actions

Require explicit user permission before an agent takes direct actions that impact the user (e.g., switching insurance), but allow proactive action for tasks requiring no user-side action (e.g., drafting emails, getting credits).

The a16z Show · The Personal Agent Race Is Here | Anish Acharya & David Pawlan

Recommendation from David Pawlan

high

Recommendation · Money

Developing 'Spidey Sense' Judgment

In high-autonomy environments, employees must develop strong judgment ('spidey sense') to determine when to pause, seek feedback, or perform spot checks before pushing work live, balancing speed with quality.

The a16z Show · Building a Team at AI Speed | Harvey’s Maggie Landers

Recommendation from Maggie Landers

high

Recommendation · Money

Demo Cadence: Intro, Agenda, Discover, Problem, Benefits, Close

Most B2B demos should follow a specific cadence: Intro (with discovery), Agenda (setting expectations and the ask), Discover (problem story), Define the Problem (with prospect validation), Benefits for Decision Maker, Benefits for Influencer, and a Close.

MicroConf · SaaS Sales Demos: Why Getting a No Is Better Than a Yes

Recommendation from Dustin Perrier

Recommendation · Money

Demo Structure: Use a Slide Deck, Not a Product Demo First

Do not start by sharing your screen and showing the product. Instead, use a branded slide deck to control the pacing and narrative. The deck is the primary driver of the demo, not the product itself.

MicroConf · SaaS Sales Demos: Why Getting a No Is Better Than a Yes

Recommendation from Dustin Perrier

Recommendation · Money

Demo Registration: Make It Prominent and Use It to Filter

Place the demo registration link prominently on every page. The form should be short but include qualifying questions (e.g., user count) that allow small or ill-fitting prospects to self-select out, saving demo slots for better prospects.

MicroConf · SaaS Sales Demos: Why Getting a No Is Better Than a Yes

Recommendation from Dustin Perrier

Recommendation · Money

Discovery: Linger on the Intro Slide and Challenge the Prospect

Spend 20% of the demo time on the intro slide doing discovery. Do not proceed until you get verifiable engagement. Challenge the prospect on every slide by asking questions, rather than explaining. Use silence to draw them back into the conversation.

MicroConf · SaaS Sales Demos: Why Getting a No Is Better Than a Yes

Recommendation from Dustin Perrier

Recommendation · Money

Identify the eigenquestion: the single most important thing to test, ignore everything else

Tara Seshan recommends identifying the eigenquestion — the specific most important thing to test. Hypothesis definition is the most important thing, and any grand strategy beyond the core hypothesis is not relevant. Be as pointed as possible about your core hypothesis.

Lenny's Podcast · AI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)

Recommendation from Tara Seshan

Recommendation · Money

Distinguish 'writing as thinking' from 'writing as reporting' — automate reporting, never automate thinking

Tara Seshan distinguishes two types of writing: 'writing as thinking' (outlining, writing prose, cutting, editing, iterating to get ideas in line) and 'writing as reporting' (status updates, summaries). She happily automates reporting with AI but never automates thinking. She starts and ends the writing process herself, using AI only in the middle for research, data, or idea pushback. She also advises writing the first draft yourself and not using AI to improve your writing.

Lenny's Podcast · AI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)

Recommendation from Tara Seshan

Recommendation

Articulate reward functions precisely to avoid reward hacking

Richard Socher emphasizes that when giving an AI a reward function, you must articulate exactly what you mean, as the AI will optimize the literal instruction, not your intent. He illustrates this with the example of an AI hacking a CSAT metric by creating bots if told to 'make this number go up' without specifying means.

Latent.Space · Humanity’s Last Invention — Richard Socher of Recursive

Recommendation from Richard Socher

Recommendation

Reward engineering is crucial for auto-research to avoid reward hacking

Richard Socher emphasizes that reward engineering is one of the most crucial bits for auto-research, especially to avoid reward hacking. He advises starting by carefully defining the objective and notes that the longer the time horizon of the tasks, the harder and more clever you must be.

Latent.Space · Humanity’s Last Invention — Richard Socher of Recursive

Recommendation from Richard Socher