Recommendation · MindAlex Hormozi defines patience as doing one thing for a long period of time and delaying gratification. He advises sticking with things for extended periods and not starting new projects. He also describes patience as giving yourself 'surface area of thinking' by revisiting the same problem with different perspectives over time.
Alex Hormozi · 3 signs you're going to be a successful entrepreneur.
Recommendation · MindAlex Hormozi defines frustration tolerance as being willing to get 'kicked in the face or the gonads for extended periods of time' and continuing anyway. He advises sitting with discomfort and continuing to take the actions you thought you should take before your emotions tried to convince you otherwise.
Alex Hormozi · 3 signs you're going to be a successful entrepreneur.
Recommendation · MindAlex Hormozi advises against changing your business just to alleviate how you feel emotionally. He states that taking action to alleviate anxiety does not mean the action was correct, and that constantly shifting priorities whiplashes your team and customers, leading to confused messaging and marketing.
Alex Hormozi · 3 signs you're going to be a successful entrepreneur.
Report · MoneyEvery agent action at Orca produces a change log of what changed, who changed it, from which part of the system, and when; the audit logging applies to human actors too — UI changes affecting a client are logged the same way, and the same audit trail also feeds the agentic loop by informing agents that a change happened.
Latent Space · Agents Prompt People: Using Agents for Long-Horizon Financial Work
Report · MoneyWhenever data leaves Orca's system (web search, sending a calendar invite), deterministic checks for PII plus a consensus of small models guard against data exfiltration.
Latent Space · Agents Prompt People: Using Agents for Long-Horizon Financial Work
Report · MoneyExternal inputs (emails, documents, phone calls) are prompt-injection targets, so Orca runs verifications at the edges to detect prompt injection.
Latent Space · Agents Prompt People: Using Agents for Long-Horizon Financial Work
PrincipleThe agent never gets to say approved for grants touching sensitive data, migrations that change intent, or when it is not confident in its own verdict — no matter how confident it sounds. When a change alters what the data means rather than just how it is moved or stored, a person decides. When the agent is not confident in its own verdict, it does not get to round up to approve — it escalates to a human.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleUse deterministic scripts to check facts against a source of truth (like a schema registry or catalog) as layer one of PR review. Configure a failed deterministic check to turn CI red and block the PR until it is fixed. Run a script that checks the catalog directly to verify every table has an owner and any PII is labeled.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleChetas Joshi endorses a three-tier routing principle: when a fact can be proven against a source of truth, let a script prove it and clear the PR without human involvement; when a decision requires context and judgment, route it to an agent; when a change is a protected call carrying real risk, route it to a human reviewer. Every fact a script proves is review work that never needs AI, keeping system cost down.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleUse an AI agent as layer two of PR review, focused on the risk that deterministic checks cannot catch. Have the agent review the PR, score the risk, and either approve or escalate to a person. The agent produces a risk score that drives the next step.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
PrincipleTo measure whether the reviewer is doing a good job, look at two kinds of mistakes: bad PRs it approved (a miss — dangerous because the change ships) and fine PRs it unnecessarily sent to a person (a false escalation that costs a person's time). The matrix separates four possibilities. Measure the false escalation cost the same way you measure a miss.
AI Engineer · 60% of Our Data PRs Now Merge With No Human Review - Chetas Joshi, Robinhood
Principle · MoneyThe auto-research layer should only edit a controlled policy surface — memory writing policy, retrieval policy, communication prompts, belief trust rules, source attribution, and replanning triggers — rather than rewriting the whole application. This gives the search process room to improve behavior while preventing it from gaming the evaluation.
AI Engineer · Long-Horizon Agents Need Experiments, Not Just Prompts — Erina Karati
Principle · MoneyWhen an AI model applies a percentage increase, check it was applied to the correct sub-population rather than the whole population, and compare any stated percentage against the real number disclosed in filings before accepting the forecast.
AI Engineer · Right Answer, Wrong Reasons: AI Forecasts Costco - Hassan El Mghari & Omar El Mghari, Together AI
Principle · MoneyWhen auditing an AI-generated financial model, check that every assumption is backed up with a real number from a citation, not just a citation without a number. AI can fail to identify the correct number even when the citation is real, and can draw assumption boundaries (e.g., 5-30%) that exclude the actual disclosed number (45%).
AI Engineer · Right Answer, Wrong Reasons: AI Forecasts Costco - Hassan El Mghari & Omar El Mghari, Together AI
Creator perspective · MoneyExtra agents pay off the most in planning on ProgramBench, where discovery is the bottleneck. Allocating more agents to planning achieves almost the same performance as the heavy-prompt (PIV) setup at half the cost. Each planning dollar buys almost 1.6x more reward than any other stage (implementation or verification). Planning delegation yields 2.4x more reward per call than delegating throughout implementation.
AI Engineer · How to Get Frontier Coding Results at 70% Lower Cost - Nouha Dziri, Cohere
Creator perspective · MoneyReplacing Sol workers with cheap Luna workers while keeping Sol as lead achieves almost the same performance at 60% less cost. A strong lead (Sol) with cheap workers (Luna) achieves the same performance as Sol plus Sol but at much lower cost. A 16-planner team (Sol lead, Luna planners) is 38% cheaper than an all-Sol planning team and 70% cheaper than workers at every stage (PIV).
AI Engineer · How to Get Frontier Coding Results at 70% Lower Cost - Nouha Dziri, Cohere
Principle · MoneyJesse Hall advises keeping the gap between the caller's last syllable and the agent's first audible sentence roughly inside 1.5 seconds so the call still holds, and notes that getting down to 500-600 milliseconds between turns is when the conversation really starts to feel human-like.
AI Engineer · How to Build Voice Agents That Don't Fall Apart When You Interrupt - Jesse Hall, LiveKit
Principle · MoneyProspects without pull can be enthusiastic — 'that's the future of our industry' — but aren't saying 'I need this right now'; leaned-out enthusiasm like 'let me think about it' or 'send more info' is not pull, and neither is 'we could use this in Q4.'
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Principle · MoneyRob Snyder recommends using the pull framework to stake out a claim — state which specific person is stuck — and then go test whether they try to buy before a product exists. If prospects try to buy, you might have something; then see if it repeats. If a prospect doesn't try to buy after the Hail Mary question, treat it as a learning signal that you might not have pull: 'cool, we've learned something.' He notes the Hail Mary question has led startups to completely pivot and take off.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Principle · MoneyRob Snyder advises selling to the specific individual at a company who has a project on their to-do list they can't do themselves — 'You're not selling to SpaceX. You're selling to the head of legal at SpaceX who has a project on her to-do list.' This operationalizes the person-level view of pull into a targeting rule for sales.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)