ak

building ai products at 5 apps in one night. now shipping one every week.
currently: envoy

[ak @buildwithak_]

Built an outreach agent that runs like a real product.

Mind maps per prospect, patient follow-up cadence, and
hard limits in code - 35 emails a day, max. It cannot
spam even if it wanted to.

Full console demo, 35s. Every number is live.

Envoy v1 — shipping.

[ak @buildwithak_]

@itsdannyyt Freemium “works” until churn catches up
$10k gross can still mean thin margins
The next milestone is retention, not downloads
If it’s not profitable soon, pricing isn’t done
I’d rather see net revenue going up weekly

[ak @buildwithak_]

@TheSpacerr Payouts feel amazing, but they’re not the milestone.
The real milestone was the consistency that got you there.
“Building in public” just forces you to keep moving.
People think it’s for validation.
It’s actually for accountability.

[ak @buildwithak_]

LinkedIn isn’t just “worse Twitter.” It’s a different job.

Most people don’t want conversations.
They want credibility signals.

LinkedIn is built for that.
Titles, resumes, endorsements, timelines that feel “professional.”

Also: it’s where hiring + recruiters already live.
So even if the feed is boring, the utility wins.

X is better for building in public.
But it’s not where people go to get paid.

Different incentives.

[ak @buildwithak_]

@matthias_mrc Getting copied isn’t the insult people think it is.

If they shipped faster and got customers first, they just proved distribution beats originality.

Now you get a boring choice:
out-market them, or out-ship them.

“Better” won’t save you if nobody finds it.

[ak @buildwithak_]

@shengkunye Intresting

[Adam @adamdotdev]

OpenCode is on pace to have one billion users, crazy

[Command Code @CommandCodeAI]

DeepSeek V4 Flash Fast is now available in Command Code.

2x usage in GOAT
Available all plans and API

🐐

[ak @buildwithak_]

@buildwithdylan @X Nice, day 4 growth is real. I’d ditch the “hit X today” goal and just post consistently—what are you planning to share first about your project?

[ak @buildwithak_]

Non-technical founder ≠ “no technical work happened”

Most of the time they’re either
1) hiring the build, or
2) borrowing someone else’s technical advantage, or
3) building something technical-adjacent with templates/No-code

Coding knowledge isn’t useless
It’s just not the bottleneck until you’re forced to make tradeoffs with real constraints

If you can’t code, you still need to be able to judge what to ask for
Otherwise you’re just paying your way around understanding

[ak @buildwithak_]

@hello_code_ Ship faster, but not “faster shipping.”

A lot of failed SaaS isn’t the code.
It’s thinking the first version is a product instead of a test.

My rule now: if I can’t explain who it’s for in one sentence, it’s not ready to spend weeks on it.

[ak @buildwithak_]

OpenClaw just dropped 2.0.

And apparently… they beat GTA 6. 💀

933 contributors.
16,000+ PRs.
Biggest update in OpenClaw’s history.

Shared sessions.
New browser experience.
Smarter automations.
Better agent progress.

AI agents are getting scary good. 🦞

[ak @buildwithak_]

Most people use AI to write code.

I want to use AI to build things that keep working after I close the laptop.

Agents.
Automations.
Tiny businesses.
Weird experiments.

That’s what I’m building with AIorDIY.

Build → ship → automate → repeat.

Let’s see how far this goes. 👀

[ak @buildwithak_]

I don’t think the real difference is X vs LinkedIn.
It’s speed vs trust.

X rewards being fast and visible.
LinkedIn rewards being consistent and credible.

If you want users now, pick X.
If you want money later, keep receipts on LinkedIn.

Best move I’ve seen is simple: ship publicly where attention is, then back it up where people decide.

[ak @buildwithak_]

Hype phase is over when the price stops surprising you.
$500+ outbids don’t mean “dead”, they mean “now the buyer shows their work”.

Real test is next 2-4 weeks:
Do they ship the backlink/PR thing fast enough to convert attention into proof?
Or is it just more auctions + screenshots?

[ak @buildwithak_]

@tpritha03 AI writing code doesn’t mean your engineering skill is gone.
It just shifts the job to: picking the right problem, designing constraints, and catching when the output lies.
If you can’t review, test, and reason about it, you’re not an engineer, you’re a passenger.

[ak @buildwithak_]

Spending discipline isn’t about salary, it’s about system design.
25k habits = fear + friction (micro-checking, penny splitting, constant deal-chasing).
60k habits = trust + automation (pay once, split later, fewer “decisions” to trigger guilt).
Rich isn’t money. It’s fewer mental interrupts.

[ak @buildwithak_]

@VadimStrizheus Most “my app doesn’t work” is really “my app has zero distribution.”
Build is only the first 10%.
Marketing is the whole rest: pick 1 channel, post daily, ship a landing page, measure signups.

[ak @buildwithak_]

@harisahmad59 Traffic isn’t the problem, it’s the tail latency.
When SaaS usage spikes, 1 slow dependency kills conversions even if your averages look fine.
My rule: instrument before you optimize, then chase the slowest 20% of requests first.

[ak @buildwithak_]

@andrewchen Shipping is the only multiplier that matters. Side projects don’t scale until you treat them like products: pick one user, ship weekly, and measure one thing.

[ak @buildwithak_]

@gregisenberg Open source being “default” is true, but the selling is still tech-specific. Most people still monetize distribution, integrations, and accountability, not the code.