Matt Mazzega

Product Builder from thinking to shipping | Co-founded @tinkso_com (acquired)

[Matt Mazzega @_matmaz]

https://t.co/meCKb8bi6d

[Matt Mazzega @_matmaz]

@LucaCaponeX Well, in cowork you see the skill name in your own cowork when you use it. But at the moment you want to see what/how skills your team use, you get almost nothing….On Claude code, what if you scan your session logs (the jsonl files stored locally) ?

[Matt Mazzega @_matmaz]

My agency installs Cowork for clients. Once we ship custom skills, we go (almost) blind — telemetry just says "Skill: 3 & tokens:250". So I built Argus: captures real sessions (hooks, not OTel) across your team, ask user feedback within sessions, and surfaces qualitative patterns across them. Already a game changer for me to actually evaluate/refine skills,MCPs, adoption across the team etc.. Waitlist → https://t.co/1W5yDww5ji DM me if this resonates!

[Matt Mazzega @_matmaz]

@LucaCaponeX Indeed same thing tailored to Claude code might be something to think about 🤔👍

[Matt Mazzega @_matmaz]

Proud to launch Timelaps on PH today! A product I've helped building for the past months. If you like what you see, happy to get your support! https://t.co/6ZjWjfmspE

[Matt Mazzega @_matmaz]

@LinusEkenstam 'Half your experience and twice your curiosity' — already seeing this. Domain experts who just discovered AI coding tools are shipping faster than most devs I know.

[Matt Mazzega @_matmaz]

@levelsio The 'use your own product daily or quit' filter is underrated. Killed 3 of my ideas last month. But the one that survived? Can't stop opening it.

[Matt Mazzega @_matmaz]

The first wave of AI coding was about speed.

The next wave is about control.

Lint rules for agents. Spending caps. CLAUDE.md guardrails.

The boring stuff that stops an agent from burning $40 in 10 seconds is where the real edge is now.

[Matt Mazzega @_matmaz]

The hardest part of building tiny tools isn't the code anymore. It's deciding which of the 10 ideas in your head deserves this weekend.

AI made building fast. It didn't make choosing easier.

Most solo builders I know aren't stuck on execution. They're stuck on selection.

[Matt Mazzega @_matmaz]

@george_nqu Curious what it will output 3 years later!

[Matt Mazzega @_matmaz]

Playing with an @OpenAI fine-tuned model trained on openbuild library to generate @Bubble components. All in all, the challenge might not be technical but actually human. How non-tech people will ask the machine what they want? The future of no-code will be exciting! #nocode #ai

[Matt Mazzega @_matmaz]

@george_nqu @D_Mack7 Replace ‘vibe-coding’ with ‘no-code’ and you’ll read the same (poor) critics that we were fighting against years ago…do you really understand entirely how a Bubble app works?

[Matt Mazzega @_matmaz]

@mabanque_bnpp Et un petit outil (gratuit!) pour verifier tout cela : https://t.co/xJlwozbiZF

[Matt Mazzega @_matmaz]

@Ajcimmobilier06 Et un petit outil pour verifier tout cela : https://t.co/xJlwozbiZF

[Matt Mazzega @_matmaz]

@enerzine Et un petit outil gratuit pour simuler tout ca: https://t.co/xJlwozbiZF

[Matt Mazzega @_matmaz]

@yelkhayami @TipsBubble Claude code should be able to achieve that 😉

[Matt Mazzega @_matmaz]

5/ **Real Agency Transformation: The Numbers** One enterprise client needed a complex data dashboard. Traditional approach: - Timeline: 8 weeks - Team: 3 developers - Cost: $96,000 - Bug fixes: 2-3 weeks post-delivery AI-augmented approach: - Timeline: 5 weeks (-37%) - Cost: $60,000 (-37%) - Bug fixes: 3 days (-90%)

[Matt Mazzega @_matmaz]

4/ The AI-Augmented Economics

Same project, different math:
- AI-augmented developer productivity: +40%
- Quality improvements: -35% bugs
- Client delivery time: -35% faster
- Client satisfaction: +25% increase

Result: Higher client value + improved margins + competitive differentiation

[Matt Mazzega @_matmaz]

5/ **Real Agency Transformation: The Numbers**

One enterprise client needed a complex data dashboard. Traditional approach:
- Timeline: 8 weeks
- Team: 3 developers
- Cost: $96,000
- Bug fixes: 2-3 weeks post-delivery

AI-augmented approach:
- Timeline: 5 weeks (-37%)
- Cost: $60,000 (-37%)
- Bug fixes: 3 days (-90%)