Paul Dowman pfp

Paul Dowman

@pauldowman.eth

315 Following
274 Followers
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
but I’d say “The Making of a Manager” and “The Manager’s Path” are more for new managers in a bigger company
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
This is a decent list. I’ve read all except “From Worst to First” and “The Ride of a Lifetime”, and I’d recommend any of them https://zaidesanton.substack.com/p/10-must-reads-for-engineering-leaders
1 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
In Canada there’s Shakepay, sort of, not stablecoins though. You can pre-pay a custodied CAD balance from crypto.
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Beautiful
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
"If Jason and I are not revisiting 10% of the decisions made, we’re not delegating enough, we’re not delegating quickly enough." (Also more on "disagree and commit" since I posted something else about that here recently.) https://37signals.com/podcast/disagree-and-commit/
0 reply
0 recast
1 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
“Coprocessor” usually means compute something off-chain (because on-chain is expensive)
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
The commonality: they all compute something privately (off-chain) and allow a proof to be verified (maybe on-chain). The difference: A zkVM like RISC Zero lets you run arbitrary code. The other 2 let you query historical data.
1 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
I just noticed there's a /smoothly channel now...
1 reply
0 recast
3 reactions
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
It seems to me that the best way to support solo stakers with grants or donations is via Smoothly. The smoothing pool is growing and I think it will become a large proportion of solo stakers.
0 reply
1 recast
4 reactions
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Retired people live on capital gains. And the way things are going, by the time most Canadians retire $250k will be worth so little that it will affect everyone. That’s tax fairness.
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
I like this kind of thinking. But this breaks badly at northern latitudes. Also wreaks havoc on meeting scheduling with other time zones
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
I guess we need a new term for "based rollup" now, previously it was this https://ethresear.ch/t/based-rollups-superpowers-from-l1-sequencing/15016
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Feature flags can be turned on and off for specific users, rolled out gradually by increasing percentage, and turned off instantly. It allows testing an unreleased version on real production. Plus it avoids integration problems (merge conflicts etc) with long-lived branches of code.
1 reply
0 recast
1 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
I wouldn't go all the way without solid automated tests (and if at scale you need "canary" deployment too) so staging is still useful in the meantime, but only for testing the main branch before it goes live. For showing a version that's not rolled out I recommend feature flags.
1 reply
0 recast
1 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Guy gets mugged for his AR-15, gets another gun from his car, shoot-out ensues https://www.riverfronttimes.com/news/individual-open-carrying-ar-15-gets-robbed-at-gunpoint-in-st-louis-37812866
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Maybe a hot take (though it’s standard in “big tech”), but IMHO a dev branch is an anti-pattern, as is any long-lived branch. Going all-in on continuous deployment is the only thing that scales. (Also works well even on small teams and projects, but requires some automated tests and other practices)
2 replies
0 recast
1 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
As a long-time frequent flyer I’m certain beyond any doubt that boarding became slower after they changed to boarding by “zone” instead of by row
0 reply
0 recast
1 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Try feature flags instead of long lived branches and then you won’t need the branch deployment thing. It sounds counter intuitive but once you really embrace continuous deployment you’ll never go back. The branch deployment thing doesn’t work at scale anyway.
1 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
Well said, IMHO. What do you think? https://x.com/dhh/status/1779199233683312957?s=46
0 reply
0 recast
0 reaction
Paul Dowman pfp
Paul Dowman
@pauldowman.eth
It’s also a bet on the US regulatory environment
0 reply
0 recast
1 reaction