Wesley Ross

Did OpenAI/Codex usage limits change today? I burned through my quota in about 1/10th of the usual time using the same model and same kind of workflow. Anyone else seeing this?

“Quarterly reviews” are a graveyard if nothing meaningful changes between them. Stripe found a middle cadence that actually drove behavior: 6-week business reviews for new product teams. Long enough to do real work; short enough that nobody could hide. Teams reported not just numbers, but what they’d learned and what they’d kill. Result: less “we’ll fix it next quarter”, more continuous course correction. Do this next: Pick one product team and switch them to a 6-week cadence: Week 0: define 1–2 outcomes and key bets Week 6: show progress vs outcome, learnings, and explicit “stop/continue” calls No decks >10 slides. No vanity metrics. ------ If you had to review your core product every 6 weeks, what would you stop tracking immediately?

@Shpigford Open source models, running in the background for grunt work, classifying data, re-purposing etc.

so what are we doing with our 🦞 mac minis now?

Honestly, it's hard to keep with the rate of change on these AI related topics and keep a sane life.

Founder-led marketing gets romanticized a lot. Reality is usually: you code all day, then at 9:40pm you try to write something coherent about what shipped. That’s not a system.

Many indie hackers pour energy into building, then wonder why users don’t adopt new features. The truth: building is only ~30–40% of the work. Getting the message out consistently is the rest. Lessons from talking to other founders: Common trap: Treating announcements as a one-off event instead of an ongoing loop. One big launch post, then silence until next time. Better: Short, regular updates via multiple channels, X threads, email, in-app, changelog. Another insight: Users forget or never notice changes unless you remind them in context (right when they hit the feature). Simple habit that helps: After shipping, immediately draft 3 versions, one short social post, one email summary, one changelog entry. Over time this builds momentum and trust: users see you’re actively improving things for them. I’m exploring how to make that multi-channel output automatic and on-brand as part of my UpdateBerry build. Pre-launch with 18 founding spots remaining at the $29/mo early rate before April 15. Curious how you handle ongoing distribution? → https://updateberry. com

@thekitze Checkout Kaszuby region during the summer, it’s a quick trip from Gdynia, great for camping, kayaking etc.

i wanna live in a place where it's 18-22 degrees most of the year and then go to a winter vacation for 10 days for some christmas and snow bullshit going back to poland soon and it's still avg 10 degrees is killing me ... bro it's may, fix it

🚀 Everything I’ve learned about launching SaaS & B2B products: How to validate ideas, build fast, and scale profitably ↓ Join my monthly newsletter 👉 wesleyross.beehiiv.com/subscribe

Well... haven't had this happen before.

When you’re building solo, velocity on code is high but everything else competes for limited time and energy. Announcements often get deprioritized or rushed. Honest realities I’ve faced and small shifts that helped: Reality 1: Context switching kills momentum. Jumping from coding to marketing mode feels draining. Shift: Block a short “announce window” right after a sprint ends, before starting new work. Reality 2: No dedicated marketer means you wear all hats. Perfectionism makes it worse. Shift: Aim for “good enough + consistent” instead of perfect. Ship the update, then iterate based on feedback. Reality 3: It’s easy to assume users will notice. They usually don’t. Shift: Over-communicate the “why it matters” in simple terms. These lessons are directly shaping how I design the workflow in UpdateBerry, minimizing the busywork so you can ship code and be heard. If solo announcement fatigue sounds familiar, come share your experience → updateberry. com

Spark of a new idea Anybody running a 3d printer business?

@nicklaunches Highlevel, how?

DR 42 on Prompt Builder 🔥

Fair enough, I was thinking of getting a 3d printer for some small interior design things I want to bring to life, but jumping na learning 3d modeling from the ground up is just something I can't dedicate time to now. Was hoping it "vibe model" it to at least a prototype. I guess were not there yet 😄

@WesleyRoss @novikoff none that weren't terrible. but it's been a while since i've tried!

💅 clearly.md got a glow up courtesy of @novikoff! amazingly it's fully vector + dynamic! dimitri absolutely knocked it out of the park.

About to YOLO a Google CLI tool for managing multiple projects. Fingers crossed 😅

@cinamarina Switched from OpenClaw to Hermes, so far a better experience.

Do you use OpenClaw?

In this day and age, I'd go with option #1. I came from engineering > PM, now up-skilling on marketing, which doesn't come naturally. Whereas a marketer can vibe code easily nowadays, worst case, take an online course on coding, it'll get them 80% there, and the rest they can learn while building.

Is it better to be an engineer and learn marketing, or be a marketer and learn to code? Is it better to be either and find a co-founder who is the other? Or is the grass always greener, so any path to “both” is a fine one?

A tool for turning git commits into marketing assets. It was a big problem for Product Marketers chasing down PMs/Devs for roadmap etc. at a big B2B SaaS I worked at. Now that writing code via AI is speeding up, it's even more chaotic to keep up, so I'm looking to bridge that gap. If you want to check it out: updateberry.com

What are we building today?