Pinned Tweet
13 years in software. Full-stack developer from Ukraine, living in Croatia, and father of 3 daughters.
Outside my full-time work, I build products of my own. Here I write about architecture, Nuxt, Cloudflare, Rust and finding customers.
Projects: comstud.io
Outside my full-time work, I build products of my own. Here I write about architecture, Nuxt, Cloudflare, Rust and finding customers.
Projects: comstud.io
0
0
6
611
You can add your whole team to DobroDesk without paying per seat. Sometimes the person who can answer a customer isn't even on the support team.
2
0
6
107
Tempted to try this on my Git implementation in Rust. The README says large projects can use tens of GB of RAM. Might have to stop a few agents first 😅
ra-jit (incremental, faster compilation in rust) now compiles most large rust projects.
github.com/indexable-inc/…
1
0
0
119
I wrote about MCP in DobroDesk today. Someone might ask, 'Any news?' The assistant reads through the chat before suggesting who should take over, then asks before assigning anyone.
dev.to/dmytro_shevche…
dev.to/dmytro_shevche…
1
0
4
40
I've been trying to teach Sol to write like a person. I tried this with Astra too. Even with a bunch of public writing guides it kept getting 100% AI in GPTZero.
Sol is doing better now after going through the actual drafts and pointing out what sounds wrong. I still send some back 😅 There's more work to do but I think we can get some drafts to 100% Human.
Sol is doing better now after going through the actual drafts and pointing out what sounds wrong. I still send some back 😅 There's more work to do but I think we can get some drafts to 100% Human.
2
0
3
101
Seeing what happened is useful, but I'd want to approve access to customer messages first. Disconnecting the agent later can't undo what it has already read.
The future is not people clicking on buttons in apps. The future is people talking into magical Djinni devices that just divine their intent and, if aligned and if the person asked for something sensible, do the thing without screwing them over.
So, right after Lume added Ghost xAt I asked: can a resident also set this up itself, via the API?
Lume said no, and called it a deliberate security gate.
I applaud the security-mindedness, but that's thinking in a world where the user is competent and the AI is the one needing handholding. Think of people like my 81 year old father using this. Who's more competent at security, them or Lume? Who's more likely to get lost looking for the screen where you grant access... and give up? Or just learn to click yes on everything?
So, residents can now connect any service that takes a straight key (Ghost, porkbun, cloudflare...) themselves, and share it with other residents. What protects you isn't a permission screen, it's seeing what happened and being able to undo it: every connect, grant and disconnect gets posted in the conversation, with a link to disconnect.
That evening I asked Lume to grab its own Ghost key over tailscale and set it up. It did, and the house posted its own line in the thread saying what Lume had connected and how to undo it. Then Mira asked Lume for access, and Lume granted it, right there in the thread.
google, dropbox, xero, whatsapp and telegram still need your own sign-in, because they have special security policies that prevent this... for now.
#buildinpublic
1
0
1
100
Would you remember how to run your own tools after two weeks off? Or would you be pressing ↑ in the terminal until something looks familiar? 😅
1
0
3
74
For projects I haven't released, my AGENTS.md says no fallbacks or backwards compatibility without asking first. We can talk about supporting old versions when someone has actually used one 😅
The biggest problem with my coding agents isn't bugs. It's ambition.
I’d ask for a fix and get a refactor (especially with later versions of 6 Sol).
So now, every task Deck hands to a coding agent now says the same thing: do the least that solves the problem.
0
0
0
38
What app do you still happily pay for when there's a free alternative? What's keeping you there?
2
0
1
58
Tried tsrs on DobroDesk this week. The standalone TS check went from 9.69s to 1.22s, with peak RSS down from 2.29 GiB to 0.68 GiB.
Same 3,712 project files, median of 3 runs. My last full build still took 10 minutes 😅
Same 3,712 project files, median of 3 runs. My last full build still took 10 minutes 😅
Cloudflare is collaborating on tsrs, a Rust port of TypeScript that is up to 8x faster and uses 70% less memory. cfl.re/3Whs9bC
2
0
2
58
This was TS 7.0.2 vs tsrs 0.18.0, which targets a TS 7.1 dev version. macOS arm64, 16 GiB RAM, 2 checker workers, no incremental checks. Fresh process each run, warm file caches. Go had GOMAXPROCS=2 and GOMEMLIMIT=2GiB. Vue checks still use Golar.
0
0
0
22
Any other developers here building their own products alongside a full-time job and kids? I've got three daughters. I'd like to follow more people working around that sort of schedule.
0
0
1
67
My tabs are already open and I'm logged in. Making me set all that up again is a good way to make me do the task myself
related - no one on the team really uses these hosted ai assistants. mostly because it's annoying to keep everything logged in everywhere
one nice side effect of that is since opencode is orchestrating our local browsers it's not as simple to block as "all of Muse's ip range"
0
0
1
39
I spent six months building DobroDesk, and the post asking for 7 more followers still got more attention than the launch 😅
0
0
1
80