Craig B. (@CraigBarrDev) header
Craig B. (@CraigBarrDev) profile picture

Building power tools for knowledge workers.

Australia
Joined April 2013
Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)Photo by Craig B. (@CraigBarrDev)
a guess but this time next year, probably we review session prompts, tool calls and arch visualisations we move up the review stack because the alternative is undifferentiated valueless slop , it’s close to delivering nothing
totally aligned with the no ai-slop manifesto slopmanifesto.com/
1
1
80
even w/ an army of agentic tools - every online business will need a cyber-manager, just someone who keeps on top of the agents, stream of CVEs, actionable items and makes sure the agent are doing their job! (because the cost of them not doing their job will be too high)
61
did @jasonfried say he was going to open source it or did I dream that ? either way it’s been a month and I want ! wonder if anyone tried to have a crack and open sourced it
For years, maybe a decade, I've wanted to build a writing tool that just did a few things I couldn't find anywhere else. A system that mimicked how I think when I write. A system that provided clear affordances, features, and places to do the things I wanted to do. A way to play with my writing as I'm writing it. A way to settle into just the right way to say something. It never made sense to make it a 37signals product because it's not commercially viable, nor would it be worth pulling people off their other work to hack on this. Plus, it would have taken months the old way. So this weekend I just made it myself, with Claude's help. It's called Write_On and the short video walks you through it. Essentially it's "alternative control" at the word, sentence, and paragraph level, plus a way to dim stuff back, and stash stuff near by. Not version control, but alternative control. You'll see what I mean when you watch the video below. So here it is. Write_On.
98
The trunk-based development debate started by @thorstenball has clearly touched a nerve. So I thought, why don't we look at some numbers? I built a little simulator with the help of our dear colleague Opus to compare four different integration strategies: 1. Trunk-based development (@thorstenball's horse) 2. Merge queue as the only CI gate (what I'm currently using) 3. PR-based, requiring a fresh base 4. An agentic merge queue, with built-in conflict resolution (dream option, can't buy it anywhere yet) You can play with the simulator here: lpalmieri.com/simulations/co…
22
16
10
240
40,032
probably should have seen this coming, herdr tui (upstream) just dropped a desktop app. same named binary, so the open source desktop predecessor that was not "affiliated with Herdr" but called Herdr literally gets overwritten might have to rename the other so i can try them out side by side. names still the biggest problem in computer science
calling it. this is the best cross-agent multiplexer do your self a favour and start coding w/ it now
1
197
i've called the @fabienpenso one "HerdrWorking" because its the only working, the new @herdrdev desktop didn't load any of the existing sessions. has the herdr core fragmented somehow? i'm a bit perplexed tbh
1
1
4
353
all of these AI bangers are breaking my daily workflow. i have generally consumed music via adding to my spotify playlist, but we are shipping bangers faster than they can appear in spotify ... so should i just cancel my spotify and use youtube music? atm a custom solution would be better at integrating music from all sources (local, third-party, traditional distributors etc) - take not spotify PMs (move fast i'll prob stay) youtube.com/watch?v=zONGVaON…
104
a banger... even though i don't fully agree with the bleak premise! youtube.com/watch?v=zONGVaON… i explain why below
it surprises me sometimes that anyone could be worried about being replaced by ai. here's why. if you're in gergely's first group, you've probably already discovered that the best results come from building on ai's work, refining and compounding it. not because the old ways of working still apply unchanged, but because meaningful outcomes still require craft. the tools, techniques and even the nature of the work are changing dramatically. but the idea that you can simply throw a goal at ai and consistently get exactly what you and your users need? experience tells us otherwise. there's still an art to this. and the most exciting part is that we haven't even figured out what that art is yet. we're discovering it as we go. the second group will be fine too, once the first starts turning that discovery into clearly defined roles, backed by evidence of what works. we've always had pioneers who thrive on figuring out what's next but would go mad having to operate it day to day. ai isn't eliminating that distinction. it's giving us an entirely new frontier on which to discover it.
84
it surprises me sometimes that anyone could be worried about being replaced by ai. here's why. if you're in gergely's first group, you've probably already discovered that the best results come from building on ai's work, refining and compounding it. not because the old ways of working still apply unchanged, but because meaningful outcomes still require craft. the tools, techniques and even the nature of the work are changing dramatically. but the idea that you can simply throw a goal at ai and consistently get exactly what you and your users need? experience tells us otherwise. there's still an art to this. and the most exciting part is that we haven't even figured out what that art is yet. we're discovering it as we go. the second group will be fine too, once the first starts turning that discovery into clearly defined roles, backed by evidence of what works. we've always had pioneers who thrive on figuring out what's next but would go mad having to operate it day to day. ai isn't eliminating that distinction. it's giving us an entirely new frontier on which to discover it.
I'm starting to sense a divide in the tech industry: - People thriving on autonomy who are reinventing how they work and what tools they use with AI: LOTS of work, but they're enjoying it and getting very efficient - People used to being mentored / trained / or simply told what to do and how. Frustrated with the lack of guidance from management on what to do with AI; also frustrated by seeing some of their peers thriving; increasingly bitter about AI and the state of the industry, as a whole Neither group seems to be aware that there's another group, or have empathy for the other group
1
101
I know I'm supposed to think "wow, how cool of slack!" but as an Aussie who enjoyed a solid 20 blissfully halloween-free years on the foreshore, watching Instagram turn it into an annual arms race of manufactured enthusiasm has been quite something and now workplace software wants in 🏳️🏳️🏳️
65
.env files for dev. lots of people completely over-engineer here and miss the point. it’s generally not bad to commit a local only .env that everyone uses, it’s not your prod environment. stop introducing 3rd party secret managers for local dev or sending .env files around. if you genuinely can’t commit it to the repo your problem is that you are missing safe non-prod deps, fix that first
Highly recommend architecting your projects in such a way they can be easily run in dev without any setup or env vars Then when u need to run it with the envs use some secrets manager so u just need one key (infisical is the best I’ve found for this)
76
great read. juniors have the benefit of not having learnt all the ways something is “impossible”
The coding interview is dead, at least for my team at Anthropic. I've seen enough to know that it no longer provides signal. I removed it from loops this month. I have a spicy, optimistic take - I think there's an opportunity for junior engineers: felixrieseberg.com/hunger-be…
61
Hear me out. Just spin up an agent for each dependency you have and have it be the babysitter for that dependency. It is a surprisingly effective pattern. It only opens PRs when I’m actually vulnerable and can even find vulns before they are lifted from embargo.
fyi dependabot time to open a PR is median 8 days
6
8
68
6,065
bummer i was hoping it was just me
in brisbane, australia... it’s not possible to get stable internet on a train commute into the city. it drops out in the same 5 km stretch from the bayside to the city every day. completely reproducible. solution is to wait out those 15 mins on the train before it gets good. is it foolish to expect an olympic city to be capable of providing an internet connection on a daily commute?
68
in brisbane, australia... it’s not possible to get stable internet on a train commute into the city. it drops out in the same 5 km stretch from the bayside to the city every day. completely reproducible. solution is to wait out those 15 mins on the train before it gets good. is it foolish to expect an olympic city to be capable of providing an internet connection on a daily commute?
1
139
you might not want to hear this: your ability to collaborate with AI indefinitely isn’t guaranteed. it is something you’ll have to earn, and keep, through discipline quizzes. explainers. microworlds. whatever it takes to leave each AI session with genuine understanding because the goal isn’t just to keep a human in the loop it is to remain a human capable of contributing to the next loop
1
93
I have been working on this Cookslop, an app for helping you with recipe ideas, shopping and cooking. You can use it for free right now with this invite link. cookslop.com/invite/TGui3pxQ…
2
1
5
167
announcer on this brisbane train now spruiking industrial action updates, that was weird "we have restored 91% of services across brisbane" "the next station is southbank"
1
132