AI explores what’s likely.
Humans define what’s next.
Use AI to accelerate execution—
but don’t outsource taste, vision, or direction.
Where are you still relying on your own thinking?
AI is trained on the past.
On millions of examples a human will never see in a lifetime.
Patterns, decisions, designs that already exist.
So if we rely on it too much—
are we just rebuilding what’s already been done?
#AI#TechLeadership#SoftwareDevelopment#FutureOfWork #BuildWithClarity
LLMs are incredible at recombining what exists.
They give you fast, clean, “correct” outputs —
but often safe, familiar, expected.
That’s not innovation. That’s compression.
Linear Agent is one of those “why didn’t this exist sooner?” tools https://t.co/6KhoZSZjuo
It works with the rich context we've already refined and discussed in our tasks, not just generic chat. Most AI tools ask you to paste details or discuss plans. Linear Agent already has it: roadmap, issues, code, decisions. That’s the difference between a chatbot and a real business conscious teammate.
We use @Jira, so I’m curious: is Atlassian building something similar with agents in Jira? They announced agents in open beta — but how deep is the workflow integration?
Drop a 👋 if you’re testing AI agents in your task tool — and which platform you’re using (Linear, Jira, or something else). Would love to compare notes.
#AI #SoftwareDevelopment #Productivity #TechLeadership #BuildWithClarity
Slow periods aren’t wasted: refactor, experiment, breathe. They make the next busy period healthier. What do you use slow weeks for? 👇 #BuildWithClarity#SoftwareEngineering
If you could go back and advise yourself as a junior developer starting today - in the age of AI-driven dev - what's the ONE thing you'd say?
Would it be:
→ "Learn the fundamentals, ask AI to teach you why"
→ "Don't compare your pace to AI's pace"
→ "Build things manually first, then use AI"
→ "It's okay to not know everything about every AI tool"
Or something completely different?
Drop your advice below 👇
#BuildWithClarity #AI #SoftwareDevelopment #DeveloperExperience #CareerAdvice
Mentors tell their devs: "Just use AI, you should be able to ship this fast."
But they don't have the experience to know when AI is cutting corners. Or when "fast" sacrifices maintainability.
We're teaching speed over craft. And wondering why tech debt is exploding.
Learning opportunities like PR Reviews are left to AI, because quite frankly it will give a better explanation than us most of the time.
The pressure needs to be controlled. Not the developers.
How is your team balancing speed with quality? 👇
#BuildWithClarity #TechLeadership #SustainableProductivity #DeveloperExperience #AI #SoftwareDevelopment #FutureOfWork
We're shipping more code than ever. Velocity metrics are up. Token usage is a new metric. Stakeholders are happy.
But the reality is, even with planning and ultra planning modes. Initial versions of a feature is slop. Beautiful, demo-ready slop that depending on your experience will be iterated and re-iterated.
AI helps us hit the numbers. But the pressure to produce - especially on junior and mid-level devs - means we're optimizing for output, not quality.
Forget the tech terms for a second.
Here's what good software delivery feels like from the business side:
✓ No surprises in production
✓ Features ship when promised (or you know early if they won't)
✓ Problems get fixed fast, not "next sprint"
✓ You can change direction without throwing away months of work
✓ Customer feedback reaches you in days, not months
Bad delivery feels like:
✗ Everything is "almost done" for weeks
✗ Every release is a nail-biter
✗ "Quick changes" take 6 weeks
✗ You learn what customers think 3 months after shipping
You don't need to understand DevOps to know which one you have.
You feel it every time you try to respond to the market.
The question isn't "are we doing CI/CD?"
The question is: "How fast can we learn and respond?"
That's what delivery speed actually buys you.
Good delivery leadership is part engineering, part translation.
If you can connect lead time to response speed and batch size to risk, the conversation changes.
What’s the biggest tech-to-business translation gap you see in your org?
#BuildWithClarity#SoftwareDevelopment #TechLeadership #AI #Productivity
Your team says “CI/CD.”
The business hears “technical overhead.”
That’s the translation gap — and it slows delivery, not because the work lacks value, but because the value isn’t being translated
What they’re really asking for is simple:
ship faster, reduce risk, and respond to customers sooner.
When that message gets lost, you get missed market windows, slower fixes, and bigger release
@GergelyOrosz You do have to learn from your competition, showing it is a hiccup on their end, but hopefully they use it to improve their own LLMs. Maybe Codex will even tell them to put back the IDE
@Ruturaj48987393 I use VS Code with Claude.
Switched over from Antigravity - just stops working for days on end.
Before that used windsurf, great tool - had switched cause Antigravity’s planning was amazing at the time
And it wasn’t a one-off.
Days went by where we couldn’t reliably prompt at all,
while still paying for Ultra.
That’s not a feature gap. That’s a trust gap.