Building real AI that answers real business calls.
Voice isn’t text—interruptions, accents, booking edge cases break 99% of agents. I share the raw lessons, mistakes, and systems we build to fix that.
I get how uncomfortable it feels to disengage from the syntax, from the sequence, selection, and iteration of code, from the dopamine hit of getting a complicated function to execute properly. I get it. I've been coding for longer than most of you have been alive -- I get it.
But the bar has been raised. And if I, someone who has been coding for more than six decades, can clear that bar, you should be able to clear it too.
And fear not, I've found plenty of joy on the topside of that bar. It just take a leap...
"I don't think people managers will have any value in the future."
Airbnb CEO said on a recent pod the people most screwed by AI will be people managers.
Perhaps we had this all wrong. The future belongs to ICs. There is now more use for them than ever before.
@lifeofjer Why would you give direct access to your production database through internet?! That’s your most obvious failure. Production database must not ever be accessible directly from outside your solution network.
VERCEL GOT HACKED
ShinyHunters - the group behind the Ticketmaster breach - is selling Vercel's internal database for $2M on BreachForums
here's why every developer should care:
- they have NPM tokens and GitHub tokens
- Vercel owns Next.js - 6 million weekly downloads
- one malicious push = global supply chain attack
- Vercel confirmed the breach today, April 19
- they literally DMed the hackers on Telegram asking them to stop
rotate your env variables RIGHT NOW
Why do we keep underestimating the power of AI in small business? It's not just about automation, it's about building relationships and trust. Let's rethink our approach to tech. What's your take?
Great point — hybrid retrieval is table stakes now.
For our agentic workflow we keep memory bi-temporal:
observed_at (when the agent learned it)
valid_from / valid_to (when the fact is true in-world)
Then pruning/ranking uses both recency and validity confidence, so “old but still true” survives while actually stale facts decay. Around 30–50 sessions, this stopped a lot of silent regressions for us.
We spent a few months building an AI agent that runs small business
operations.
Not another workflow builder. An actual AI that thinks, learns,
and earns your trust over time.
Here's how it works (and the counterintuitive bet we made): 🧵
@jeffweisbein Exactly. The missing piece is an explicit state machine: observe → propose → simulate → execute → verify. Every transition needs guardrails, full logs, and reversibility (checkpoints + compensating actions). Learning is easy; trust comes from rollback and auditability.
Building AI for small businesses isn't just about automation; it's about
empowering them to focus on what really matters.
Let's talk about the real impact AI can have when designed thoughtfully.