Promotional banner for the “AX Discovery Series – Builders Track” session titled “AI, code quality & your dev role: Insights from CTOs.” The event is scheduled for November 19th, 2025, at 9:00 am PT / 12:00 pm ET / 5:00 pm UTC. The banner features photos of three speakers: one wearing a black hat, one with long hair in a colorful shirt and a dark cap, and one with short blond hair in a black jacket, all against a dark background with teal design accents.

Join Dana Lawson (Netlify), @danielleleong.com (FireHydrant), and @krider2010.bsky.social (Sailhouse) as they unpack what AI really means for developers today.

Learn how to stay creative, productive, and in control of your code in the AI-augmented era.

🗓️ Nov 19
🔗 www.netlify.com/webinars/ai-...

Very much looking forward to this! Sign up link in the thread.

A couple of Negronis and some goldfish crackers. How’s your Sunday afternoon going?

(For the tech hustlers, I already used some AI tooling this morning - the afternoon is for relaxing 😉)

LLMs aren’t APIs. You call, you get 200 and vibes. With GPT-5, prompts that worked yesterday may fail today; without you touching code. Models drift. Users blame you. Build for it.

Everyone’s trying to build the optimal AI workflow.

Agent infra, eval loops, prompt specs, all wired up before anything works.

Using coding assistants is either wild west or over-specced.

Try things first. Optimise later.

Engineering is the intersection of Curiosity and Capability

When your curiosity outstrips your capability, you expand your capability and grow

When your capability expands, it allows you to see farther and grows your curiosity

Don't allow your tools take away both your Curiosity and Capability

Everyone thinks AI means fewer engineers. But building secure, resilient systems is still hard. AI gives you leverage. Got a 10% output boost? Use it to build better, ship faster, explore more. Replace waste; not talent!

Don’t forget all the nuances and human quirks! But also that makes it interesting. No one team or org is the same.

"You're absolutely right".

I mean it's nice to hear that. But I kinda wish I wasn't so right and didn't need to make these corrections 🤣 Thanks Claude Code!

Debugging async systems feels like ghost hunting. Logs say it ran. Retries say it didn’t. The agent? Forgot the plan halfway through.

You’re stuck piecing together timelines like a true crime board. And half the time, you still don’t know what happened.

What’s your move when async goes sideways?

Vibe reply-guys

“Your agent doesn’t need memory.”

Cool.

Why is it fixing and breaking the same file 3 times in a row?

What’s the weirdest state bug you’ve hit in an AI system?

Glue code is never on the roadmap. But it always shows up.

And if you ignore it long enough… it ends up running your system.

What’s the glue you forgot was critical?

A very good question! Did you get to go? I have FOMO.

No standups. No tickets for tickets. No “please justify this $12 tool.”

Just two people building the thing they wish already existed.

What would you nuke from your team’s process if you could?

Event Infrastructure for Modern Systems AI-native apps generate 50x more events — but most teams are still relying on brittle, legacy infra. Here’s how to modernize your event backbone without breaking everything.

Most event systems were designed for a time when “real-time” meant “sometime this hour.” Now your agents are spinning off retries, branches, fallback calls, and memory loops.

We wrote about why that’s breaking things, and how to fix it without crying: sailhouse.dev/blog/event-i...

Correctness is a nice to have at the moment 😳

Not every event-based system needs a streaming platform.

Kafka’s great if you’ve got a team.

Most of us? We just want something that works, retries, and doesn’t explode in prod. What do you use?

Done well it can accelerate with aligned handoff. Back in the day, I left PRs for colleagues to review while I was offline. If something shipped product reviewed. Next day I picked up review fixes/merges. Jumped on prod feedback, reviewed things left for me. Rinse & repeat. We overlapped 4 hrs IIRC.

Leading teams teaches you new pain:

- Tech debt as strategy
- COGS lurking behind every infra choice
- Simplicity beats perfect

What’s the leadership lesson you wish you’d learned sooner?

Bank holiday Monday. Drinking a very nice Marzen while working on marketing things for @sailhouse.dev . I know how to have a good time 🤣

Retries aren’t free.

Naive retries = failure cascades, duplicated requests, hidden bugs.

Retries baked into infra = graceful failure, predictable outcomes.

Been bitten by retries in production? Tell me your story.

Distributed monoliths: shiny microservices outside, shared DB and deploy pain inside. Infra should be leverage.

Been there? Tell me your best "this was totally decoupled, I swear" story.

You know you love it! Also thank you. Or is it because you need me to ship a preview and not break prod? 😂

That was one I got today!

I am reminded once again how easily even cursor code can get distracted and just want to complete tasks and stub random things out rather than doing a great job. Perhaps we are closer to the average programmer AGI than we realise 😂

But what does ChatGippity say?

32 primary aged new books.

36 secondary school aged books. All new.

I had a clear out today.

Here are two lots of books. I'm giving them away for free, cos I want to.

The first pic to a primary school.
The second pic to a secondary school.

If you'd like them for your school get in touch. I'll choose the schools tomorrow

UK only

PLEASE SHARE

Are you trying to start a food war?!

Everything had to become a (side)hustle :(