PokeTrade's pack opening no longer takes the whole app down on Android test devices that can't draw it.
The 3D renderer aborts on some graphics drivers, and no try/catch can stop that. So my agent draws one test frame in a throwaway process first. If that process dies, the app lives and hides the pack.
When a crash can't be caught, let something disposable crash first.
BugBoard 0.5.5 is out right now: one bugboard for your team and your AI agents.
Agents connect over MCP (75 tools) from Claude Code, Cursor, VS Code, Windsurf or Codex. Six free tools work without an account, like a duplicate check before you file a bug.
BugBoard is the tool I've been using for over a year now for all my projects.
One PokeTrade CI job went from 34 minutes to under 18.
It built four app images one after another. Now each builds in its own job, side by side, so the wait is only the slowest one.
Same builds, half the wait, on every pull request my agents open.
1,154 of the 3,865 events in BugBoard's analytics came from localhost. The script tag had no list of allowed domains, so our own development was recorded as site traffic. The fix is one attribute naming the site's own domain. It is merged, not released yet.
This is a lesson to adjust your analytics platform ( for me is Umami ) to not record AI activity.
I built a column into BugBoard that has to stay empty. New tasks land in Triage. A dispatcher tells the project's agent within a minute. The agent starts the task or parks it: Waiting, or Blocked when it needs me. After 10 minutes in Triage the dispatcher reminds the agent.
I am using self hosted Coder with Claude connected with MCP on BugBoard to set this up. That way I manage over 5 projects this is basically my new team of engineers
Working a 9-5 is 50% work 50% doing the therapist to the ones above you. Meetings 3-4 hours daily because your boss did not fuck his wife last night.
This is why working for yourself is 10X better , prefer contract work over 9-5
Also in Greece you get paid pennies and your manager usually takes Xanax to wake up in the morning.
Right now big tech are racing to who is going to make the best AI model. The AI race for us is who can spent more on AI because the more you can spent the bigger advantage you have against your competitors.
A single developer can outpace big tech companies now with AI because big companies cannot adapt fast enough to AI Driven Development , indie hackers or small teams can.
Cloud development is the future - nobody will work on their laptops anymore in a few years from now. Having 3-4 big machines writing code 24/7 and testing non stop is much better than having a laptop
I already have 144GB 72vCore 1.2TB NVME servers and all of my development is through there.
What evaluation do you think trustmrr could reach ? I mean you’ve made 3 million from indie hacking and you get an offer for 2.5mil
It’s not like Facebook when Zuckerberg rejected a big offer - we are talking for a static website here it doesn’t have a long lifespan probably.
Maybe I would do the same and not sell I don’t know 🤷
@Bagdex This is why the global banking system is going to collapse because all banks suck. The usually block payments that they don't like. This is why everyone is using Revolut in Europe
The future of payments is in crypto either we like it or not
@marclou Why don’t you try to vibe code a real business from a fresh account nobody knows ? I think most of you saas-influencers have a huge advantage compared with someone with no fame.
What do you think ? Does your influencer status plays a role in the success of the business?
Cool app it is open source, I actually did a contribution as well. E2E encrypted communications over BLE
It is only for Android right now I installed it with Obtainium
https://t.co/q5WG9F27ef
https://t.co/tmZXKGClAe
Έφτιαξα τον ΠΥΡΣΟ 🔥
Δωρεάν εφαρμογή Android για μηνύματα όταν πέσει το ίντερνετ. Τα μηνύματα ταξιδεύουν από κινητό σε κινητό με Bluetooth, με post-quantum κρυπτογράφηση. Χωρίς servers, χωρίς λογαριασμούς.
https://t.co/zsqV9b7SoZ
It's not your AGENTS.md you need a complete architecture. Here's mine :
- Main Agent ( Opus 5.5 )
- Task Agent ( Sonnet 5.5 )
- Performance Agent ( Sonnet 5.5 )
- Security Agent ( Sonnet 5.5 )
- Test Agent ( Sonnet 5.5 )
I run a Coder Cluster on a bunch of VMs I only talk on the main agent with special instructions to basically
- Do Research and HandOff everything on multiple task agents and work in parallel
- When task agents are finished they open a PR
- Performance and Security Agents review and comment inside the PR on a loop until nothing more id there to fix
- Test Agent tests everything on a loop until everything is working with evidence in json/mp4/jpeg depending on tasks nature
The result I get back most likely it is working with no bugs. If i approve Main Agent deploys everything.
Also I have cron jobs for Daily, Weekly and Monthly security checks which the check Grafana, GlitchTip, Crash Logs and automatically open PRs, fix and deploy everything.
This allows me to merge daily over 100 PRs across multiple projects of mine with minimal bugs.
BTW A vibe coder can't setup such ecosystem.
I always asked my self why everyone uses traditional cloud and they don't just self host everything.
For me this is stupid = $230 a month to $4,300
I guess for some people money just don't matter. I use OVH and Contabo for everything. I rent over 500GB of RAM, over 5TB of NVME and over 190 vCPU cores for less than 600$ per month
The only downside is that I have to setup all devops myself.
And I see startups and companies have a 50k bill per month on AWS, CICD etc with half the machines I own.
BTW The article is good I agree all of these things speed up my CI as well by a lot.
I always asked my self why everyone uses traditional cloud and they don't just self host everything.
For me this is stupid = $230 a month to $4,300
I guess for some people money just don't matter. I use OVH and Contabo for everything. I rent over 500GB of RAM, over 5TB of NVME and over 190 vCPU cores for less than 600$ per month
The only downside is that I have to setup all devops myself.
And I see startups and companies have a 50k bill per month on AWS, CICD etc with half the machines I own.
BTW The article is good I agree all of these things speed up my CI as well by a lot.
@orama_network@rootwalletio Read this again in 3 years from now then you will understand the vision behind Orama Network which literally translates to "Vision"
2027 is going to be double than 2026. People don't understand that the longterm solution to everything is Orama Network. Orama is not just a decentralized network it is a way of building applications that are local first and they work offline. Servers are becoming more and more expensive and companies will start to realize that they will have to change the architecture of their apps so user devices store/process data as well.
Up until today everything was done by the server, servers in the future will be contract signers. 90% of people globally have computers and phones that they don't even utilize 10% of their power.
This is how AnChat and RootWallet are working.
Github will just not be able to keep up with the growth and they will need to make everything decentralized and distributed when they realize that it will be already too late for them.
Orama Network has seen that years ago and we are already working towards this future.
The only reason this wont happen is that if somehow chips become 20X cheaper and we have 20X more production than today, I don't see that happening anywhere near the next 10 years and even if it does happen still it will not be enough.
Decentralized architectures and WEB3 is the future !