Excited to be launching Basiq.Work — AI agents & automations to help product & eng teams move faster, stay focused, and kill blockers.
Taking on the CTO role, and honestly, couldn’t be more fired up to build.
Sound on for the announcement video. 🥁
Week’s takeaways:
– The rabbit hole of product & eng improvements is deep.
– Everyone wants faster feature delivery—no shock there.
– 2-week rule is sparking great conversations.
– MVP tests on AI audio for feature qualification—promising.
– Veo 3? Impressive.
AI is already helping product & eng teams — but there’s a ton left to build.
If we want faster feature delivery, tools should attack the manual junk in the dev process.
Start with PRDs. Or planning. Or handoffs.
As AI platforms compete for dominance, I’ve been thinking: what would actually make me loyal to 1 provider over another?
Price still carries a lot of weight — especially at scale — but maybe the real differentiator is devEx. If building becomes frictionless that might be enough to tip the scale.
@developer.microsoft.com I am trying to get approved for an MPC account for a domain which is **hosted on your platform** and yet you keep rejecting me saying I am not providing the necessary proof of ownership. Can you help?
Sorry what are we talking about?
The difference between “has agreed” and “the deal has closed.”
It was reported 3 weeks ago that OpenAI decided to buy Windsurf - and many on social media treat it as fact that it happened.
But until OpenAI and Windsurf announce something: it’s not official + might not be true.
Working with Turso on my new project: SQLite in production (yikes!).
Honestly kinda awesome. Forces you to rethink architecture — many small DBs, fast local reads/writes.
Most exciting part?
Delete your account = I delete your whole DB. Instantly gone. Clean.
What could go wrong.
I think the opportunities in a startup can’t be matched by a large established firm. Doesn’t mean it’s the right choice but I do that explains why people still take the risk
Wrestled for an hour on my front end today with … mailto links 🤦 - how have we not figured this out yet
Working through a theory:
The #1 killer of Product/Eng progress is long-running features.
My radical take:
⚠️ Any feature that takes >2 weeks? Don’t start it.
💀 Any feature that goes >2 weeks? Kill it.
Extreme?
– They suck up time + resources
– Quick wins usually deliver more value
Thoughts?
I just talked to an AI for 20 minutes (audio not chat) to work through a technical problem. Wild.
Agile was supposed to help us ship faster.
Instead, we got bloated rituals, Jira fatigue, and more overhead than output.
Most leaders I know don’t follow Agile dogma anymore.
The future? Leaner workflows, less process, more delivery.
AI is magical—until you ask it to debug your k8s cluster and it starts gaslighting you with fake YAML. We need to increase the deep technical voices in the room because if you haven’t gone 8 prompts deep into a tech rabbit hole, you haven’t really met the limits yet
We’re still just scratching the surface of what AI-generated video can do—but the quality is already staggering. This was generated in minutes with Google Gemini.
I’m curious to see how this impacts startup marketing in the coming months.
Wait, innies are just docker containers for a person
I keep thinking about the coming collision between pre-AI and post-AI apps. Integrating new tech into existing architectures has always been harder than building fresh. Will AI-first apps outpace those trying to retrofit, or will the incumbents find a way?
Hot take: If you’re early in your career, work from an office every chance you get.
Your network is everything when it comes that 2/3rd job. People remember those they see every day > ones they occasionally chat with on Zoom.
Folks you work with now will move all over—make those connections count.
San Diego at its finest
I feel we’ve lost the thread with OAuth. I get it—auth is hard but signing in just to get a code to exchange for an access token and an ID token, plus a refresh token you need to get a new access token, which you then have to store indefinitely for continuous access? Madness.
LLMs work with known data, which isn’t a flaw—it’s how most progress happens. Regular advancements come from refining what exists, but IME big leaps often come from unknown insights breaking convention. AI could get there, but how it makes that leap is still unclear to me
Engineers showing up to the company Christmas party …
Arc has really grown on me—especially how they’ve nailed Spaces. Separating work this seamlessly is something I didn’t realize I needed.
But with all that promise, it was honestly sad to watch @t3.gg video about the state of the browser. Worth a watch: youtu.be/dPUzOQdUFSg?...
Hey Bluesky! I’m Joe—2x founder, CTO, developer. I’ve spent a lot of time scaling startups and leading eng teams, but what I love most is working with other builders and learning from the community.
Still figuring out the best way to plug in here—if you’re working on something cool let’s connect!
