Method · BodyThe best approach to establishing training zones is to take a lab step test first for control and guidance. About a week after the lab test, do a field test outdoors to check whether treadmill heart rate is comparable to outdoor heart rate and perceived effort. Thereafter, use lactate testing to cross-check intensity once in a while, not on every session, since testing is expensive.
hedged
Method, not a verified protocol.
Garage Strength · The Genius Who Solved Endurance Training w/ Dr. Marius Bakken
Method · BodyIn a normal winter training week for elite athletes, Bakken prescribes double threshold — two threshold sessions in one day — on Tuesdays and Thursdays, with the harder X session on Saturday where you really push. The two weekly threshold sessions are split into intervals: one short interval session (45 seconds to one minute) and one longer session (six-minute intervals), keeping total running time in threshold sessions to about 25 to 35 minutes. The X session is a hill workout in winter and more flat work in summer when preparing for track workouts.
Method, not a verified protocol.
Garage Strength · The Genius Who Solved Endurance Training w/ Dr. Marius Bakken
Method · MoneyWhen starting to optimize an AI system with RSI, begin with prompt updates because they are the easiest and most straightforward way to keep iterating. After prompt optimization, the next step (L1 ladder) is to optimize the structural context of the AI system. For the L2 ladder, expand scope to automatically design the entire workflow, a kind of meta information design. After workflow design, the next step is to optimize the harness code itself.
Method, not a verified protocol.
AI Engineer · An Agent That Fixes Its Own Mistakes - Saloni Gandhi & Ziwei Guo, LinkedIn
Method · MoneyThe RSI process is broken into seven steps. (1) Run the agent and generate samples: given a query, generate output and log the processes in traces (e.g., LangSmith traces) — these trajectories are what you optimize. (2) Run the traces through the evaluation engine (e.g., the LLM judge) to get judge output showing where the system is good and where it is not. (3) Agent reflection / root-causing: understand the root cause of bad output — agent, model, prompt, tool, or other issue; the root cause can even be a judge issue or rubric issue. (4) Based on the root cause, generate an edits proposal — currently just prompts, but in future a new workflow design, harness code, or even optimizer code. (5) Validate the edits, since some are good and some bad, using the same evaluation engine (samejudge
Method, not a verified protocol.
AI Engineer · An Agent That Fixes Its Own Mistakes - Saloni Gandhi & Ziwei Guo, LinkedIn
Method · MoneyRun the agent and generate samples logged in traces; run the traces through the evaluation engine (LLM judge) to see where the system is good and not; perform agent reflection (root causing) on judge output; generate an edits proposal based on the root cause; validate the edits using the same evaluation engine (same judge, same rubrics); after the prompt is updated, run evaluation again on a holdout set to confirm the new system performs better; then make the updates the new baseline and continue iterating at a good cadence.
Method, not a verified protocol.
AI Engineer · An Agent That Fixes Its Own Mistakes - Saloni Gandhi & Ziwei Guo, LinkedIn
Method · MoneyCluster the failing rows from production traces by failing dimension using pure clustering without any LLM calls; within each cluster run an LLM to diagnose each row and assign a root cause plus a tag; merge near-duplicate issues because LLM-generated root causes are free tags with slightly different wording; then group and rank canonical issues by frequency to identify major bottlenecks.
Method, not a verified protocol.
AI Engineer · An Agent That Fixes Its Own Mistakes - Saloni Gandhi & Ziwei Guo, LinkedIn
Method · MoneyCollect a bunch of production traces from your agent, get a human to annotate about 50 of those trajectories with very detailed feedback (e.g. this is a long response, this is a short response, this is a good response, this uses this terminology), use GEPA to optimize an LLM-as-a-judge prompt, then use that judge prompt to go back and optimize your agent and deploy it, forming a data flywheel for continuous improvement.
Method, not a verified protocol.
AI Engineer · Beating RL With Reflection: GEPA and Optimize Anything — Lakshya A. Agrawal, GEPA
Method · MoneyFor subjective tasks that are hard to evaluate, GEPA can learn evals for the task from production traces. Collect a bunch of production traces from your agent and have a human annotate about 50 of those trajectories with very detailed feedback — specifying things like response length preferences, quality judgments, and terminology usage. Once you have human annotations, use GEPA to optimize an LLM-as-a-judge prompt. Use the optimized judge prompt to go back and optimize your agent, then deploy it, creating a data flywheel of continual improvement. This eval-learning flywheel is a successful paradigm that some leading teams in production are already using.
Method, not a verified protocol.
AI Engineer · Beating RL With Reflection: GEPA and Optimize Anything — Lakshya A. Agrawal, GEPA
Method · MoneyErina Karati lays out an ordered recipe for long-horizon agents: (1) freeze the harness, (2) define scenarios, (3) log traces, (4) score behavior, and (5) expose only a small policy surface. The steps are presented as an ordered method for making agent behavior measurable and safely improvable.
Method, not a verified protocol.
AI Engineer · Long-Horizon Agents Need Experiments, Not Just Prompts — Erina Karati
Method · MoneyJesse Hall lays out a benchmarking method: take a complete working agent with a real backend, put simulated callers against it (callers that pursue goals, change their minds, and introduce background noise), define each scenario as what the caller wants, what actions the agent should take, and what outcomes should be, then score those outcomes; swap the model stack while keeping the same agent and scenarios so every model sees identical input, and run the loop again.
Method, not a verified protocol.
AI Engineer · How to Build Voice Agents That Don't Fall Apart When You Interrupt - Jesse Hall, LiveKit
Method · MoneyTo pick models, test the conversation: take a complete working agent with a real backend and put simulated callers against it. Simulated callers can pursue goals, change their mind, and have background noise added to make the scenario real-world. Run text-based simulations to test business logic, and audio-based simulations to test speech recognition, turn-taking, and the entire voice pipeline. Define each scenario as: what the caller wants, what actions the agent should take, and what the outcomes should be. Score agents on outcomes — in a hotel benchmark, did the correct reservation hit the database; in a financial benchmark, did the correct authorized amount get charged. Don't score on completion alone, because agents can game completions — verify they took the right steps: check that a
Method, not a verified protocol.
AI Engineer · How to Build Voice Agents That Don't Fall Apart When You Interrupt - Jesse Hall, LiveKit
Method · MoneyWhen the happy path doesn't happen and a prospect gives a lukewarm feel, Rob Snyder recommends throwing the 'Hail Mary': asking 'Is there anything on your to-do list that you would pay me to do right now — that you would pay somebody to take off your plate?' He reports that quite a few startups have totally pivoted and taken off from this question.
The pivot outcomes are described anecdotally.
Method, not a verified protocol.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Method · MoneyRob Snyder describes a sales call as mostly two steps: pull discovery, then the conceptual pitch of how you might help. Approach it leaned back and excited to chat — 'let's see if I might be able to help you' — not wired up to convince; if they have pull, they'll convince you to let them buy. Open by asking why they took the call, which ideally lets them fill out the entire pull framework. If they say 'I don't know, you reached out to me,' use what they just told you and propose it as the agenda, getting their thumbs up. Then earn the right to ask discovery questions by asking permission — this prevents the call from feeling like an interrogation, because if it does, they're not telling you anything. Make it a 30-minute calendar block (smaller blocks don't get seen), and be iffy about a on
None stated.
Method, not a verified protocol.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Method · MoneyRob Snyder walks through a call script: ask why the prospect took the call, ask what they're trying to do but have been unable to do with existing tools, ask whether the problem must be solved now, and ask what other options they've looked into and whether those are good enough.
presented as the guest's own practiced script
Method, not a verified protocol.
Lenny's Podcast · How to find product-market fit (but, actually) | Rob Snyder (Author, The Power of Pull)
Method · MoneyBuild a five-step regression loop: (1) select a dataset by pulling raw real production traces; (2) add a scoring mechanism by bringing in evaluators; (3) run a baseline experiment by running the agent against all production traces and datasets, then running the eval testing suite on the output — in the SDK via run_experiment, passing in the dataset, task, and evaluation suite; (4) add the one change, rerun the evaluation, and compare eval results; (5) automate the regression process by pushing the experiment into CI (e.g., GitHub Actions) so each pull request triggers it. Use observability as a way to build a robust regression process that feeds automatically into CI/CD.
Method, not a verified protocol.
AI Engineer · Your Agent Fix Just Broke Something Else - Sofia Jakovcevic, Arize AI
Method · MoneyTo bring in a new hardware backend: characterize the hardware, then specify what each of its operations means by equating them to operations the compiler already understands via rewrite rules. New rewrite rules propagate out to the rest of the equalities already specified in the compiler, making it a fairly low-effort process even for very different backends.
Method, not a verified protocol.
Y Combinator · Not All Open Model Tokens Are the Same | YC Paper Club
Method · MoneyHarneet Kaur created a benchmark that assesses each layer of the AI reasoning process to improve assessment of the models' process. On the extraction layer, LLMs can mess up, though with public companies it's pretty solved; extraction with private loans, Excel documents, and PDFs can still be a challenge. The reconciliation or derivation layer comes after extraction — having EBITDA and debt, you can derive leverage. The judgment layer asks whether, given the leverage, this is a distressed credit or not. To get to the prediction layer, you must first assure the data points are good, the reconciliation layer is good, and there is some sort of reasonable judgment. When analyzing a distressed credit, you should look at whether you agree or disagree with the different assumptions at each layer.
Method, not a verified protocol.
AI Engineer · Can AI Understand a Distressed Credit? - Harneet Kaur, Axim Intelligence
Method · MoneyOnce you know who your audience is, figure out what they like. Associate your brand with what your audience likes by making content about the thing they like, making products they like (free and paid), and appearing next to the people they like.
Method, not a verified protocol.
Alex Hormozi · How I Would Build a Personal Brand in 2027 (If I Had to Start Over)
Method · MoneyStart AI evaluation with a clear requirement, such as 'advice must match the approved policy'. Challenge the AI chatbot with edge-case questions (e.g., about reimbursement after travel) to expose incorrect advice before deployment. If eval tests expose incorrect AI advice, add a policy check or human escalation as a control. After adding controls, test again to verify the controls catch errors before they reach the customer.
Method, not a verified protocol.
AI Engineer · 1,368 AI Failures and What They Teach About Evals — Vishal Srivastava & Tanmay Sah, New York Life
Method · MoneyStart by looking at your system claim and testing it to try to break it, then identify where the system claims break. Map out the uncertain region and high-risk region where the AI system's performance cannot be trusted, creating a full failure region map. Then either restrict the AI system from operating in failure territories or add supporting systems you are confident of, and iterate until reaching the desired confidence level. Apply this build-break-fix-break-again approach before deployment to expose failures that narrow evaluations miss in the real world.
Method, not a verified protocol.
AI Engineer · 1,368 AI Failures and What They Teach About Evals — Vishal Srivastava & Tanmay Sah, New York Life