Sarosh Waiz

What's harder: getting good data or getting people to use it?

A product failing in one market is expensive. That same product succeeding after repositioning into a different market is proof the problem was framing, not function. What's the best example you've seen of this?

The self-reinforcing hiring loop: require category experience, get category-standard ideas, conclude that category experience is essential. Repeat until every competitor looks the same.

Service businesses should tie sales compensation to customer satisfaction at month six. Watch how fast the handoff problem gets solved.

Most AI roadmaps assume that once leadership approves and the tech works, adoption follows. It almost never does. The roadmap forgot to account for human self-interest.

@ajmahler 💯

@saroshws High end customers gladly pay more for discernment and discretion.

When you hire a service provider, do you prefer knowing the hourly rate or the total project cost?

@coreyganim People don't pay for monitoring. They pay for someone who can tell them what to do with what they found.

the business hiding in this repo: 1. pick a niche (real estate agents, ecommerce brands, SaaS founders) 2. use this tool to monitor Reddit, X, and YouTube for mentions of their brand, competitors, and industry keywords 3. pipe the results into an AI agent that writes a daily briefing 4. charge $500-$1,500/mo per client for "market intelligence as a service" your client gets a daily report they'd never have time to build themselves. you set it up once and it runs on autopilot. 5 clients = $2,500-$7,500/mo recurring. zero API fees.

@itsalexvacca Impressions and pipeline are measured in different currencies. Most B2B teams optimize one and wonder why the other doesn't follow.

@aakashgupta Most teams stay at 80% reliable because they iterate on prompts and ignore the evaluation layer. The measurement system is the real work.

I took Karpathy's loop and applied it to the thing every team using AI agents struggles with: getting prompts from 80% reliable to 95%. The pattern is identical. One file changes. One metric scores it. The agent makes one edit per round, tests it, keeps winners, reverts losers. 12 experiments per hour. 100 overnight. Instead of optimizing a training script, the target is any prompt or system instruction you use repeatedly. Customer support agent prompts. Internal workflow automations. Data extraction pipelines. Code review instructions. Anything where you've written a prompt, gotten it to "good enough," and moved on because manual iteration hit diminishing returns. The setup takes three things. The target prompt you want to improve. 2-3 realistic test inputs, the kind of request that would actually hit the prompt in production. And 3-6 binary yes/no checks that define quality. Did the output meet the format constraint? Did it follow the specific instruction? Did it avoid the failure pattern you keep seeing? The loop: Execute the prompt 30 times across all test inputs. Score every output against the checklist. Analyze which criterion fails most. Mutate one thing in the prompt. Check if the score improved. If yes, git commit. If no, git reset. Repeat until you're above 95%. What you wake up to: the improved prompt saved separately, original untouched. A results.log showing every round's score. A changelog explaining what worked, what didn't, and why. The insight Karpathy landed that transfers beyond ML: if you can score it, you can autoresearch it. Training loss is a score. A binary checklist on prompt output quality is also a score. The loop doesn't care what it's optimizing. It only needs a number that goes up or down. Prompt engineering today looks like software before unit tests. Manual tweaking, vibes-based evaluation, no version control, no systematic iteration. The Karpathy loop applied to prompts turns it into an engineering discipline with measurable improvement per iteration. Every team running AI agents has prompts that work "well enough." The gap between well enough and reliable is exactly the gap this loop closes while you sleep.

@thetripathi58 This is why talent acquisition can't fix a broken operating model. Most problems aren't people problems.

A brilliant statistician who spent 50 years studying why massive engineering projects fail realized one terrifying truth: Individual incompetence is almost never the actual problem. His name is W. Edwards Deming, the man who famously rebuilt Japan's post-war manufacturing empire from scratch. He argued that we obsess over individual performance and completely ignore the environment. Here are 4 operational frameworks he used to build elite, failure-proof organizations:

@jimheskel The plan that lives in a document never gets tested. The one that ships gets corrected by the market.

Planning feels safe. Shipping builds businesses. More on closing the knowing-doing gap: jimheskel.com

Most people already know they should start. They're choosing not to. Not because of clarity. Not because of timing. Not because of money. Because starting means finding out if they're as good as they think they are. Let's not call it preparation. This is protection.

I think that is the right way to go about it. But I was reading an article a few weeks ago, and it argued against the idea of hourly rates because it undervalues the intellectual capital a service provider offers as part of the project delivery. And due to this mindset, even highly complex services are compared to physical labour.

@saroshws Service provider here (contractor). The hourly rate is given when there are too many unknowns. A project price is given for two reasons, payment up front cause the client is sketchy or the provider is sketchy, and when maximum profit is expected cause fee unknowns.

This week we're walking through through 2 things: 1). How The Story Gap works as a practical diagnostic you can use on anything. 2). Apply it to the biggest default story forming right now: AI adoption. open.substack.com/pub/thestrateg…

Large companies worry about what happens if someone makes a bad decision fast. Solo operators worry about what happens if a good decision takes too long. Both fears are real. One costs more.

Measurement timelines create the perfect excuse. A quarterly dashboard shows Q3. A strategic bet plays out over 18 months. Leaders use that gap to ignore anything inconvenient.

The best hire a large company can make is someone who spent three years working alone. They've seen what speed looks like without organizational drag.

AI adoption without org restructuring is just an expensive way to produce better PowerPoint presentations.

Competitor analysis tells you where the category already is. Studying a brand from a completely different industry tells you where it could go. What's one outside brand that changed how you think about yours?