Matthew Lebrasseur

CTO at GoEasyCare with over 2 decades of experience in technology and management. Building and leading high-performing teams to deliver innovative solutions.

[Matthew Lebrasseur @TrueCTO]

@mipsytipsy Powerful abstractions help when you can still see how they fail. If the team cannot read what went wrong underneath, on-call turns into folklore and stress. I prefer tools that keep the system understandable. That is how small teams stay secure and ship with confidence.

[Matthew Lebrasseur @TrueCTO]

This is systems over memory, said in data-platform terms. If the write path decides what mattered too early, you lose the evidence you need when production disagrees later. Keep the event. Rebuild the view. Put the commitments in a system so judgment stays available for the hard cases.

[Matthew Lebrasseur @TrueCTO]

@maxedapps This is practical advice for anyone learning the craft. Decently sized tasks, clear tests and a path you can explain on Monday beat a big rewrite most of the time. Coding is learnable. Good systems around the work are what keep the learning useful.

[Matthew Lebrasseur @TrueCTO]

Agents getting better is real and useful. Ownership of production still sits with people. A good harness helps. So do tests that catch what almost works. Someone still has to be accountable when it reaches users. I treat AI as a tool for small teams, not a replacement for that judgment.

[Matthew Lebrasseur @TrueCTO]

This is close to how I try to run priorities. When code gets easier to produce, saying no becomes more important, not less. Review may get cheaper. Owning what ships still takes judgment. Being explicit about what is not a priority this week is one of the simplest systems I know. Saying yes to everything feels responsible. It usually isn’t.

[Matthew Lebrasseur @TrueCTO]

I keep coming back to this as a systems problem. Multiple tools can work if the contracts between them are clear. When they are not, you end up with several versions of the truth and nobody owns the seams. One coherent product is not about tidy logos. It is about protecting the team’s judgment and time.

[Matthew Lebrasseur @TrueCTO]

@MichaelArnaldi This matches what I have seen on small teams. A language change can help, but the bigger win is usually understanding the hot path and removing work you did not need. Measure first. Keep the evidence. Then decide what is worth rewriting.

[Matthew Lebrasseur @TrueCTO]

I have a lot of respect for people who choose to stay close to the work. Hands-on time keeps judgment honest, especially when the tools are moving this fast. If the org cannot make room to learn and ship with new systems, good people will go where they can. For me, building and leading still belong together. When they drift apart, priorities get fuzzy fast.

[Matthew Lebrasseur @TrueCTO]

This is a useful reminder for anyone leading a small team. Internal tools do not need to look like a public platform. Complexity you do not need usually shows up later as slower delivery and harder decisions. I try to keep the architecture matched to the job in front of us, then put the real commitments in a system so judgment stays free for the hard calls.

[Matthew Lebrasseur @TrueCTO]

Wide ownership creates a constant pull to react to whatever is loudest.

Two habits keep that from turning into noise.

First: don't rely on memory for commitments and deadlines.
Give them a system so you can stay focused on judgment.

Second: be explicit about what is not a priority this week.

Saying yes to everything feels responsible.
It usually isn't.

Clarity beats activity when the stakes are high.

[Matthew Lebrasseur @TrueCTO]

Ugly trend from Canadian LTC and disability services: we still treat staffing as a memory sport.

Homes stitch payroll to time to scheduling to agency across four tools that disagree by 7 a.m. Then we buy another dashboard and call it transformation.

The good news is boring. One operating system for the workforce. The bad news is most “AI scheduling” still assumes the roster was honest to begin with.

[Matthew Lebrasseur @TrueCTO]

That survey is the industry in one sentence. Workforce is still the fire. AI is now the agenda item.

I am not anti-model. I am anti-deck. If the night supervisor still has to text a group chat to fill a shift, the priority was theatre.

Fix the schedule and the call-ins first. Get one source of truth. Then the model has something real to chew on.

[Matthew Lebrasseur @TrueCTO]

This is the split we live in healthcare workforce software.

A model can draft a roster in seconds. Keeping it legal and fair and still fillable when three PSWs call in before breakfast is the actual product.

If nobody owns production after the demo, you did not ship. You staged a play.

[Matthew Lebrasseur @TrueCTO]

This is the part hospital dashboards rarely show.

The ER is full because the home is empty of skilled hours. You cannot flow a 90-year-old home in a week if nobody is scheduled to meet them at the door.

We spend a fortune on the acute end and then hope the community schedule holds together with voicemail and goodwill. It does not.

[Matthew Lebrasseur @TrueCTO]

This is the most honest staffing post I have seen this year.

Homes we work with treat overtime like a people problem. It is often a planning problem that started in a spreadsheet nobody on the floor was allowed to question.

If the budget never created the positions, no scheduler on earth can try harder their way out of Sunday night.

[Matthew Lebrasseur @TrueCTO]

@elonmusk @grok @bot @elonmusk when is it coming to Grok Business 🧐?

[Matthew Lebrasseur @TrueCTO]

🚨 CONFIRMED MALWARE — Fake “GPT-5.5 Local” GitHub repo https://t.co/1pnwDiQWhA
File: gpt-5.5-Thinking_x64.7z → gpt-5.5-Thinking_x64.exe
Sandboxes confirm: • Process injection into ipconfig.exe • VM/sandbox evasion (reads VMware registry keys) • Spyware (ipconfig fingerprinting + credential APIs) • 67 MITRE ATT&CK techniques
Hybrid Analysis report: https://t.co/lqxKSQHezA.
This is NOT GPT-5.5, it’s malware riding today’s real OpenAI release. DO NOT DOWNLOAD. Report the repo.
#AIscam #GPT55 #CyberSecurity @OpenAI @github

[Matthew Lebrasseur @TrueCTO]

Problems aren't the real issue, it's how you handle them. I dwell on what I can change, let go of what I can't. Shifts stress massively.

Your strategy for uncontrollable stuff? 👇

[Matthew Lebrasseur @TrueCTO]

@str8edgeracer This is hilarious. Start indoctrinating these little one early, half the adult population still stares at their pay stub like it’s written in ancient Klingon.

[AdvantAge Ontario @AdvantAgeOnt]

Stay on top of changes to Ontario laws and regulations, including the new PSW Authority, case law updates, and more! Human Resources Law: The Year in Review (2025) is now streaming at Learning Hub sponsored by @GoEasyCare

https://t.co/5CJZrGQ7EO