Maciej Myslinski (@macmys) header
Maciej Myslinski (@macmys) profile picture

Maciej Myslinski

@macmys

I build the AI systems I talk about. Spec-driven development, agents in production, honest numbers. https://teapaw.org/t.co/ce9kkQfcsR co-founder · PhD · Leuven, BE · skiing

 Leuven, Belgium
linkedin.com/in/macmys
Joined April 2011
  • Tweets781
  • Following853
  • Followers363
Anthropic and GitHub both shipped "what did the AI cost, per piece of work" this week. Neither shows whose budget the saving lands in. That's still a spreadsheet someone has to own.
Screenshot of Anthropic's Claude Enterprise smart reports: a Workstreams panel sorted by percent of spend, listing work such as launching a marketing campaign, preparing daily briefings and developing website modules, each with its share of sessions and its dollar spend. There is no column for savings.

ALT Screenshot of Anthropic's Claude Enterprise smart reports: a Workstreams panel sorted by percent of spend, listing work such as launching a marketing campaign, preparing daily briefings and developing website modules, each with its share of sessions and its dollar spend. There is no column for savings.

5
We bid a fixed price on a small app. The client chose an offer with zero build cost: the developer takes a share of revenue per active user. Cheaper code doesn't remove the risk. The pricing model decides who carries it.
2
Engineering got a new line in its budget: the token bill. The hours it saved aren't a line anywhere. They turn into money only when another department changes.
Diagram: AI agents speed up Engineering and its token bill rises. Work flows to Back office (hours saved: hire fewer? move people?) and to Sales, where cards pile up (learn AI? new roles? new process? check the AI's work?). Who owns the change next door?

ALT Diagram: AI agents speed up Engineering and its token bill rises. Work flows to Back office (hours saved: hire fewer? move people?) and to Sales, where cards pile up (learn AI? new roles? new process? check the AI's work?). Who owns the change next door?

5
AI made engineering faster. The queue moved next door: sales now has to keep up with what engineering ships, and the back office has to decide what to do with the hours it saved.
3
My agent drafts my comments. One rule: quote the post from the platform the comment goes on. Same Ethan Mollick post, two platforms: the LinkedIn version had a fifth line the X version didn't, and that line was the point.
A writing rule from the file my agent drafts from, dated 2026-08-19: a LinkedIn post quoted from anywhere but LinkedIn is a partial quote. The brief quoted Mollick's X version, four lines; the LinkedIn post had a fifth: 'But a bit weird to be out of a job I have held for decades.' The missing line was the one the comment existed to deliver.

ALT A writing rule from the file my agent drafts from, dated 2026-08-19: a LinkedIn post quoted from anywhere but LinkedIn is a partial quote. The brief quoted Mollick's X version, four lines; the LinkedIn post had a fifth: 'But a bit weird to be out of a job I have held for decades.' The missing line was the one the comment existed to deliver.

9
A rule that took me a summer: say the drafted comment aloud to the person who posted. If it opens with "Agree with…", it's a label. Nobody talks like that.
The rubber-duck rule from my brief-mode checklist: say the comment aloud to the person who posted it; if the opening clause is a category word like 'Agree with…', it is a label, not a reaction, and reads as AI on sight. A person opens with 'Impressive, and I watched it twice' or 'I don't buy the second half.' The posture is spoken, not labelled.

ALT The rubber-duck rule from my brief-mode checklist: say the comment aloud to the person who posted it; if the opening clause is a category word like 'Agree with…', it is a label, not a reaction, and reads as AI on sight. A person opens with 'Impressive, and I watched it twice' or 'I don't buy the second half.' The posture is spoken, not labelled.

5
My morning agent never places a comment, never sends a message, never follows anyone. It can propose anything, because nothing leaves without me.
Rule 3 of 5 from the CLAUDE.md file that runs my branding agent: relationship actions (comments, DMs, connection requests) are prepared by agents, placed by Maciej. Never automate them; never log into or scrape LinkedIn/X.

ALT Rule 3 of 5 from the CLAUDE.md file that runs my branding agent: relationship actions (comments, DMs, connection requests) are prepared by agents, placed by Maciej. Never automate them; never log into or scrape LinkedIn/X.

5
My morning agent doesn't tell me what to read. It tells me what to say. Every weekday at 06:45 a job on my Mac runs Claude Code against a list of people and topics I follow. The first section is called "Do these": three items, five minutes each, the text already written.
Item 2 of the 'Do these' section of my 13 October morning brief: a Marty Cagan essay, the note that no LinkedIn post of it was found, the drafted comment in a blue box, and the check lines under it: plain version, rubber duck, hostile pass.

ALT Item 2 of the 'Do these' section of my 13 October morning brief: a Marty Cagan essay, the note that no LinkedIn post of it was found, the drafted comment in a blue box, and the check lines under it: plain version, rubber duck, hostile pass.

39
Briefing by @TBrijs on how to use @roadvikings at #platoontest #selfdriving cars @mat_maj @Aon_Belgium @vivadriveio
Image from Maciej Myslinski (@macmys)
7
6
Ready 4 #platoontest of self-driving cars by @Aon_Belgium with @imob @roadvikings @vivadriveio @startitkbc @iMinds
Image from Maciej Myslinski (@macmys)
1
6
7
Load more